Systematic KMTNet Planetary Anomaly Search. XI. Complete Sample of 2016 Subprime Field Planets In-Gu Shin1 , Jennifer C. Yee1 , Weicheng Zang1,2 , Cheongho Han3 , Hongjing Yang2 , Andrew Gould4,5, Chung-Uk Lee6 , Andrzej Udalski7 , Takahiro Sumi8 (Leading authors), Michael D. Albrow9 , Sun-Ju Chung6 , Kyu-Ha Hwang6, Youn Kil Jung6,10, Yoon-Hyun Ryu6, Yossi Shvartzvald11 , Sang-Mok Cha6,12, Dong-Jin Kim6, Hyoun-Woo Kim6, Seung-Lee Kim6, Dong-Joo Lee6, Yongseok Lee6,12, Byeong-Gon Park6, Richard W. Pogge5 (The KMTNet Collaboration), Przemek Mróz7, Michał K. Szymański7 , Jan Skowron7 , Radosław Poleski7 , Igor Soszyński7 , Paweł Pietrukowicz7 , Szymon Kozłowski7 , Krzysztof A. Rybicki11,7 , Patryk Iwanek7 , Krzysztof Ulaczyk13 , Marcin Wrona7 , Mariusz Gromadzki7 (The OGLE Collaboration), and Fumio Abe14, Ken Bando8, Richard Barry15, David P. Bennett15,16 , Aparna Bhattacharya15,16, Ian A. Bond17, Hirosane Fujii14, Akihiko Fukui18,19 , Ryusei Hamada8, Shunya Hamada8, Naoto Hamasaki8, Yuki Hirao20 , Stela Ishitani Silva15,21, Yoshitaka Itow14 , Rintaro Kirikawa8, Naoki Koshimoto8 , Yutaka Matsubara14 , Shota Miyazaki22 , Yasushi Muraki14 , Tutumi Nagai8, Kansuke Nunota8 , Greg Olmschenk15 , Clément Ranc23, Nicholas J. Rattenbury24, Yuki Satoh8 , Daisuke Suzuki8 , Mio Tomoyoshi8, Paul. J. Tristram25, Aikaterini Vandorou15,16, Hibiki Yama8, and Kansuke Yamashita8 (The MOA Collaboration) 1 Center for Astrophysics, Center for Astrophysics | Harvard & Smithsonian, 60 Garden Street, Cambridge, MA 02138, USA 2 Department of Astronomy, Tsinghua University, Beijing 100084, Peopleʼs Republic of China 3 Department of Physics, Chungbuk National University, Cheongju 28644, Republic of Korea 4 Max Planck Institute for Astronomy, Königstuhl 17, D-69117 Heidelberg, Germany 5 Department of Astronomy, The Ohio State University, 140 W. 18th Avenue, Columbus, OH 43210, USA 6 Korea Astronomy and Space Science Institute, Daejeon 34055, Republic of Korea 7 Astronomical Observatory, University of Warsaw, Al. Ujazdowskie 4, 00-478 Warszawa, Poland 8 Department of Earth and Space Science, Graduate School of Science, Osaka University, Toyonaka, Osaka 560-0043, Japan 9 University of Canterbury, Department of Physics and Astronomy, Private Bag 4800, Christchurch 8020, New Zealand 10 Korea University of Science and Technology (UST), 217 Gajeong-ro, Yuseong-gu, Daejeon 34113, Republic of Korea 11 Department of Particle Physics and Astrophysics, Weizmann Institute of Science, Rehovot 76100, Israel 12 School of Space Research, Kyung Hee University, Yongin, Kyeonggi 17104, Republic of Korea 13 Department of Physics, University of Warwick, Gibbet Hill Road, Coventry, CV4 7AL, UK 14 Institute for Space-Earth Environmental Research, Nagoya University, Nagoya 464-8601, Japan 15 Code 667, NASA Goddard Space Flight Center, Greenbelt, MD 20771, USA 16 Department of Astronomy, University of Maryland, College Park, MD 20742, USA 17 Institute of Natural and Mathematical Sciences, Massey University, Auckland 0745, New Zealand 18 Department of Earth and Planetary Science, Graduate School of Science, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-0033, Japan 19 Instituto de Astrofísica de Canarias, Vía Láctea s/n, E-38205 La Laguna, Tenerife, Spain 20 Institute of Astronomy, Graduate School of Science, The University of Tokyo, 2-21-1 Osawa, Mitaka, Tokyo 181-0015, Japan 21 Department of Physics, The Catholic University of America, Washington, DC 20064, USA 22 Institute of Space and Astronautical Science, Japan Aerospace Exploration Agency, 3-1-1 Yoshinodai, Chuo, Sagamihara, Kanagawa 252-5210, Japan 23 Sorbonne Université, CNRS, UMR 7095, Institut d’Astrophysique de Paris, 98 bis bd Arago, 75014 Paris, France 24 Department of Physics, University of Auckland, Private Bag 92019, Auckland, New Zealand 25 University of Canterbury Mt. John Observatory, P.O. Box 56, Lake Tekapo 8770, New Zealand Received 2024 January 8; revised 2024 April 2; accepted 2024 April 6; published 2024 May 16 Abstract Following Shin et al. (2023b), which is a part of the “Systematic KMTNet Planetary Anomaly Search” series (i.e., a search for planets in the 2016 KMTNet prime fields), we conduct a systematic search of the 2016 KMTNet subprime fields using a semi-machine-based algorithm to identify hidden anomalous events missed by the conventional by-eye search. We find four new planets and seven planet candidates that were buried in the KMTNet archive. The new planets are OGLE-2016-BLG-1598Lb, OGLE-2016-BLG-1800Lb, MOA-2016-BLG-526Lb, and KMT-2016-BLG- 2321Lb, which show typical properties of microlensing planets, i.e., giant planets orbit M-dwarf host stars beyond their snow lines. For the planet candidates, we find planet/binary or 2L1S/1L2S degeneracies, which are an obstacle to firmly claiming planet detections. By combining the results of Shin et al. (2023b) and this work, we find a total of The Astronomical Journal, 167:269 (30pp), 2024 June https://doi.org/10.3847/1538-3881/ad3ba3 © 2024. The Author(s). Published by the American Astronomical Society. Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence. Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. 1 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-6000-3463 https://orcid.org/0000-0001-6000-3463 https://orcid.org/0000-0001-6000-3463 https://orcid.org/0000-0002-2641-9964 https://orcid.org/0000-0002-2641-9964 https://orcid.org/0000-0002-2641-9964 https://orcid.org/0000-0003-0626-8465 https://orcid.org/0000-0003-0626-8465 https://orcid.org/0000-0003-0626-8465 https://orcid.org/0000-0003-0043-3925 https://orcid.org/0000-0003-0043-3925 https://orcid.org/0000-0003-0043-3925 https://orcid.org/0000-0001-5207-5619 https://orcid.org/0000-0001-5207-5619 https://orcid.org/0000-0001-5207-5619 https://orcid.org/0000-0002-4035-5012 https://orcid.org/0000-0002-4035-5012 https://orcid.org/0000-0002-4035-5012 https://orcid.org/0000-0003-3316-4012 https://orcid.org/0000-0003-3316-4012 https://orcid.org/0000-0003-3316-4012 https://orcid.org/0000-0001-6285-4528 https://orcid.org/0000-0001-6285-4528 https://orcid.org/0000-0001-6285-4528 https://orcid.org/0000-0003-1525-5041 https://orcid.org/0000-0003-1525-5041 https://orcid.org/0000-0003-1525-5041 https://orcid.org/0000-0002-0548-8995 https://orcid.org/0000-0002-0548-8995 https://orcid.org/0000-0002-0548-8995 https://orcid.org/0000-0002-2335-1730 https://orcid.org/0000-0002-2335-1730 https://orcid.org/0000-0002-2335-1730 https://orcid.org/0000-0002-9245-6368 https://orcid.org/0000-0002-9245-6368 https://orcid.org/0000-0002-9245-6368 https://orcid.org/0000-0002-7777-0842 https://orcid.org/0000-0002-7777-0842 https://orcid.org/0000-0002-7777-0842 https://orcid.org/0000-0002-2339-5899 https://orcid.org/0000-0002-2339-5899 https://orcid.org/0000-0002-2339-5899 https://orcid.org/0000-0003-4084-880X https://orcid.org/0000-0003-4084-880X https://orcid.org/0000-0003-4084-880X https://orcid.org/0000-0002-9326-9329 https://orcid.org/0000-0002-9326-9329 https://orcid.org/0000-0002-9326-9329 https://orcid.org/0000-0002-6212-7221 https://orcid.org/0000-0002-6212-7221 https://orcid.org/0000-0002-6212-7221 https://orcid.org/0000-0001-6364-408X https://orcid.org/0000-0001-6364-408X https://orcid.org/0000-0001-6364-408X https://orcid.org/0000-0002-3051-274X https://orcid.org/0000-0002-3051-274X https://orcid.org/0000-0002-3051-274X https://orcid.org/0000-0002-1650-1518 https://orcid.org/0000-0002-1650-1518 https://orcid.org/0000-0002-1650-1518 https://orcid.org/0000-0001-8043-8413 https://orcid.org/0000-0001-8043-8413 https://orcid.org/0000-0001-8043-8413 https://orcid.org/0000-0002-4909-5763 https://orcid.org/0000-0002-4909-5763 https://orcid.org/0000-0002-4909-5763 https://orcid.org/0000-0003-4776-8618 https://orcid.org/0000-0003-4776-8618 https://orcid.org/0000-0003-4776-8618 https://orcid.org/0000-0002-8198-1968 https://orcid.org/0000-0002-8198-1968 https://orcid.org/0000-0002-8198-1968 https://orcid.org/0000-0003-2302-9562 https://orcid.org/0000-0003-2302-9562 https://orcid.org/0000-0003-2302-9562 https://orcid.org/0000-0002-9629-4810 https://orcid.org/0000-0002-9629-4810 https://orcid.org/0000-0002-9629-4810 https://orcid.org/0000-0001-9818-1513 https://orcid.org/0000-0001-9818-1513 https://orcid.org/0000-0001-9818-1513 https://orcid.org/0000-0003-1978-2092 https://orcid.org/0000-0003-1978-2092 https://orcid.org/0000-0003-1978-2092 https://orcid.org/0009-0005-3414-455X https://orcid.org/0009-0005-3414-455X https://orcid.org/0009-0005-3414-455X https://orcid.org/0000-0001-8472-2219 https://orcid.org/0000-0001-8472-2219 https://orcid.org/0000-0001-8472-2219 https://orcid.org/0000-0002-1228-4122 https://orcid.org/0000-0002-1228-4122 https://orcid.org/0000-0002-1228-4122 https://orcid.org/0000-0002-5843-9433 https://orcid.org/0000-0002-5843-9433 https://orcid.org/0000-0002-5843-9433 https://doi.org/10.3847/1538-3881/ad3ba3 https://crossmark.crossref.org/dialog/?doi=10.3847/1538-3881/ad3ba3&domain=pdf&date_stamp=2024-05-16 https://crossmark.crossref.org/dialog/?doi=10.3847/1538-3881/ad3ba3&domain=pdf&date_stamp=2024-05-16 http://creativecommons.org/licenses/by/4.0/ nine hidden planets, which is about half the number of planets discovered by eye in 2016. With this work, we have met the goal of the systematic search series for 2016, which is to build a complete microlensing planet sample. We also show that our systematic searches significantly contribute to completing the planet sample, especially for planet/ host mass ratios smaller than 10−3, which were incomplete in previous by-eye searches of the KMTNet archive. Unified Astronomy Thesaurus concepts: Gravitational microlensing exoplanet detection (2147) 1. Introduction Since 2016, the Korea Microlensing Telescope Network (KMTNet; Kim et al. 2016) has operated a microlensing survey to detect exoplanets using their near-continuous observations toward the Galactic bulge. As of 2023, the KMTNet has contributed to the discovery/characterization of more than 135 microlensing planets.26 Initially, the planetary events were identified by a traditional method, i.e., “by-eye” search. The human dependence of that method, which relies on the experience or insight of operators, is difficult to quantify, and there may exist missing or hidden planets. Thus, we conduct a series of works called “Systematic KMTNet Planetary Anomaly Search” to find hidden planets in the KMTNet data archive in order to build a complete microlensing planet sample. The complete sample can be used to construct well- defined samples of planets for statistical studies such as the planet frequency and mass-ratio distribution of planetary systems in our Galaxy. To systematically search anomalous events in the KMTNet data archive, we use a semi-machine-based algorithm called AnomalyFinder (AF; Zang et al. 2021, 2022) instead of the by-eye search. The AF search is separately conducted for each year and cadence. The nominal cadences of the KMTNet observations have two categories, which are high cadence (Γ= 2.0–4.0 hr−1 for prime fields) and low cadence (Γ= 0.2–1.0 hr−1 for subprime fields). The detailed information of the KMTNet fields is described in Kim et al. (2018). Based on the AF searches, we conducted detailed light-curve analyses for the identified anomalous events. The parts of this AF series have been published or submitted. Indeed, from the systematic search, we can find hidden planets that are missing from the by-eye search. Shin et al. (2023b) reported five planets, which were newly found in the 2016 prime fields. Ryu et al. (2024) report three new planets found in the 2017 prime fields.27 Gould et al. (2022), Wang et al. (2022), and Hwang et al. (2022) reported a total of 12 new planets, which were discovered in the 2018 prime fields. Jung et al. (2022) reported six new planets found in the 2018 subprime fields. Zang et al. (2021, 2022) and Hwang et al. (2022) reported a total of seven new planets discovered in the 2019 prime fields. Jung et al. (2023) reported five new planets found in the 2019 subprime fields. Lastly, Zang et al. (2023) present seven new planets having q< 10−4, which were identified by the AF in the KMTNet data archive observed from 2016 to 2019. Although our systematic search works are not complete, yet, we have found a total of 45 hidden planets in the KMTNet archive, which amounts to about 33% of the total microlensing planets discovered from 2016 to 2022. Following the work of Shin et al. (2023b), we conduct the AF search for 2016 subprime fields to find hidden planetary systems. The AF identifies a total of 113 anomalous events in the fields, including the recovery of all previously published planetary events identified by eye (i.e., eight events). Among them, we find that 83 events were caused by binary-lens systems (i.e., q> 0.06) from the preliminary light-curve analyses using the KMTNet pipeline data. For the remaining 22 events, we conduct detailed light-curve analyses using the rereduced data sets with the best quality (see Section 2). The detailed analyses reveal that 11 events do not have possible planetary solutions (i.e., q< 0.03; see Appendix). Finally, we find four new planetary events and seven planet candidates on the 2016 subprime fields. The new planets are OGLE-2016- BLG-1598Lb, OGLE-2016-BLG-1800Lb, MOA-2016-BLG- 526Lb, and KMT-2016-BLG-2321Lb. We present the detailed light-curve analyses for these planetary events in Section 3. In this section, we also present the analyses of the planet candidates to show the possibility of planet detection. In Sections 4 and 5, we present the analyses of color–magnitude diagrams (CMDs) and lens properties of each planetary system, respectively. Lastly, we summarize our findings in Section 6. 2. Observations Although the AF identified anomalous events based on the KMTNet data archive, these events may also have been independently observed or discovered by other microlensing surveys. Thus, we gather all available data sets for each event. In Table 1, we list anomalous events that have at least one solution with q< 0.06 from the preliminary analyses along with their observational information. Note that, following the standard convention, we designate them according to the survey that first announced the event. The KMTNet data sets were obtained from three identical 1.6 m telescopes equipped with 4 deg2 wide field cameras, which are located at three sites in the Southern Hemisphere, i.e., the Cerro Tololo Inter-American Observatory in Chile (KMTC), South African Astronomical Observatory in South Africa (KMTS), and Siding Spring Observatory in Australia (KMTA). Note that, in the figures, the two-digit number after the site acronym indicates the field number of the KMTNet survey. These sites cover well-separated time zones to achieve near-continuous observations. The KMTNet observations are initially reduced using their pySIS pipeline (Albrow et al. 2009), which adopts the difference image analysis (DIA) method (Tomaney & Crotts 1996; Alard & Lupton 1998). For KMTC and KMTS observations, the KMTNet survey regularly takes one V-band observation for every 10th and 20th I-band observations, respectively (Johnson-Cousins BVRI filter sys- tem). The pipeline data are available at the KMTNet alert system (Kim et al. 2018).28 Note that we manually rereduced the KMTNet data sets for each preliminary planet candidate listed in Table 1 using the 26 We count the discovered microlensing planets using the NASA Exoplanet Archive (https://exoplanetarchive.ipac.caltech.edu/) as of 2023 October. 27 Among the three planetary events, two events were newly identified by the AF. While one event was previously identified by eye, however, this event was not published due to technical complications. 28 https://kmtnet.kasi.re.kr/ulens/ 2 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. http://astrothesaurus.org/uat/2147 https://exoplanetarchive.ipac.caltech.edu/ https://exoplanetarchive.ipac.caltech.edu/ https://exoplanetarchive.ipac.caltech.edu/ http://kmtnet.kasi.re.kr/ulens/ updated pySIS package described in Yang et al. (2024). We conduct light-curve analyses based on these tender-loving-care (TLC) reductions, which have checked the anomalous data points with the best quality. The Optical Gravitational Lensing Experiment (OGLE; Udalski et al. 2015) data sets were obtained from a 1.6 m Warsaw telescope equipped with a 1.4 deg2 field camera, which is located at Las Campanas Observatory in Chile. For the OGLE observations, it mainly takes I-band observations and periodically takes V-band observations. The OGLE observa- tions are reduced by their own DIA pipeline (Wozniak 2000). The data are available on the OGLE Early Warning System (Udalski et al. 1994).29 The Microlensing Observations in Astrophysics (MOA; Bond et al. 2001; Sumi et al. 2003) data sets were obtained from a 1.8 m telescope located at Mt. John University Observatory in New Zealand. The observations were made in the MOA red band (hereafter, referred to as R band), which has wavelength ranges of 609–1109 nm and transmission ranges of 0.0–0.978 (i.e., a rough sum of the Cousins R and I bands). The MOA observations were reduced by their DIA pipeline (Bond et al. 2001), which are available on the MOA alert system.30 3. Light-curve Analysis 3.1. Basics of the Light-curve Analysis We conduct the light-curve analysis following the proce- dures described in Shin et al. (2023b), which describes the systematic KMTNet planetary anomaly search for 2016 prime- field events. To avoid redundant descriptions for analysis procedures, we do not present the details here. However, in Table 2, we present definitions of acronyms and model parameters to describe the analysis results in the following sections. For each event, we conduct the heuristic analysis described in Ryu et al. (2022) to predict s and/or q values (also, to guess its possible degeneracy).31 We note that we test the APRX effect if the event has a relatively long timescale, which is defined as larger than tE> 15.0 days. Once we detect the APRX effect, we also test the OBT effect and xallarap effect to confirm the robustness of the APRX detection. Because the OBT can affect the APRX measurement and its uncertainty, and the xallarap can mimic the APRX effect. Lastly, if we find a planetary solution(s) from bump-shaped anomalies on the light curve, we test the 2L1S/1L2S degeneracy (Gaudi 1998) to confirm planet detection. To quantitatively compare various models for each event based on the Δχ2, we rescale the errors of each data set based on the best-fit solution using the method described in Yee et al. (2012). The error rescaling process can make each data point contribute χ2∼ 1. 3.2. Planetary Events We find four events caused by planetary lens systems that satisfy our minimum criteria to claim planet detection. For clarity, we summarize our criteria to claim planet detection as follows: Table 1 Observations of 2016 Planets and Planet Candidates Event Location Obs. Info. KMTNet OGLE MOA R.A. (J2000) decl. (J2000) (ℓ, b) AI Γ (hr−1) 0696 1598 521 18h00m45 78 -  ¢ 29 10 31. 58 (+1°. 47, − 2°. 97) 1.16 1.0 0781 1800 581 17h59m36 77 -  ¢ 30 51 55. 30 (−0°. 12, − 3°. 59) 1.72 1.0 1611 1705 526 17h46m40 39 -  ¢ 34 23 09. 89 (−4°. 54, − 3°. 02) 1.50 0.4 2321 L L 17h36m13 90 -  ¢ 25 31 29. 57 (+1°. 79, + 3°. 58) 3.88 0.4 1243 L L 17h49m55 56 -  ¢ 22 11 57. 91 (+6°. 27, + 2°. 65) 1.98 0.4 1406 0336 092 18h16m09 86 -  ¢ 25 08 02. 29 (+6°. 68, − 4°. 02) 1.18 0.4 1449 0882 L 18h14m50 42 -  ¢ 27 40 06. 71 (+4°. 29, − 4°. 96) 0.52 0.4 1609 1704 L 17h47m32 79 -  ¢ 34 42 40. 10 (−4°. 73, − 3°. 34) 1.04 0.4 1630 1408 L 17h40m38 76 -  ¢ 35 55 51. 96 (−6°. 51, − 2°. 78) 2.33 0.4 2399 L L 17h42m13 00 -  ¢ 22 56 15. 14 (+4°. 72, + 3°. 79) 1.76 0.4 2473 L L 17h41m24 88 -  ¢ 32 21 32. 40 (−3°. 39, − 1°. 03) 4.93 1.0 0255 0620 183 18h11m21 81 -  ¢ 26 50 24. 11 (+4°. 66, − 3°. 88) 1.06 0.4 0913 L L 17h35m43 08 -  ¢ 29 12 14. 87 (−1°. 37, + 1°. 69) 3.02 1.0 1004 1432 L 17h34m01 67 -  ¢ 26 56 02. 98 (+0°. 34, + 3°. 23) 2.41 1.0 1222 L L 17h44m29 20 -  ¢ 24 43 38. 10 (+3°. 46, + 2°. 41) 2.89 1.0 1326 1844 L 17h59m12 49 -  ¢ 33 10 20. 78 (−2°. 17, − 4°. 66) 1.42 1.4 1425 L L 18h11m30 35 -  ¢ 26 38 02. 62 (+4°. 85, − 3°. 81) 1.03 0.4 1433 0982 L 18h09m56 99 -  ¢ 26 45 28. 01 (+4°. 58, − 3°. 57) 1.35 0.4 1461 1517 L 18h13m49 10 -  ¢ 28 24 48. 60 (+3°. 53, − 5°. 11) 0.43 0.4 2067 1258 L 17h44m15 45 -  ¢ 26 31 13. 73 (+1°. 90, + 1°. 52) 3.18 1.0 2256 L L 17h36m36 13 -  ¢ 25 59 10. 61 (+1°. 45, + 3°. 26) 2.98 1.0 2331 L L 17h43m39 02 -  ¢ 26 07 51. 31 (+2°. 16, + 1°. 84) 2.87 1.0 Note. The boldface indicates the “discovery” name of each event. The horizontal lines separate planetary events, planet candidates, and nonplanetary events (see Appendix). 29 https://ogle.astrouw.edu.pl/ogle4/ews/ews.html 30 https://www.massey.ac.nz/~iabond/moa/alerts/ 31 For the details of the heuristic analysis, the heuristic analysis was originally introduced in Hwang et al. (2022). The formalism was modified for better approximation, which is described in Gould et al. (2022). Ryu et al. (2022) presented a unified formalism of the analysis. 3 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. http://ogle.astrouw.edu.pl/ogle4/ews/ews.html https://www.massey.ac.nz/~iabond/moa/alerts/ (a) The mass ratio of the best-fit planetary solution must be smaller than 0.03 (i.e., q< 0.03). (b) Competing binary-lens solutions can be resolved by Δχ2> 10.0. (c) If the 2L1S/1L2S degeneracy exists, 1L2S can be resolved by Δχ2> 15.0. These are the criteria used in other works and for the construction of the statistical sample in Zang et al. (2024). They are somewhat arbitrary, but for the construction of a statistical sample, the most important thing is that they can be incorporated into a sensitivity analysis in a straightforward way. In addition, we provide sufficient information on each event to allow a different choice of criteria. We present the details of the light-curve analysis for each planetary event in the following sections. 3.2.1. OGLE-2016-BLG-1598 As shown in Figure 1, the light curve of OGLE-2016-BLG- 1598 (which we identified as KMT-2016-BLG-0696) exhibits a shallow-dip anomaly near the peak (i.e., HJD′= 7636.0∼ 7639.0), which shows clear residuals from the 1L1S model (i.e., cD =- 171.91L1S 2L1S 2 ). The anomaly can be explained by two 2L1S models caused by the inner/outer degeneracy. Although the degenerate models cannot be resolved (i.e.,Δχ2= 8.5), both cases indicate that the lens system is a planetary system (i.e., q< 0.03) as presented in Table 3. Thus, we conclude that OGLE-2016-BLG- 1598 was caused by a planetary lens system. Indeed, the heuristic analysis (tanom= 7637.5, τanom=− 0.1045, and uanom= 0.2436) predicts =- †s 0.886, =+ †s 1.129, and q∼ 6.4× 10−4. The predicted q is consistent with the empirical q= 6.44× 10−4 value. Also, the empirical = =†s s s 0.833inner outer is similar to the predicted - †s value. Because of the relatively long timescale (tE∼ 38 days) for all cases, we test the annual APRX effect. As shown in Figure 2, we find the χ2 improves by ∼20.7 when we consider the APRX effect, which mostly comes from the OGLE data. However, the improvement of the OGLE data is inconsistent with the KMTNet and MOA data. Moreover, there is no improvement in the case of the KMTC data although the KMTC data have similar coverage to the OGLE data. Thus, we separately conduct APRX modeling using KMTC and OGLE only. We find that the OGLE-only case favors too large APRX values (i.e., |πE|> 2.82), which are unreliable. In contrast, the KMTC-only case shows that the APRX values are consistent with a nondetection (i.e., (πE,E, πE,N)∼ (0.0, 0.0) within 1σ level). The inconsistency between OGLE and KMTNet data of both the χ2 improvements and the APRX measurements indicates that the APRX effect of this event is unreliable. We test again the APRX effect using rereduced OGLE data. Even though we use the best-quality data sets, we have the same results from the test. Hence, we conclude that the STD models should be the fiducial solutions for this event. 3.2.2. OGLE-2016-BLG-1800 In Figure 3, we present the light curve of OGLE-2016-BLG- 1800 (which we identified as KMT-2016-BLG-0781), which shows deviations (HJD′= 7651.0–7657.0) from the 1L1S model. The anomaly can be explained by the 2L1S models that fit better by Δχ2= 196.33 compared to the 1L1S fits. In Table 4, we present the model parameters of the 2L1S solutions. Indeed, the heuristic analysis predicts =- †s 0.918 Table 2 Definition of Acronyms and Model Parameters Acronym Definition nLmS Number of lenses (n) and sources (m), which are included for models STDa Static model without any consideration of acceleration for the lens, source, and observer APRX Model considering the annual microlens-parallax (APRX) effect (Gould 1992) OBT Model considering the effect of the orbital motion (OBT) of lens system Model parameter Definition t0 Time at the peak of the light curve u0 Impact parameter in units of θE tE Time during which the source travels the angular Einstein ring radius (θE) s Projected separation between lens components in units of θE q Mass ratio of the lens components defined as q ≡ Msecondary/Mprimary α Angle of the source trajectory with respect to the binary axis of lens components (e.g., 0 is toward the planet or π is toward the host if q < 1) ρ* Angular source radius (θ*) scaled by θE, i.e., ρ* ≡ θ*/θE t0,Sm (m = 1, 2) Times of closest approach to the lens by the first and second sources, respectively u0,Sm (m = 1, 2) Impact parameter between the lens and the first and second sources, respectively qflux Flux ( f ) ratio of the binary sources defined as qflux ≡ fS2/fS1 sn (n = 1, 2) Projected separations between lens components, i.e., sn indicates the separation between m1 and m(n+1) where n = 1, 2 qn (n = 1, 2) Mass ratios of the lens components defined as qn ≡ m(n+1)/m1 ψ Orientation angle of m3 measured from the m1 − m2 axis with the m1 origin πE,E East component of the microlens-parallax vector, πE ≡ (πE,N, πE,E), projected on the sky πE,N North component of the πE ds/dt Changes of s in time (year) caused by the orbital motion of the lens system dα/dt Changes of α in time (year) Note. a We conduct modeling using the static case as the standard (STD) model. 4 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. and =+ †s 1.090 from τanom=− 0.100, uanom= 0.172, which is similar to the value of = =- + †s s s 0.911 from the models. We find that the s± cases of the 2L1S solutions cannot be resolved (Δχ2= 0.92). However, the mass ratios of both solutions indicate that the lens system consists of a planet and a host star. Thus, we conclude that OGLE-2016-BLG-1800 was caused by the planetary lens system. Because the timescales of both cases are relatively long (i.e., tE∼ 20 days), we test the APRX effect for this event. However, we find the χ2 improvement is negligible, only Δχ2= 0.74. Thus, we treat the STD cases as the fiducial solutions for this event. Also, for both cases, the ρ* is not measured as expected from the non-caustic-crossing geometries (see geometries in Figure 3). 3.2.3. MOA-2016-BLG-526 As shown in Figure 4, in the light curve of MOA-2016- BLG-526 (which we identified as KMT-2016-BLG-1611), two KMTC points near the peak exhibit an anomaly from the 1L1S model.32 Based on the TLC reductions, we investigate these Figure 1. Light curve of OGLE-2016-BLG-1598 with 2L1S and 1L1S models. We also present caustic geometries of the 2L1S models. Table 3 The Parameters of Degenerate 2L1S Models for OGLE-2016-BLG-1598 Parameter Outer Inner χ2/Ndata 4393.288/4395 4401.751/4395 Δχ2 L (best fit) 8.463 t0 [HJD′] 7641.470 ± 0.037 7641.471 ± 0.035 u0 0.225 ± 0.007 0.219 ± 0.007 tE [days] 37.513 ± 0.810 37.833 ± 0.814 s 0.955 ± 0.057 0.727 ± 0.030 q (×10−4) 6.437 ± 3.738 12.211 ± 3.543 á ñqlog10 −3.003 ± 0.152 −2.943 ± 0.131 α [rad] 5.137 ± 0.019 5.145 ± 0.016 ρ*,limit <0.083 <0.059 Note. HJD′ = HJD−2,450,000.0. We note that the ρ* is not measured. Thus, we present 3σ upper limits on the ρ* values (i.e., ρ*,limit). 32 We note that MOA data did not cover the anomaly part although the MOA first announced this event. Also, the data have systematics that might be caused by the faintness of the source or bad weather conditions. Thus, we do not include the MOA data in the analysis. 5 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. points to check whether or not the anomaly is reliable. We find that the anomalous points have “normal” photometric proper- ties compared to the average photometric properties of this event. Quantitatively, the FWHM values are 3.12 and 3.22, respectively, which are better than the average FHWM value (i.e., 3.59). The background levels of the two points are 1190.7 and 999.4, respectively. These are also better than the average value (i.e., 1370.4; we removed points that were observed during full moon phases for this average). In addition, while the photometric properties of the two points are better than average, they are also not extreme. We also visually inspected the images to confirm that there is nothing unusual about them. Hence, we conclude that the anomalous points are robust. Thus, we conduct the 2L1S modeling to describe the anomaly. We find that the 2L1S models can perfectly explain the anomaly, which shows better fits by Δχ2∼ 84 compared to the 1L1S model. Moreover, although the coverage of the anomaly is sparse, as we will show below, we find that all nonplanetary Figure 2. APRX test of OGLE-2016-BLG-1598. The upper two panels show the cumulative Δχ2 plot between the APRX and STD models with the light curve. The lower three panels show APRX contours obtained using all data (left), OGLE only (middle), and KMTC only (right), respectively. 6 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. cases (i.e., cases having > -( )qlog 2.010 ) are disfavored by 6σ level. Hence, only planetary solutions can explain the “reliable anomaly” on the light curve. Because of the sparse coverage, we find that there exist several degenerate solutions as presented in Table 5. Indeed, we predict =- †s 0.954, =+ †s 1.048, and q∼ 2.9× 10−4 from the heuristic analysis (τanom=− 0.026, and uanom= 0.094). The - †s is consistent with the = =- -( ) ( )†s s A s C 0.954 and the s value of the s− (B) case. The + †s is also consistent with = =+ +( ) ( )†s s A s B 1.049. The predicted q is similar to empirical q values of the s− cases by a factor of ∼2. For the s− case, we find several degenerate solutions within Δχ2< 1.0. These solutions show three categories of geometries as shown in Figure 5. The A, B, and C families are produced by the different source trajectories, which travel over the inner, intermediate, and outer parts of the caustics, Figure 3. Light curve of OGLE-2016-BLG-1800 with both 2L1S solutions compared to the 1L1S models. Table 4 The Parameters of 2L1S Solutions for OGLE-2016-BLG-1800 Parameter s− s+ χ2/Ndata 6117.000/6145 6117.919/6145 Δχ2 L (best fit) 0.919 t0 [HJD′] 7657.025 ± 0.026 7656.985 ± 0.030 u0 0.128 ± 0.007 0.139 ± 0.009 tE [days] 20.415 ± 0.847 19.844 ± 0.878 s 0.686 ± 0.030 1.211 ± 0.066 q (×10−4) 57.528 ± 14.341 65.310 ± 23.391 á ñqlog10 −2.229 ± 0.104 −2.133 ± 0.134 α [rad] 5.403 ± 0.016 5.392 ± 0.016 ρ*,limit <0.057 <0.071 Note. HJD′ = HJD−2,450,000.0. We note that the ρ* is not measured for both 2L1S cases. We present 3σ upper limits on the ρ* values (i.e., ρ*,limit). 7 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. respectively. Indeed, this kind of degeneracy was introduced in the analysis of OGLE-2017-BLG-0173Lb (Hwang et al. 2018). Thus, we adopt the Δξ (≡u0cscα− [s− s−1]) parameter described in Hwang et al. (2018) to separate and extract each family (see dotted lines in the upper left panel of Figure 5). We present the best-fit solution for each family as a representative. For the s+ case, we find two solutions caused by the inner/ outer degeneracy, which cannot be distinguished (i.e., Δχ2= 0.2). In Figure 6, we present the light curves of s+ solutions with their geometries. For consistency, we also present xD – ( )qlog10 space to show the locations of each solution, which are clearly divided into two categories. Although there exist several degenerate solutions with Δχ2 1.0, all solutions have mass ratios less than 0.03. Thus, we conclude that this event was caused by a planetary lens system. Because of the relatively long timescale (tE∼ 20 days) for all solutions, we test the APRX effect for this event. However, we find a negligible χ2 improvement of Δχ2= 1.55 compared to the best fit of STD solution and no meaningful constraints on πE. Thus, we treat the STD models as our fiducial solutions. Note that, because of the sparse coverage, we cannot measure the ρ* for all STD cases even though some cases show caustic- crossing features. Figure 4. Light curve of MOA-2016-BLG-526 with 2L1S solutions compared to the 1L1S models. 8 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. Lastly, the s+ solutions exhibit a bump-like anomaly, which can yield a 2L1S/1L2S degeneracy. Thus we check whether or not the 1L2S model can explain the anomaly. We find that the 1L2S model cannot explain the KMTC point at HJD′= 7637.4903, which shows a shallow dip relative to the 1L1S fits. Also, we find that the 1L2S interpretation is fine- tuned to describe the KMTC point at HJD′= 7637.6021. That is, to fit this point, the 1L2S model has ~ -( )q 10flux 4 , which is nonphysical. Thus, we conclude that there is no 2L1S/1L2S degeneracy for this event despite the fact that, due to the lack of covered data points, Δχ2= 11.7 between the 2L1S and 1L2S models, which is smaller than our formal threshold. We consider MOA-2016-BLG-526 to be a clear planet detection because all classes of solutions are planetary, and the anomalous points are real. However, this planet also illustrates that the utility of a criterion requiring at least three data points contributes to the signal for planets that are part of statistical samples (e.g., as in Shvartzvald et al. 2016). In this case, the fact that there are multiple solutions is a product of there only being two points on the anomaly with the result that there is substantial uncertainty in, e.g., the mass ratio of the planet. As seen in Figure 4, a third observation at almost any point during the anomaly could have differentiated between the families of the solution. In conclusion, whether or not this particular planet should be included in a given statistical sample must be carefully evaluated based on the criteria for defining that sample, and, conversely, this event serves as a good edge case to consider when defining such criteria. Ultimately, whether or not this particular planet is included in a given statistical sample will depend on what criteria are chosen to define that sample (such as the minimum number of data points in an anomaly), but it is outside of the scope of this work to define exactly what those criteria should be. 3.2.4. KMT-2016-BLG-2321 As shown in Figure 7, the light curve of KMT-2016-BLG- 2321 exhibits an apparent anomaly at HJD′∼ 7621.5 that has a short duration (∼0.35 days). We find that the anomaly can be explained by 2L1S models with caustic-crossing geometries. In Table 6, we present the best-fit parameters of the 2L1S solutions. Indeed, we predict =- †s 0.912, and =+ †s 1.097 from the heuristic analysis (τanom=−0.0737, tanom= 0.1853). The predicted + †s value corresponds with the empirical value of = =†s s s 1.096outer inner . Although the 2L1S solutions caused by the inner/outer degeneracy cannot be distinguished (Δχ2= 0.66), the mass ratios of both solutions indicate that the event was caused by a planetary lens system, i.e., ~ -( )q 10 3 . Because of the long timescales (∼57 days) for both solutions, we test the APRX effect. However, we find no χ2 improvement (the STD best-fit solution shows better fits than the APRX model by Δχ2= 0.33). Even though we addition- ally include the OBT effect (i.e., APRX+OBT model), we find a negligible χ2 improvement of Δχ2= 1.28 and no meaningful constraints on πE. Thus, we conclude that the higher-order effects are not available for this event. We note that, despite caustic-crossing features, the ρ* measurements are uncertain because the data coverage is not optimal. Because of the caustic-crossing feature, we expect the 2L1S/ 1L2S degeneracy will not be an obstacle to claim planet detection. However, because the coverage is not optimal, we check the 1L2S model for confirmation. As expected, we find the 1L2S model is disfavored by Δχ2= 132.56, which cannot explain the caustic-crossing feature despite the nonoptimal coverage. 3.3. Planet Candidates We find seven planet candidates among the 11 events, which are analyzed using the TLC data sets. These events have the possibility to be caused by a planetary lens system. However, these candidates cannot satisfy all our criteria to firmly claim planet detection. For example, there exist competing binary- lens solutions that cannot be resolved, or there is the 2L1S/ 1L2S degeneracy to prevent claiming the planet detection. Although we cannot firmly claim planet detection, there still remains the possibility that these events might be caused by a planetary system unless we have clear evidence against this conclusion. Hence, we report these planet candidates with the details of the light-curve analyses for the record, in case there is an opportunity to conclusively reveal their nature in the future. 3.3.1. KMT-2016-BLG-1243 The light curve of KMT-2016-BLG-1243 exhibits a flat- topped deviation at the peak as shown in Figure 8. Such an anomaly may be caused solely by finite-source (FS) effects or by the combination of FS effects and the central caustic of a 2L1S model. In Table 7, we present the model parameters of these various competing models. When fitting for a 1L1S+FS model, we find χ2= 685.63. Fitting for 2L1S models, (i.e., s± cases), we find solutions that fit better than the 1L1S+FS model by Δχ2= 9.6. These solutions imply that the lens system consists of binary stars (see Table 5 The Parameters of Degenerate 2L1S Solutions for MOA-2016-BLG-526 Parameter s− (A) s− (B) s− (C) s+ (A) s+ (B) χ2/Ndata 2218.429/2220 2218.431/2220 2218.648/2220 2219.283/2220 2219.463/2220 Δχ2 L (best fit) 0.002 0.219 0.854 1.034 t0 [HJD′] 7638.076 ± 0.021 7638.075 ± 0.021 7638.078 ± 0.022 7638.056 ± 0.019 7638.059 ± 0.019 u0 0.086 ± 0.005 0.087 ± 0.005 0.090 ± 0.005 0.085 ± 0.005 0.086 ± 0.005 tE [days] 19.483 ± 0.771 19.446 ± 0.758 19.037 ± 0.774 19.558 ± 0.752 19.348 ± 0.754 s 0.944 ± 0.005 0.954 ± 0.003 0.965 ± 0.006 1.062 ± 0.013 1.036 ± 0.014 q (×10−4) 1.760 ± 0.467 1.184 ± 0.329 1.593 ± 0.524 0.605 ± 0.620 0.475 ± 0.751 á ñqlog10 −3.772 ± 0.119 −3.957 ± 0.128 −3.787 ± 0.138 −4.096 ± 0.229 −4.081 ± 0.239 α [rad] 5.109 ± 0.016 5.104 ± 0.016 5.103 ± 0.017 1.855 ± 0.018 1.855 ± 0.018 ρ*,limit <0.007 <0.009 <0.009 <0.007 <0.007 Note. HJD′ = HJD−2,450,000.0. We note that the ρ* is not measured for any 2L1S case. We present 3σ upper limits on the ρ* values (i.e., ρ*,limit). 9 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. Table 7). However, we also find that there exist competing 2L1S models (Δχ2< 4.8) that indicate that the lens is likely to be a planetary system (i.e., q< 0.03). Indeed, we predict =- †s 0.990, and =+ †s 1.011 from the heuristic analysis (τanom= 0.0067, uanom= 0.0211), which is similar to the º =- + †s s s 1.024 for the combination of P3 and P4 cases. In Figure 9, we present the residuals of the anomaly part for all degenerate models with their caustic geometries. By comparing them, we find the χ2 difference mostly comes from fits between HJD′= 7643.5∼ 7646.0. However, because of the sparse coverage, the Δχ2 of all degenerate cases is smaller than our χ2 criterion (Δχ2= 10.0) to claim a planet detection. In particular, the best-fit model of the planet case shows only Δχ2= 1.8. Lastly, we note that we test the APRX effect because of the long timescales (i.e., tE> 70 days). However, we find negli- gible χ2 improvement of 2.8 compared to the STD best-fit case. However, in addition to the χ2 criterion, we can also apply a Galactic prior to these solutions. In the case of the planetary models, the model parameters imply that μrel= θ*/(ρ*tE)∼ 0.08mas yr−1 by assuming a dwarf source (i.e., θ*∼ 0.5 μas). From the argument following Equation (22) of Gould (2022), the probability of such an exceptionally small μrel is m ~ ´- -( )6 mas yr 1.8 10rel 1 2 4. Likewise, the 1L1S+FS model is equally unlikely that it has similar parameters. Hence, although this event is technically a planet candidate based on Δχ2, those solutions are extremely unlikely to be the true solution after taking physical considerations into account. Figure 5. Light curve of MOA-2016-BLG-526 with the family of degenerate 2L1S s− models with the xD – ( )qlog10 space. In the xD – ( )qlog10 space (upper left panel), each color represents Δχ2 � n2 from the best-fit χ2 where n = 1 (red), 2 (yellow), 3 (green), 4 (light blue), 5 (blue), and 6 (purple), respectively. 10 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. 3.3.2. OGLE-2016-BLG-0336 As shown in Figure 10, the light curve of OGLE-2016-BLG- 0336 (which we identified as KMT-2016-BLG-1406) shows an apparent bump-shaped anomaly at the peak (HJD′= 7481.7), which was covered by KMTC and KMTS observations. We find that the anomaly can be explained by several models presented in Table 8. Similar to the case of MOA-2016-BLG- 526, there exist three 2L1S solutions caused by different caustic geometries (i.e., (A) caustic-crossing, (B) inner, and (C) outer trajectories). These cases cannot be resolved (i.e., Δχ2 1). Indeed, we predict =- †s 0.919, and =+ †s 1.088 from the heuristic analysis (τanom= 0.0280, uanom= 0.1684). The + †s is well consistent with the best fit of s+= 1.089. We present the xD – ( )qlog10 space to show the locations of these degenerate cases (see the right upper panel in Figure 10). Although we cannot resolve the degeneracy, the mass ratios of all 2L1S solutions imply that the lens is likely to be a planetary lens system (i.e., q< 0.03). However, the bump-shaped anomaly is a typical type to have the 2L1S/1L2S degeneracy. We find that the 1L2S model can describe the anomaly well. Moreover, the Δχ2 compared to the 2L1S best-fit model is only 1.13. Because there are only weak constraints on ρ*,S1 and ρ*,S2, and a relatively large separation between the two sources (Δu∼ 0.17), we cannot place any additional meaningful constraints from physical considerations. Based on currently available data sets and analysis results, we cannot resolve the 2L1S/1L2S degeneracy for this event. Thus, we treat this event as a planet candidate unless we have additional evidence to rule out the 1L2S solution. Note that we have checked the APRX effect for this event because of the relatively long timescale (tE∼ 25 days). We find the χ2 improvement of 14.83 for the APRX-included model. Figure 6. Light curve of MOA-2016-BLG-526 with the degenerate 2L1S s+ models and the xD – ( )qlog10 space. The color scheme of the xD – ( )qlog10 space is identical to that of Figure 5. 11 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. However, we find that the χ2 improvements between data sets are inconsistent. Indeed, the STD model shows better fits for the KMTC data that yields Δχ2∼ 10.0. In contrast, for the other data (OGLE, MOA, and KMTA), the APRX model shows better fits that yield Δχ2∼ 8.0, 12.0, and 4.0, respectively. For KMTS, there is no χ2 improvement. This inconsistency makes us suspect the APRX detection is unreliable, similar to the case of OGLE-2016-BLG-1598. Also, these improvements only come from the baseline, which can have systematics. Thus, we conclude that the APRX measurement is not robust. The STD models should be the fiducial solutions for this event. 3.3.3. OGLE-2016-BLG-0882 The light curve of OGLE-2016-BLG-0882 (which we identified as KMT-2016-BLG-1449) shows anomalies at the peak, which have complex features consisting of three bump- shaped anomalies as shown in Figure 11. We find no 2L1S models that can correctly describe the anomalies. Thus, we try to describe the anomalies using 2L2S and 3L1S interpretations. We find the best-fit 2L2S model can describe all anomalies, which implies that the lens system consists of binary stars (i.e., q∼ 0.3). However, we also find that there exist competing solutions having Δχ2< 10.0. In Table 9, we present these degenerate 2L2S solutions. Among them, one case satisfies our Figure 7. Light curve of KMT-2016-BLG-2321 with the 2L1S solutions compared to the 1L1S models. 12 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. mass ratio criterion for planet detection (i.e., q∼ 0.01< 0.03). The Δχ2 between binary and planetary solutions is only 5.8, which is not enough to distinguish them. In addition, because the complex anomaly could be described by the 3L1S interpretation, we try to find a possible planetary solution. We find a plausible 3L1S model that can describe the anomalies (see Figure 11 and Table 9). This 3L1S model implies that the third body is likely to be a planet (i.e., q2∼ 0.011). However, this 3L1S model shows Δχ2> 32.0 compared to the best-fit 2L2S models. If we consider the satisfied 2L2S models, the 3L1S has worse fits by Δχ2> 23.0. Thus, the 3L1S case can be nominally ruled out considering our χ2 criterion. However, we do not ignore the possibility of the 3L1S solution because of two reasons. First, our search for 3L1S models was not exhaustive because of the technical difficulty of conducting a full search of the six parameters required to describe the two companions and source trajectory angle (i.e., s1, q1, s2, q2, ψ, and α), which are most sensitive to explaining the anomaly. Thus, there may exist alternative 3L1S solution(s) having better χ2. Second, for this event, the data sets have systematics on the anomaly part that are not explained by any model. Hence, our χ2 criteria may not be sufficient in this case. Thus, we present the 3L1S planetary solution as one alternative possibility of planetary systems that could produce the anomaly. Indeed, if we rule out this 3L1S case, there still remains the binary/planet degeneracy in the 2L2S solutions. Thus, we treat this event as a planet candidate including the possible 3L1S solution. Lastly, we note that we test the APRX effect because the models show that the timescales are longer than 32 days. However, we find only negligible χ2 improvement (i.e., Δχ2∼ 4.7) when the APRX effect is considered. Thus, we conclude the STD models are fiducial solutions for this event. 3.3.4. OGLE-2016-BLG-1704 The light curve of OGLE-2016-BLG-1704 (which we identified as KMT-2016-BLG-1609) shows apparent deviations from the 1L1S fit. The anomaly can be explained by various models. In Figure 12, we present these models with their caustic geometries. As shown in Table 10, the best-fit model (see the (A) case) implies that the lens system consists of binary stars (i.e., q∼ 0.53). However, there exist degenerate models having Δχ2< 10.0. The mass ratio of the (B) case nominally indicates that the lens is likely to be a binary star system. However, this model is caused by the Chang & Refsdal lensing (Chang & Refsdal 1979), which has large uncertainties in the (s, q) parameters. Hence, the mass ratio satisfies our mass ratio criterion (i.e., q< 0.03) within 1σ. For the (C) case, the mass ratio indicates the lens system could have a planet. The (D) solution can be nominally resolved by Δχ2= 13.1, which is slightly larger than our χ2 criterion. However, by considering the systematics in the data sets, we cannot firmly rule out this case. Thus, we present this planet-like case for completeness. For the (C) and (D) cases, the heuristic analysis (τanom= 0.0217, uanom= 0.0752) predicts =- †s 0.963, and =+ †s 1.038, which is similar to the empirical value of = =- + † ( ) ( )s s s 1.035, C , D . Lastly, we find that a 1L2S model can also explain the anomaly. The Δχ2 between the best-fit and 1L2S models is only 3.4, which cannot be resolved. Thus, we treat this event as a planet candidate because of the binary/planet and 2L1S/ 1L2S degeneracies. We note that we have tested the APRX effect because of the relatively long timescales (i.e., tE> 32 days). We find the negligible χ2 improvement of 5.0 when the APRX effect is included. Thus, we conclude that the STD models are the fiducial solutions for this event. 3.3.5. OGLE-2016-BLG-1408 OGLE-2016-BLG-1408 (which we identified as KMT-2016- BLG-1630) is a long timescale event that has an anomaly at the peak on the light curve. In Figure 13, we present the light curve with the 2L1S and 1L1S models of the STD and APRX cases. Because of the long timescale (i.e., tE> 96 days), we find that the APRX effect is essential to describe the observed light curve. In particular, as shown in Figure 13, it is impossible to describe the 2017 data without the APRX effect. Also, the 2L1S models with the APRX effect are the only interpretations that can explain the anomaly at the peak. However, we find that several 2L1S APRX models can describe the whole light curve, which cannot be distinguished from each other. In Figure 14, we present these degenerate solutions with their caustic geometries. We also present model parameters for the cases in Table 11. The best-fit case indicates that the lens could be a planetary system (i.e., ~ -( )q 10 3 ). There exist five competing planetary cases caused by the close/ wide (Griest & Safizadeh 1998) and ecliptic (Smith et al. 2003; Jiang et al. 2004; Poindexter et al. 2005) degeneracies. Although, among the planetary cases, the wide u0± cases can be nominally resolved by Δχ2> 10.0, we present them for completeness and comparison to the binary-lens cases. Despite the best-fit model implying that the lens has a planet, we find that there also exist competing binary-lens cases having Δχ2 5.4. In particular, the best fit of the binary case shows only Δχ2= 0.9. We note that we conduct tests for the APRX effect because the effect is essential to finding the solutions. First, we have tested the OBT effect, which can affect the APRX measure- ment. We find no χ2 improvement when the OBT effect is considered (i.e., Δχ2 [OBT−APRX]= 0.3). Moreover, we find that the OBT effect does not affect the uncertainties of the APRX measurement. Second, we have tested whether the xallarap effect can mimic the APRX effect. Similar to the OBT case, we find that the xallarap effect does not improve the fits (i.e., Δχ2 [xallarap−APRX]= 0.4). Also, as shown in Figure 15, the best-fit xallarap model has P= 1 yr, which is consistent with the orbital period of the Earth. Both facts imply Table 6 The Parameters of Degenerate 2L1S Solutions for KMT-2016-BLG-2321 Parameter Outer Inner χ2/Ndata 882.415/883 883.079/883 Δχ2 L (best fit) 0.664 t0 [HJD′] 7625.748 ± 0.210 7625.684 ± 0.216 u0 0.166 ± 0.018 0.168 ± 0.020 tE [days] 56.847 ± 4.562 56.966 ± 6.387 s 1.039 ± 0.014 1.157 ± 0.012 q (×10−4) 12.326 ± 2.992 12.703 ± 3.200 á ñqlog10 −2.967 ± 0.115 −2.948 ± 0.118 α [rad] 1.993 ± 0.020 1.981 ± 0.021 ρ*,limit (×10−3) <1.466 <1.268 Note. HJD′ = HJD−2,450,000.0. We note that ρ* is not measured for any case. We present 3σ upper limits on the ρ* values (i.e., ρ*,limit). 13 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. that the effect on the light curve is caused by APRX rather than xallarap. Hence, we conclude that the APRX models are the fiducial solutions for this event. We also note that we can measure ρ* for only the resonant (u0± ) cases induced by the caustic-crossing geometries. For other cases, we cannot robustly measure the ρ* because of the non-caustic-crossing geometries. Lastly, we check the 2L1S/1L2S degeneracy because the bump-like anomaly can be explained by the 1L2S interpreta- tion. We find that the 1L2S model with the APRX effect shows better fits by Δχ2= 6.7 compared to the best fit of the 2L1S APRX models. Thus, this 1L2S model can be an alternative solution for this event. At this moment, we cannot resolve both planet/binary and 2L1S/1L2S degeneracies for this event because of insufficient Δχ2. Hence, we treat OGLE-2016- BLG-1408 as a planet candidate. 3.3.6. KMT-2016-BLG-2399 The light curve of KMT-2016-BLG-2399 shows a bump- shaped anomaly on the rising part (HJD′∼ 7626). As shown in Figure 16, the anomaly can be described by a binary-lens model that contains a low-mass object (i.e., q∼ 0.057). We also find that planet-like models can plausibly describe the anomaly. In Table 12, we present the model parameters of possible solutions for this event. Indeed, the heuristic analysis (τanom=− 0.2813, uanom= 0.2924) predicts =- †s 0.864, and =+ †s 1.157, which is consistent with º =+ -( ) ( )†s s A s C 0.864. In addition, we Figure 8. Light curve of KMT-2016-BLG-1243 with 2L1S binary and planet models compared to the 1L1S model. The 1L1S model includes the finite- source effect (FS). 14 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. Table 7 The Parameters of Degenerate Solutions for KMT-2016-BLG-1243 Case 2L1S: Binary 2L1S: Planet 1L1S Parameter B1 B2 P1 P2 P3 P4 Parameter FS χ2/Ndata 676.060/677 676.254/677 677.904/677 680.261/677 680.065/677 680.823/677 χ2/Ndata 685.628/677 Δχ2 L (best fit) 0.194 1.844 4.201 4.005 4.763 Δχ2 9.568 t0 [HJD′] 7644.621 ± 0.022 7644.596 ± 0.020 7644.588 ± 0.019 7644.587 ± 0.019 7644.583 ± 0.019 7644.598 ± 0.019 t0 [HJD′] 7644.599 ± 0.018 u0 0.018 ± 0.001 0.015 ± 0.001 0.023 ± 0.002 0.022 ± 0.002 0.024 ± 0.002 0.024 ± 0.002 u0 0.023 ± 0.002 tE [days] 74.153 ± 4.402 85.562 ± 5.320 72.871 ± 4.421 77.779 ± 4.802 70.841 ± 4.591 71.117 ± 4.656 tE [days] 74.794 ± 4.593 s 0.162 ± 0.023 8.260 ± 1.095 - +1.167 1.002 0.584 - +0.407 0.248 2.387 - +2.945 2.812 2.006 - +0.356 0.223 4.462 L L q 0.200 ± 0.138 0.336 ± 0.104 ´- + -( )0.759 100.759 45.951 4 - +0.008 0.008 0.004 - +0.027 0.027 0.024 - +0.025 0.025 0.022 L L á ñqlog10 −0.632 ± 0.184 −0.559 ± 0.157 −3.359 ± 0.955 −3.358 ± 1.013 −3.311 ± 1.014 −3.333 ± 0.983 L L α [rad] 4.043 ± 0.057 4.016 ± 0.053 −6.219 ± 3.490 6.058 ± 3.137 3.811 ± 3.638 −2.417 ± 3.479 L L ρ* <0.016 <0.014 0.031 ± 0.002 0.029 ± 0.002 0.032 ± 0.002 0.032 ± 0.002 ρ* 0.030 ± 0.002 Note. HJD′ = HJD−2,450,000.0. For the ρ* parameter, the inequality sign indicates the upper limit on ρ* (i.e., 3σ), because we cannot robustly measure ρ* for those cases. 15 T h e A stro n o m ica l Jo u rn a l, 167:269 (30pp), 2024 June S hin et al. find that the bump-shaped anomaly can also be plausibly described by a 1L2S model, which showsΔχ2= 14.2 compared to the best-fit model. We note that the planet-like cases are borderline given our criteria. First, for the B case, the mass ratio is ∼0.030, which is consistent with the q criterion, while the C case does not satisfy the q criterion. However, the C model shows a very short timescale (i.e., tE∼ 8 days) with a relatively small q value (i.e., q∼ 0.049), which implies the component of the lens system would be a planet. Second, both cases are nominally resolved by the χ2 criterion (i.e., Δχ2= 10.0). However, the B case (Δχ2= 10.2) is very close to our χ2 criterion. By considering the systematics in the data, we cannot firmly rule out the model based on current data. We note that the B model exhibits a sharp bump at HJD′∼ 7620. However, there are no available data points observed by either KMTNet or OGLE. Even if we can rule out the C and 1L2S cases by simply adopting our criteria, there still remains a possible planet case (i.e., the B case) that cannot be clearly ruled out. Thus, we treat this event as a planet candidate. Note that we have tested the APRX effect for this event because the best-fit solution has a sufficiently long timescale (i.e., tE∼ 19 days) that the APRX effect may be detected. However, we find a negligible χ2 improvement of 0.9 when we consider the APRX effect. Thus, the STD cases are the fiducial models for this event. Finally, we note that we can measure the ρ* values for the A (caustic-crossing) and C (buried caustic) cases (see caustic geometries in Figure 16). 3.3.7. KMT-2016-BLG-2473 The light curve of KMT-2016-BLG-2473 exhibits anomalies from the 1L1S model (Δχ2= 171.0) during HJD′= 7500∼ 7520, as shown in Figure 17. The anomalies can be explained by a 2L1S model (note that the heuristic analysis is not valid for this event). The mass ratio of this best-fit model indicates that the lens system is likely to be a planetary system (i.e., q∼ 0.011). However, we find that a 1L2S model is also able to plausibly describe the anomaly. In Table 13, we present the model parameters of the 2L1S and 1L2S models. The 2L1S and 1L2S models themselves show a clear difference at HJD′∼ 7505.0, which seems to be a shallow bump-shaped anomaly. However, the Δχ2 between them is only 10.3, which does not satisfy our criterion to resolve the 2L1S/1L2S degeneracy. The small Δχ2 is caused by severe systematics in data sets because the event experienced heavy extinction (i.e., AI∼ 4.9). Thus, we treat this event as a planet candidate because we do not have any conclusive evidence to resolve the 2L1S/1L2S degeneracy. Note that we have tested the APRX effect because of the long timescale (i.e., tE∼ 47 days). We find a small χ2 Figure 9. OGLE-2016-BLG-1258: Residuals of each case shown in Table 7 with its caustic geometry. We show the residuals for the zoom-in part of Figure 8. 16 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. improvement of 5.8 when we consider the APRX effect. However, the improvement comes from the baseline, which has severe systematics. Thus, we conclude that the APRX effect is not robust. Hence, the STD models are the fiducial solutions for this event. Lastly, we note that the ρ* can be measured for the 2L1S case from the caustic-crossing feature. Figure 10. Light curves of OGLE-2016-BLG-0336 with degenerate models and their residuals. We present xD – ( )qlog10 space to show the local minima for the 2L1S models. We also present their caustic geometries. Table 8 The Parameters of Degenerate Solutions for OGLE-2016-BLG-0336 Parameter A: 2L1S B: 2L1S C: 2L1S Parameter 1L2S χ2/Ndata 7847.419/7875 7848.177/7875 7848.525/7875 χ2/Ndata 7848.545/7875 Δχ2 L (best fit) 0.758 1.106 Δχ2 1.126 t0 [HJD′] 7480.996 ± 0.009 7481.006 ± 0.009 7481.009 ± 0.009 t0,S1 [HJD′] 7480.995 ± 0.010 u0 0.166 ± 0.002 0.164 ± 0.002 0.164 ± 0.002 u0,S1 0.168 ± 0.006 tE [days] 24.513 ± 0.227 24.621 ± 0.230 24.623 ± 0.230 tE [days] 24.495 ± 0.246 s 1.089 ± 0.003 1.007 ± 0.016 1.174 ± 0.019 t0,S2 [HJD′] 7481.700 ± 0.012 q (×10−4) 0.229 ± 0.032 1.462 ± 0.319 1.577 ± 0.340 u0,S2 −0.004 ± 0.003 á ñqlog10 −4.619 ± 0.059 −3.835 ± 0.110 −3.802 ± 0.115 qflux 0.002 ± 0.001 α [rad] 1.396 ± 0.003 1.399 ± 0.004 1.400 ± 0.004 ρ*,S1 <0.190 ρ* 0.010 ± 0.001 <0.011 <0.011 ρ*,S2 <0.009 Note. HJD′ = HJD−2,450,000.0. For the ρ* parameter, the inequality sign indicates the upper limit on ρ* (i.e., 3σ), because we cannot robustly measure ρ* for those cases. 17 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. Figure 11. Light curves of OGLE-2016-BLG-0882 with degenerate models and their residuals. We also present the caustic geometries of each case. Table 9 The Parameters of Degenerate Solutions for OGLE-2016-BLG-0882 Parameter A: 2L2S B: 2L2S C: 2L2S Parameter 3L1S χ2/Ndata 1431.945/1436 1439.626/1436 1437.736/1436 χ2/Ndata 1471.948/1436 Δχ2 L (best fit) 7.681 5.791 Δχ2 40.003 t0,S1 [HJD′] 7520.424 ± 0.152 7521.433 ± 0.284 7521.529 ± 0.035 t0 7524.747 ± 0.133 u0,S1 0.086 ± 0.007 0.064 ± 0.006 0.350 ± 0.021 u0 0.095 ± 0.007 tE [days] 38.718 ± 2.156 55.660 ± 4.965 32.083 ± 0.923 tE 54.345 ± 2.877 s 0.543 ± 0.012 2.651 ± 0.269 1.186 ± 0.012 s1 1.523 ± 0.033 q 0.286 ± 0.017 0.568 ± 0.148 0.010 ± 0.001 q1 0.074 ± 0.005 á ñ( )qlog10 −0.530 ± 0.024 −0.274 ± 0.111 −1.976 ± 0.042 á ñ( )qlog10 1 −1.156 ± 0.029 α 5.223 ± 0.018 5.430 ± 0.067 1.561 ± 0.007 α 4.740 ± 0.019 t0,S2 [HJD′] 7524.833 ± 0.220 7526.798 ± 0.473 7530.741 ± 0.225 s2 1.004 ± 0.002 u0,S2 0.178 ± 0.014 0.122 ± 0.020 0.091 ± 0.012 q2 0.011 ± 0.002 qflux 6.476 ± 1.186 5.081 ± 3.829 0.252 ± 0.031 á ñ( )qlog10 2 −1.972 ± 0.062 ρ*,S1 <0.004 <0.004 0.003 ± 0.001 ψ 4.589 ± 0.029 ρ*,S2 <0.047 <0.039 0.023 ± 0.015 ρ* (×10−4) 11.194 ± 2.436 Note. HJD′ = HJD−2,450,000.0. For the ρ*,S1 and ρ*,S2 parameter, the cases with inequality signs are upper limits (i.e., 3σ), because we cannot robustly measure the source sizes. 18 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. 4. CMD Analysis We cannot securely measure ρ* for any of the four planetary events. We can determine only upper limits on the ρ* values. However, we can apply the ρ* distributions as constraints on the Bayesian analysis by including information on the angular source radius (θ*) of each event in the analysis. Thus, we carry Figure 12. Light curves of OGLE-2016-BLG-1704 with degenerate models and their residuals. We also present the caustic geometries of each 2L1S case. Table 10 The Parameters of Degenerate Solutions for OGLE-2016-BLG-1704 Case 2L1S: Binary 2L1S: Planet-like 1L2S Parameter (A) s− (B) s+ (C) s− (D) s+ Parameter (E) χ2/Ndata 1791.049/1796 1791.171/1796 1793.005/1796 1804.127/1796 χ2/Ndata 1794.442/1796 Δχ2 L (best fit) 0.122 1.956 13.078 Δχ2 3.393 t0 [HJD′] 7638.124 ± 0.198 7638.611 ± 0.086 7638.207 ± 0.107 7638.209 ± 0.090 t0,S1 [HJD′] 7637.907 ± 0.193 u0 0.125 ± 0.023 0.028 ± 0.010 0.155 ± 0.023 0.072 ± 0.011 u0,S1 0.238 ± 0.069 tE [days] 39.040 ± 4.575 40.442 ± 6.261 32.127 ± 3.267 50.567 ± 7.298 tE [days] 33.855 ± 4.459 s 0.363 ± 0.054 5.792 ± 0.443 0.723 ± 0.066 1.481 ± 0.051 t0,S2 [HJD′] 7639.517 ± 0.061 q 0.530 ± 0.240 0.076 ± 0.067 0.025 ± 0.016 0.020 ± 0.005 u0,S2 −0.006 ± 0.025 á ñqlog10 −0.352 ± 0.202 −1.121 ± 0.283 −1.474 ± 0.177 −1.702 ± 0.112 qflux 0.059 ± 0.033 α [rad] −0.408 ± 0.096 −0.303 ± 0.032 1.291 ± 0.031 5.185 ± 0.034 ρ*,S1 <0.525 ρ* <0.031 <0.017 <0.054 <0.021 ρ*,S2 <0.059 Note. HJD′ = HJD−2,450,000.0. We note that ρ* is not measured for all cases. Thus, we present 3σ upper limits on the ρ* values (i.e., ρ*,limit). 19 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. out the CMD analysis to measure the θ*. The basics of the CMD analysis are described in Yoo et al. (2004). In addition, the detailed procedures of the analysis are described in Shin et al. (2023b). In Figure 18, we present the measured locations of the centroid of the red giant clump (RGC), the source, and the blend overlaid on the CMD of each event. Although the analysis is conducted based on the multiband KMTNet observations (i.e., I and V bands), we present them in the OGLE-III magnitude system because we determine the RGC based on the OGLE-III CMD (Szymański et al. 2011). The exception is KMT-2016-BLG-2321 because the OGLE-III CMD is not available for this event, so we present the uncalibrated/dereddend KMTNet magnitudes instead. In Table 14, we present the results of the CMD analyses with the derived θ* values. We also present the lower limits on the angular Einstein ring radii (θE) and lens-source relative proper motions (μrel). Indeed, the lower limit on μrel (i.e., μrel,+3σ≡ θ*/tEρ*,+3σ) is a useful indicator to check the effect of the ρ* constraint before proceeding with the actual Bayesian Figure 13. Light curves of OGLE-2016-BLG-1408 with 2L1S and 1L1S models. For the 2L1S and 1L1S models, we present both the STD and APRX cases. 20 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. analysis. In general, we expect 1< μrel/mas yr−1< 10. Hence, if the lower limit on μrel,+3σ 1 mas yr−1, we expect the ρ* constraint to have little effect on the Bayesian result. Note that, for KMT-2016-BLG-2321, we conduct an additional analysis to check our measurement of the source color because the quality of the V-band data is low. The V-band light curve has systematics because this event experienced severe extinction (i.e., AI∼ 3.88), and the source is faint (i.e., IKMTNet∼ 21.4). Thus, we have checked our measurement using the source color estimation method (Bennett et al. 2008) and the Galactic bulge CMD (Holtzman et al. 1998) from the Hubble Space Telescope. We find that the estimated source color (i.e., (V− I)0,S=0.723± 0.055) is consistent with our measured color (i.e., (V− I)0,S = 0.763± 0.092, or (V− I)0,S=0.752± 0.090) at the 1σ level. Hence, we conclude that our measurement is reliable despite the obstacles. 5. Planet Properties The lens properties such as the mass of the lens system (ML), distance to the lens (DL), projected separation between lens components (a⊥), and lens-source relative proper motion (μrel) can be determined from q m = = = = p p q k q p q + ^ ( ) ∣ ∣ ∣ ∣ M D a sD , , , , 1 t L L au L E rel E E E E S E E where k = - M8.144 mas 1, and πS is the parallax of the source defined as πS≡ au/DS (DS is distance to the source). As shown in Equation (1), two observables (i.e., θE and |πE|) need to be measured to directly determine the lens properties. These observables may be measured from the finite-source and microlens-parallax effects, respectively. However, for the planetary events in this work, we do not have measurements Figure 14. Comparison of 2L1S APRX models with their caustic geometries for OGLE-2016-BLG-1408. 21 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. Table 11 The Parameters of 2L1S APRX Models for OGLE-2016-BLG-1408 Case Planet Binary Parameter Resonant (u0 − ) Close (u0 − ) Wide (u0 − ) Close (u0 − ) Wide (u0 − ) χ2/Ndata 2109.666/2124 2119.292/2124 2121.676/2124 2110.571/2124 2111.403/2124 Δχ2 L (best fit) 9.626 12.010 0.905 1.737 t0 [HJD′] 7653.474 ± 0.048 7653.458 ± 0.051 7653.432 ± 0.055 7653.763 ± 0.103 7653.127 ± 0.070 u0 −0.089 ± 0.002 −0.089 ± 0.002 −0.090 ± 0.003 −0.090 ± 0.003 −0.073 ± 0.005 tE [days] 96.686 ± 2.217 96.302 ± 2.287 95.185 ± 2.331 95.874 ± 2.872 117.147 ± 8.281 s 1.083 ± 0.005 0.701 ± 0.094 1.742 ± 0.235 0.263 ± 0.018 5.400 ± 0.586 q (×10−4) 18.382 ± 2.461 47.894 ± 40.051 65.010 ± 37.037 L L q L L L 0.275 ± 0.071 0.571 ± 0.199 á ñqlog10 −2.752 ± 0.061 −2.186 ± 0.216 −2.187 ± 0.171 −0.525 ± 0.093 −0.229 ± 0.136 α [rad] 4.200 ± 0.018 4.176 ± 0.020 4.179 ± 0.019 5.522 ± 0.043 2.453 ± 0.033 ρ* 0.040 ± 0.003 <0.047 <0.046 <0.044 <0.034 πE,N 0.196 ± 0.059 0.203 ± 0.063 0.228 ± 0.056 0.183 ± 0.060 0.180 ± 0.058 πE,E −0.247 ± 0.012 −0.244 ± 0.014 −0.242 ± 0.013 −0.241 ± 0.014 −0.198 ± 0.012 Resonant (u0 + ) Close (u0 + ) Wide (u0 + ) Close (u0 + ) Wide (u0 + ) χ2/Ndata 2110.464/2124 2117.795/2124 2119.989/2124 2114.903/2124 2115.099/2124 Δχ2 0.798 8.129 10.323 5.237 5.433 t0 [HJD′] 7653.541 ± 0.055 7653.510 ± 0.063 7653.522 ± 0.069 7653.637 ± 0.142 7653.124 ± 0.120 u0 0.093 ± 0.003 0.093 ± 0.003 0.093 ± 0.003 0.087 ± 0.007 0.065 ± 0.007 tE [days] 97.872 ± 2.143 96.546 ± 1.952 97.192 ± 1.903 100.772 ± 6.035 134.427 ± 9.191 s 1.086 ± 0.005 0.702 ± 0.070 1.746 ± 0.174 0.241 ± 0.021 6.090 ± 0.751 q(×10−4) 17.348 ± 2.181 48.054 ± 30.889 63.992 ± 26.388 L L q L L L 0.350 ± 0.112 0.797 ± 0.213 á ñqlog10 −2.781 ± 0.058 −2.151 ± 0.168 −2.194 ± 0.147 −0.429 ± 0.112 −0.285 ± 0.167 α [rad] 2.054 ± 0.017 2.092 ± 0.018 2.075 ± 0.016 0.708 ± 0.041 3.805 ± 0.044 ρ* 0.041 ± 0.002 <0.046 <0.044 <0.040 <0.033 πE,N 0.158 ± 0.050 0.166 ± 0.046 0.173 ± 0.045 0.064 ± 0.080 0.074 ± 0.048 πE,E −0.246 ± 0.010 −0.246 ± 0.010 −0.241 ± 0.011 −0.258 ± 0.011 −0.182 ± 0.013 Note. HJD′ = HJD–2,450,000.0. For the ρ* parameter, the cases with the inequality signs are upper limits on the ρ* (i.e., 3σ) because we cannot robustly measure ρ*. Figure 15. Test of the xallarap effect for OGLE-2016-BLG-1408, which shows χ2 value for each rotation period (P) of the binary source system. The red cross indicates the best-fit χ2 of the APRX model, which is equal to 1 yr. 22 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. of either observable. Thus, we conduct a Bayesian analysis to estimate the lens properties for the new planetary systems. We follow the formalism and procedures of the Bayesian analysis described in Shin et al. (2023a, 2023b). In Table 15, we present the lens properties estimated from the Bayesian analyses for each event. Note that we apply the tE and ρ* distribution constraints to the Bayesian analyses for all planetary events. For each event, we present several lens Figure 16. Light curves of KMT-2016-BLG-2399 with 2L1S and 1L2S models. Note model B has a sharp feature near HJD′ ∼ 7620 that lacks data coverage. We also present the caustic geometries of the 2L1S models. Table 12 The Parameters of 2L1S and 1L2S Models for KMT-2016-BLG-2399 Parameter A: 2L1S B: 2L1S C: 2L1S Parameter D: 1L2S χ2/Ndata 630.708/632 640.907/632 645.397/632 χ2/Ndata 644.879/632 Δχ2 L (best fit) 10.199 14.689 Δχ2 14.171 t0 [HJD′] 7633.696 ± 0.088 7633.798 ± 0.067 7633.450 ± 0.090 t0,S1 [HJD′] 7625.980 ± 0.097 u0 0.084 ± 0.007 0.173 ± 0.020 0.462 ± 0.061 u0,S1 0.002 ± 0.085 tE [days] 19.173 ± 1.219 13.168 ± 1.265 7.970 ± 0.587 tE [days] 11.171 ± 1.991 s 1.229 ± 0.016 1.646 ± 0.062 0.607 ± 0.024 t0,S2 [HJD′] 7634.150 ± 0.068 q 0.057 ± 0.009 0.030 ± 0.006 0.049 ± 0.008 u0,S2 0.221 ± 0.150 á ñqlog10 −1.227 ± 0.068 −1.580 ± 0.104 −1.303 ± 0.069 qflux 12.943 ± 2.610 α [rad] 2.530 ± 0.017 2.858 ± 0.015 −1.011 ± 0.031 ρ*,S1 <0.430 ρ* 0.006 ± 0.003 <0.038 0.111 ± 0.023 ρ*,S2 <1.038 Note. HJD′ = HJD−2,450,000.0. For the ρ* parameter, the cases with the inequality signs are upper limits on the ρ* (i.e., 3σ) because we cannot robustly measure ρ*. 23 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. properties because of the degenerate solutions. Thus, we present “adopted” values for ease of cataloging, which are weighted average values described in Jung et al. (2023). 5.1. OGLE-2016-BLG-1598 The planetary lens system of this event consists of a sub- Jupiter-mass planet (Mplanet∼ 0.37 or ∼0.70MJ) orbiting an early M-dwarf host star (Mhost∼ 0.55, Me) with a projected separation of ∼2.5 or ∼1.9 au. This planetary system is located at a distance of ∼5.9 kpc from us. The properties of the planetary system are those of a typical microlensing planet, i.e., a Jupiter-class planet orbiting an M-dwarf host beyond the snow line (Ida & Lin 2005; Kennedy & Kenyon 2008). 5.2. OGLE-2016-BLG-1800 For this event, the lens system is composed of a super Jupiter-mass planet (Mplanet∼ 2.49 or ∼2.77MJ) and an M-dwarf host star (Mhost∼ 0.41, Me). The planet orbits the host with a projected separation of ∼1.5 or ∼2.6 au. The system is located at a distance of ∼6.5 kpc from us. This planetary system is also one that is typical for microlensing planets. Figure 17. Light curves of KMT-2016-BLG-2473 with 2L1S and 1L2S models, their residuals, and the 2L1S caustic geometry. Table 13 The Parameters of 2L1S and 1L2S Models for KMT-2016-BLG-2473 Parameter 2L1S Parameter 1L2S χ2/Ndata 1529.729/1535 χ2/Ndata 1540.057/1535 Δχ2 L (best fit) Δχ2 10.328 t0 [HJD′] 7491.437 ± 0.206 t0,S1 [HJD′] 7515.409 ± 0.322 u0 0.088 ± 0.009 u0,S1 0.033 ± 0.036 tE [days] 47.179 ± 4.634 tE [days] 51.742 ± 11.156 s 1.217 ± 0.026 t0,S2 [HJD′] 7490.605 ± 0.160 q 0.011 ± 0.003 u0,S2 0.103 ± 0.043 á ñqlog10 −1.967 ± 0.137 qflux 7.663 ± 1.829 α [rad] 0.178 ± 0.012 ρ*,S1 <0.116 ρ* 0.019 ± 0.006 ρ*,S2 <0.275 Note. HJD′ = HJD−2,450,000.0. For the ρ* parameters, the cases with the inequality signs are upper limits on the ρ* (i.e., 3σ) because we cannot robustly measure ρ*. 24 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. 5.3. MOA-2016-BLG-526 Despite several solutions, the Bayesian results indicate that the properties of the host star are consistent, i.e., it is an M-dwarf star with the mass of ∼0.4Me. However, because of the variation in mass ratios for different solutions, the planet could be either a sub-Neptune-mass or Neptune-class planet (see Table 15). This planet orbits the host with a projected separation a⊥∼ 2 au. This planetary system is located at a distance of ∼6.9 kpc from us. 5.4. KMT-2016-BLG-2321 Bayesian results show that the lens system of this event consists of a Jupiter-class planet (Mplanet∼ 0.94 or ∼0.98MJ) orbiting a mid-K-type host star (Mhost∼ 0.73, Me) with a projected separation of ∼3.4 or ∼3.8 au. The system is located at the distance of ∼3.6 or ∼3.5 kpc. Note that, for this event, the constraints from the ρ* distributions have a major effect on the posteriors, in contrast to the other cases presented above. Indeed, we can expect the effect of the ρ* constraints to be significant as described in Section 4. Specifically, for this event, μrel,+3σ∼ 4 mas yr−1, which is much larger than 1 mas yr−1. Meanwhile, for the other events, the effects of the ρ* constraints were minor, as would be expected from lower limits of μrel,+3σ 1 mas yr−1 (see Table 14). 6. Summary and Discussion Through our systematic planetary anomaly search, we found four hidden planets and seven planet candidates in the 2016 KMTNet subprime fields. The properties of these new planetary systems are those of typical microlensing planets, i.e., giant planets orbiting M-dwarf host stars beyond their Figure 18. Color–magnitude diagrams (CMDs) of four planetary events. 25 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. snow lines. Although these new planets show typical properties discovered by the microlensing method, these are complemen- tary planet samples compared to samples discovered by other detection methods because of the different detection sensitiv- ities of each method (Clanton & Gaudi 2014a, 2014b; Shin et al. 2019). Table 14 CMD Analyses of Planetary Events Event Case (V − I)RGC (V − I)0,RGC (V − I)S (V − I)0,S (V − I)B θ* θE μrel IRGC I0,RGC IS I0,S IB (μas) (mas) (mas yr−1) OB161598 Outer 2.012 1.060 1.785 ± 0.022 0.833 ± 0.055 1.161 ± 0.008 1.080 ± 0.065 >0.013 >0.127 15.648 14.385 18.886 ± 0.006 17.623 ± 0.006 17.743 ± 0.004 L L L Inner 2.012 1.060 1.784 ± 0.022 0.832 ± 0.055 1.165 ± 0.008 1.068 ± 0.065 >0.018 >0.175 15.648 14.385 18.907 ± 0.006 17.645 ± 0.006 17.736 ± 0.004 L L L OB161800 s− 2.489 1.060 1.991 ± 0.058 0.562 ± 0.076 L 0.591 ± 0.046 >0.010 >0.185 16.296 14.450 20.139 ± 0.012 18.294 ± 0.012 L L L L s+ 2.489 1.060 1.991 ± 0.059 0.562 ± 0.078 L 0.611 ± 0.048 >0.009 >0.158 16.296 14.450 20.066 ± 0.012 18.220 ± 0.012 L L L L MB16526 s− (A) 2.203 1.060 1.767 ± 0.015 0.624 ± 0.052 L 0.716 ± 0.039 >0.102 >1.918 16.136 14.611 19.536 ± 0.007 18.011 ± 0.007 L L L s− (B) 2.203 1.060 1.763 ± 0.015 0.620 ± 0.052 L 0.713 ± 0.039 >0.079 >1.488 16.136 14.611 19.538 ± 0.007 18.012 ± 0.007 L L L s− (C) 2.203 1.060 1.767 ± 0.015 0.624 ± 0.052 L 0.728 ± 0.040 >0.081 >1.552 16.136 14.611 19.501 ± 0.007 17.976 ± 0.007 L L L s+ (A) 2.203 1.060 1.762 ± 0.015 0.619 ± 0.052 L 0.709 ± 0.039 >0.101 >1.891 16.136 14.611 19.548 ± 0.007 18.023 ± 0.007 L L L s+ (B) 2.203 1.060 1.764 ± 0.015 0.621 ± 0.052 L 0.717 ± 0.039 >0.102 >1.934 16.136 14.611 19.527 ± 0.007 18.002 ± 0.007 L L L KB162321a Outer 3.863 1.060 3.566 ± 0.077 0.763 ± 0.092 1.740 ± 0.238 0.825 ± 0.085 >0.563 >3.617 17.762 14.378 21.425 ± 0.012 18.041 ± 0.012 22.898 ± 0.151 L L L Inner 3.863 1.060 3.555 ± 0.075 0.752 ± 0.090 1.719 ± 0.240 0.818 ± 0.083 >0.645 >4.137 17.762 14.378 21.417 ± 0.012 18.033 ± 0.012 22.938 ± 0.156 L L L Notes. We use the abbreviation for event names, e.g., OGLE-2016-BLG-1598 is abbreviated as OB161598. a For KB162321, we note that the V and I magnitudes are in units of the instrumental scale of the KMTNet. Because there is no available OGLE-III catalog for this event, the magnitude system is not scaled to the OGLE-III. Table 15 Lens Properties of Planetary Events Event Constraints Case Mhost Mplanet DL a⊥ μrel Gal. Mod. χ2 (Me) (MJ/MN/M⊕) (kpc) (au) (mas yr−1) OB161598 tE + ρ* outer - +0.55 0.32 0.34 - + M0.37 0.30 0.31 J - +5.91 2.27 1.28 - +2.51 0.92 0.89 - +4.65 1.75 2.76 0.714 1.000 inner - +0.55 0.32 0.34 - + M0.70 0.46 0.48 J - +5.91 2.27 1.28 - +1.92 0.70 0.67 - +4.63 1.74 2.77 1.000 0.015 Adopted 0.55 ± 0.32 0.38 ± 0.30 MJ 5.91 ± 1.74 2.50 ± 0.88 4.65 ± 2.21 OB161800 tE + ρ* s− - +0.41 0.26 0.33 - + M2.49 1.65 2.13 J - +6.47 1.65 1.14 - +1.49 0.56 0.59 - +6.36 2.33 2.90 1.000 1.000 s+ - +0.40 0.26 0.33 - + M2.77 1.83 2.67 J - +6.49 1.63 1.14 - +2.60 0.97 1.05 - +6.43 2.35 2.93 0.850 0.632 Adopted 0.41 ± 0.22 2.59 ± 1.46 MJ 6.48 ± 1.03 1.88 ± 0.51 6.38 ± 1.94 MB16526 tE + ρ* s− (A) - +0.37 0.23 0.34 - + M1.27 0.87 1.21 N - +6.92 1.98 1.29 - +1.99 0.71 0.79 - +6.09 2.19 2.84 0.908 1.000 s− (B) - +0.37 0.23 0.34 - + M0.84 0.60 0.81 N - +6.94 1.98 1.29 - +2.00 0.73 0.81 - +6.06 2.22 2.85 0.913 0.999 s− (C) - +0.36 0.23 0.34 - + M1.12 0.77 1.14 N - +6.95 1.96 1.28 - +2.01 0.72 0.81 - +6.11 2.22 2.86 1.000 0.896 s+ (A) - +0.38 0.23 0.34 - + M0.44 0.30 0.61 N - +6.90 1.98 1.29 - +2.27 0.79 0.89 - +6.15 2.15 2.81 0.902 0.652 s+ (B) - +0.37 0.23 0.34 - + M0.34 0.22 0.70 N - +6.91 1.98 1.29 - +2.20 0.77 0.86 - +6.17 2.17 2.82 0.918 0.596 Adopted 0.37 ± 0.13 0.87 ± 0.38 MN 6.93 ± 0.75 2.07 ± 0.35 6.11 ± 1.15 KB162321 tE + ρ* outer - +0.73 0.36 0.41 - + M0.94 0.57 0.55 J - +3.64 1.60 1.84 - +3.41 1.20 0.96 - +5.78 1.55 2.79 0.992 1.000 inner - +0.73 0.36 0.43 - + M0.98 0.59 0.59 J - +3.49 1.54 1.77 - +3.81 1.36 1.10 - +6.15 1.62 2.78 1.000 0.717 Adopted 0.73 ± 0.28 0.95 ± 0.41 MJ 3.58 ± 1.22 3.58 ± 0.81 5.94 ± 1.56 Note. For the planet mass, we present the value in Jupiter (MJ), Neptune (MN), or Earth (M⊕) masses as appropriate. 26 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. In Table 16, we present all planetary events observed in 2016, including the new planets of this work. Both the by-eye and the AF methods were used to identify these planets. This work contributes 31% of the total number of planets discovered in the 2016 KMTNet subprime fields. Similarly to the contribution of this work, Shin et al. (2023b) reported five planets, which contributed 33% of the total number of planets discovered in 2016 in the prime fields. Hence, for the high- and low-cadence fields, we found a similar fraction of hidden planets. Despite the number of new planets in both fields being similar, the number of new planet candidates shows a big difference. Shin et al. (2023b) found only one planet candidate in the high-cadence fields. By contrast, we found seven planet candidates in the low-cadence fields. These events are treated as planet candidates because we cannot resolve the binary/ planet or 2L1S/1L2S degeneracy, which is caused by nonoptimal coverage of the anomalies. This fact clearly shows the importance of high-cadence observations to conclusively claim planet detections. Now that we have finished the systematic search work for both prime and subprime fields observed in 2016, 2018, and 2019, in Figure 19, we present the cumulative number of planets discovered by the AF and by eye as functions of ( )qlog10 . For each year, we find that 86%(=6/7), 55%(=6/11), and 75%(=9/12) of total planetary events having < -( )qlog 3.010 were identified by the AF method, respectively. Combining the three seasons, a 70% (=21/30) of planetary systems in the region of < -( )qlog 3.010 were discovered by the AF method rather than by eye. This is a remarkable result. Indeed, a total of 53 planetary events were identified by the conventional method (i.e., by eye) in the 2016, 2018, and 2019 seasons. However, only 17%(=9/53) of those planetary systems have < -( )qlog 3.010 . This lack of planet abundance in the region of < -( )qlog 3.010 is unexpected considering the fact that microlensing detections are only weakly dependent on the mass of the planet (∝q1/2), and KMTNet’s near- continuous observations should easily capture, e.g., the ∼8 hr signals due to ~ -( )qlog 410 planets. However, this investigation simply shows that most of the planetary systems having < -( )qlog 3.010 were just buried in the archive and missed by Table 16 Planetary Events Discovered or Recovered by the KMTNet AnomalyFinder in 2016 Event Name KMT Name KMT Field ( )qlog10 s Degeneracy Method References KB161105 ... subprime −5.19 1.14 i/o, c/w AF Zang et al. (2023) OB160007 KB161991 prime −5.17 2.83 ... AF Zang et al. (2024) OB161195a KB160372 prime −4.34 0.99 c/w, ecliptic by eye Gould et al. (2023) OB161850 KB161307 prime −4.00 0.80 i/o, ecliptic AF Shin et al. (2023b) OB161598 KB160696 subprime −3.19 0.96 i/o AF This work KB162321 ... subprime −2.91 1.04 i/o AF This work OB161067 KB161453 subprime −2.84 0.81 s-degen., ecliptic by eye Calchi Novati et al. (2019) OB161093 KB161345 subprime −2.84 1.02 ecliptic by eye Shin et al. 2022 MB16319 KB161816 prime −2.41 0.82 i/o by eye Han et al. (2018) KB162397 ... subprime −2.40 1.15 c/w by eye Han et al. (2020b) MB16532 KB160506 prime −2.39 0.65 c/w AF Shin et al. (2023b) KB161836 ... prime −2.35 1.30 c/w, ecliptic by eye Yang et al. (2020) OB161800 KB160781 subprime −2.24 0.69 c/w AF This work KB162364 ... subprime −2.12 1.17 ... by eye Han et al. (2020b) OB161227 KB161089 subprime −2.10 3.68 i/o by eye Han et al. (2020a) MB16227 KB160622 prime −2.03 0.93 ... by eye Koshimoto et al. (2017) OB160596 KB161677 prime −1.93 1.08 ... by eye Mróz et al. (2017) KB162605 ... prime −1.92 0.94 ... by eye Ryu et al. (2021) OB161190 KB160113 prime −1.84 0.60 ecliptic by eye Ryu et al. (2018) KB161397 ... subprime −1.80 1.68 c/w by eye Zang et al. (2018) OB161635 KB160269 prime −1.59 0.59 c/w AF Shin et al. (2023b) OB160263 KB161515 subprime −1.51 4.72 α-degen. by eye Han et al. (2017a) KB161107 ... subprime −1.44 0.35 c/w by eye Hwang et al. (2019) MB16526 KB161611 subprime −3.75 0.94 c/w, i/o AF This work KB160625 ... prime −3.63 0.74 c/w AF Shin et al. (2023b) OB160613b KB160017 prime −2.26 1.06 c/w by eye Han et al. (2017b) KB161751 ... prime −2.19 1.05 c/w AF Shin et al. (2023b) KB161855c ... prime −1.61 3.80 c/w, α, offset, 1L2S AF Shin et al. (2023b) KB160212 ... prime −1.43 0.83 c/w by eye Hwang et al. (2018) KB161820 ... prime −0.95 1.40 ... by eye Jung et al. (2018) KB162142c ... prime −0.69 0.97 c/w by eye Jung et al. (2018) Notes. The horizontal line separates planets expected to be part of the final statistical sample and those whose mass ratios are likely too uncertain or too large to be included. In the column of “Degeneracy,” we present the type of degeneracies for the solutions: “c/w,” “i/o,” “ecliptic,” “offset,” “α,” and “1L2S” indicate the close/ wide (c/w) degeneracy, inner/outer (i/o) degeneracy, ecliptic degeneracy of the microlens-parallax effect, offset-degeneracy, α-degeneracy (see Shin et al. 2023a), and 2L1S/1L2S degeneracy, respectively. Note that “s-degen.” indicates small/large s degeneracy (this is different from “c/w”; see Calchi Novati et al. 2019). a For OB161195, the properties of this planetary system were reported by Shvartzvald et al. (2017) and Bond et al. (2017). However, we adopt ( )qlog10 and s values from Gould et al. (2023), which reanalyze the event and measure a more precise mass ratio. b For OB160613, the event was caused by a lens system consisting of a planet and binary host stars. c For KB161855 and KB162142, these are planet candidates. 27 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. by-eye searches. This fact clearly shows the importance of our systematic search to building a complete microlensing planet sample. Acknowledgments This research has made use of the KMTNet system operated by the Korea Astronomy and Space Science Institute (KASI) at three host sites of CTIO in Chile, SAAO in South Africa, and SSO in Australia. Data transfer from the host site to KASI was supported by the Korea Research Environment Open NETwork (KREONET). This research was supported by KASI under the R&D program (project No. 2023-1-832-03) supervised by the Ministry of Science and ICT. The MOA project is supported by JSPS KAKENHI grant Nos. JP24253004, JP26247023, JP16H06287, and JP22H00153. J.C.Y. and I.-G.S. acknowledge support from N.S.F grant No. AST-2108414. Work by C.H. was supported by the grants of National Research Foundation of Korea (2019R1A2C2085965 and 2020R1A4A2002885). Y.S. acknowledges support from BSF grant No. 2020740. W.Z. and H.Y. acknowledge support by the National Natural Science Foundation of China (grant No. 12133005). W.Z. acknowledges the support from the Harvard-Smithsonian Center for Astro- physics through the CfA Fellowship. The computations in this paper were conducted on the Smithsonian High Performance Cluster (SI/HPC), Smithsonian Institution (https://doi.org/10. 25572/SIHPC). This research has made use of the NASA Exoplanet Archive, which is operated by the California Institute of Technology, under contract with the National Aeronautics and Space Administration under the Exoplanet Exploration Program. Appendix Nonplanetary Events From the preliminary analysis using the pipeline data sets, we find that some events in the 2016 subprime fields have the potential to be caused by planetary lens systems (i.e., q< 0.06). However, based on the detailed analysis using the TLC data sets, we reveal that these events were caused by binary-lens systems. Figure 19. Cumulative number of planets discovered by the AF and eye as a function of ( )qlog10 . We present 2016, 2018, and 2019 cases that have finished the systematic search for both prime and subprime fields. 28 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. https://doi.org/10.25572/SIHPC https://doi.org/10.25572/SIHPC The events cannot satisfy our criteria (i.e., no competing planetary solutions having q< 0.03 and Δχ2< 10.0). Although the scientific importance is low for these events, we briefly document these binary-lens events for the record. This documentation will be helpful to avoid redundant efforts for planet searches using the KMTNet data archive. In Table 1, we list these nonplanetary events with their observational information. A.1. OGLE-2016-BLG-0620 The overall shape of the light curve of OGLE-2016-BLG- 0620 is a 1L1S-like feature. However, the 1L1S model exhibits residuals at the rising part of the left wing and around the peak. We find that these systematic residuals can be explained by the 2L1S interpretation, which gives Δχ2= 613.2 between 1L1S and 2L1S models. The best-fit 2L1S model has (s, q)= (2.449± 0.050, 0.208± 0.013), which indicates that the lens is a binary system. We also find a planet-like model (i.e., q= 0.027± 0.004). However, this case is worse than the best fit by Δχ2= 30.7, which does not satisfy our criterion (i.e., Δχ2< 10.0). Thus, we conclude that OGLE-2016-BLG- 0620 was caused by the binary rather than a planetary system. A.2. KMT-2016-BLG-0913 The light curve of KMT-2016-BLG-0913 shows an apparent anomaly at the peak. We find the best-fit 2L1S model has (s, q)= (2.492± 0.067, 0.823± 0.082). The best-fit model indicates that the lens system consists of binary stars. We check possible planetary models. We find two possible cases with (s, q)= (1.328± 0.032, 0.044± 0.008), and (s, q)= (0.700± 0.009, 0.032± 0.004). However, both cases are disfavored byΔχ2= 104.5 and 116.7. Furthermore, the mass ratios of both cases do not satisfy our criterion. Thus, we conclude that KMT-2016-BLG-0913 was caused by a binary- lens system. A.3. OGLE-2016-BLG-1432 The light curve of OGLE-2016-BLG-1432 shows an asymmetric feature. This anomaly can be explained by a 2L1S model with (s, q)= (1.423± 0.074, 0.257± 0.063). The best-fit model indicates the lens is a binary-lens system. We also find a planetary model having (s, q)= (1.603± 0.056, (69.475± 29.048)× 10−4). However, the planetary case is disfavored by Δχ2= 78.1. Thus, we conclude that OGLE- 2016-BLG-1432 was caused by a binary. A.4. KMT-2016-BLG-1222 The light curve of KMT-2016-BLG-1222 shows a clear anomaly at the peak. We find that the 2L1S interpretation can explain the anomaly. The best-fit solution with (s, q)= (1.926± 0.117, 0.113± 0.024) indicates the lens is a binary. We also find a degenerate solution caused by the well- known close/wide degeneracy. This close solution with (s, q)= (0.584± 0.027, 0.088± 0.017) has only Δχ2= 1.28, so the degeneracy cannot be resolved. But, the close solution also indicates the lens is a binary. In addition, we find that there is no possible model having q< 0.03 based on the detailed analysis using the TLC data sets. Thus, we conclude that KMT- 2016-BLG-1222 was caused by the binary. A.5. OGLE-2016-BLG-1844 The light curve of OGLE-2016-BLG-1844 exhibits two bump-shaped anomalies at the peak. The anomaly can be described by 2L1S models with (s, q)= (4.062± 0.413, 0.313± 0.088), or (s, q)= (0.337± 0.028, 0.171± 0.034) corresponding to the s+ or s− cases, respectively. The degeneracy between the s± solutions cannot be resolved (i.e., Δχ2= 0.7). Both s± cases indicate the lens is a binary system. We find possible planetary cases (i.e., q∼ (38± 5)× 10−4) with s+ (s= 1.289± 0.027), and s− (s= 0.761± 0.015), but they are disfavored by Δχ2= 61.7 and 62.1, respectively. Thus, we conclude that OGLE-2016-BLG-1844 was caused by a binary-lens system. A.6. KMT-2016-BLG-1425 The light curve of KMT-2016-BLG-1425 shows an apparent bump near the peak. The anomaly can be described by a 2L1S model with (s, q)= (1.169± 0.057, 0.314± 0.056) (the best-fit model). We find an alternative planetary model having (s, q)= (1.279± 0.062, 0.017± 0.012). However, this case is disfavored by Δχ2= 17.7, which does not satisfy our criterion (i.e., Δχ2< 10.0). Thus, we conclude that OGLE-2016-BLG- 0882 was caused by a binary-lens system. A.7. OGLE-2016-BLG-0982 The light curve of OGLE-2016-BLG-0982 shows a clear bump-shaped anomaly near the peak. We find the best-fit solution has (s, q)= (1.989± 0.086, 0.660± 0.090), which indicates the lens is a binary system. There are no possible planetary cases (i.e., q< 0.03). Among the competing solutions, the lowest q value is 0.083± 0.011, which is disfavored by Δχ2= 16.7. Thus, we conclude that OGLE- 2016-BLG-0982 was caused by a binary system. A.8. OGLE-2016-BLG-1517 The light curve of OGLE-2016-BLG-1517 shows asym- metric deviations from the 1L1S fitting. The anomaly can be explained by the 2L1S models with (s, q)= (3.259± 0.103, 0.460± 0.083). The best-fit solution indicates the lens is a binary. We have checked for possible planetary cases. However, we find that the model having the lowest q value (i.e., q= 0.014± 0.002) is disfavored by Δχ2= 60.0. Thus, we conclude that OGLE-2016-BLG-1517 was caused by a binary-lens system. A.9. OGLE-2016-BLG-1258 OGLE-2016-BLG-1258 is a low-magnification event with a bump at the peak. The anomaly can be described by 2L1S models with (s, q)= (0.573± 0.012, 0.192± 0.031), and (s, q)= (2.002± 0.085, 0.161± 0.062) corresponding to the s− and s+ cases, respectively. The s− case shows better fits than the s+ case by Δχ2= 7.2. Both solutions imply that the lens is a binary system. We find a possible planetary model (q= 0.011± 0.003), which is disfavored by Δχ2= 27.4. This planetary case is rejected based on our criterion. Thus, we conclude that OGLE-2016-BLG-1258 was caused by a binary- lens system. 29 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. A.10. KMT-2016-BLG-2256 The light curve of KMT-2016-BLG-2256 exhibits a bump-like anomaly, which is sparsely covered by KMTC only. The anomaly can be explained by both s± models with (s, q)= (0.580± 0.043, 0.629± 0.197), and (s, q)= (3.937± 0.292, 0.360± 0.410) corresponding to the s− and s+ cases, respectively. Although the s± cases cannot be resolved (i.e., Δχ2= 1.5), both cases indicate that the lens is a binary system. We find that there is no competing planetary solution. The lowest q model (i.e., q= 0.029± 0.005) is disfavored by Δχ2= 47.9, which is clearly rejected based on our criterion. Thus, we conclude that KMT- 2016-BLG-2256 was caused by a binary-lens system. A.11. KMT-2016-BLG-2331 The light curve of KMT-2016-BLG-2331 shows a bump at the peak, which is sparsely covered. The anomaly can be described by both 2L1S models with (s, q)= (0.317± 0.025, 0.280± 0.081), and (s, q)= (5.014± 0.646, 0.478± 0.289) corresponding to the s− and s+ cases, respectively (Δχ2= 2.7). Both s± cases indicate the lens is a binary. We find a possible planetary model having q= (92.366± 14.168)× 10−4. How- ever, this planet case is disfavored by Δχ2= 57.5, which is clearly rejected based on our criterion. Thus, we conclude that KMT-2016-BLG-2331 was caused by a binary-lens system. ORCID iDs In-Gu Shin https://orcid.org/0000-0002-4355-9838 Jennifer C. Yee https://orcid.org/0000-0001-9481-7123 Weicheng Zang https://orcid.org/0000-0001-6000-3463 Cheongho Han https://orcid.org/0000-0002-2641-9964 Hongjing Yang https://orcid.org/0000-0003-0626-8465 Chung-Uk Lee https://orcid.org/0000-0003-0043-3925 Andrzej Udalski https://orcid.org/0000-0001-5207-5619 Takahiro Sumi https://orcid.org/0000-0002-4035-5012 Michael D. Albrow https://orcid.org/0000-0003-3316-4012 Sun-Ju Chung https://orcid.org/0000-0001-6285-4528 Yossi Shvartzvald https://orcid.org/0000-0003-1525-5041 Michał K. Szymański https://orcid.org/0000-0002- 0548-8995 Jan Skowron https://orcid.org/0000-0002-2335-1730 Radosław Poleski https://orcid.org/0000-0002-9245-6368 Igor Soszyński https://orcid.org/0000-0002-7777-0842 Paweł Pietrukowicz https://orcid.org/0000-0002-2339-5899 Szymon Kozłowski https://orcid.org/0000-0003- 4084-880X Krzysztof A. Rybicki https://orcid.org/0000-0002- 9326-9329 Patryk Iwanek https://orcid.org/0000-0002-6212-7221 Krzysztof Ulaczyk https://orcid.org/0000-0001-6364-408X Marcin Wrona https://orcid.org/0000-0002-3051-274X Mariusz Gromadzki https://orcid.org/0000-0002-1650-1518 David P. Bennett https://orcid.org/0000-0001-8043-8413 Akihiko Fukui https://orcid.org/0000-0002-4909-5763 Yuki Hirao https://orcid.org/0000-0003-4776-8618 Yoshitaka Itow https://orcid.org/0000-0002-8198-1968 Naoki Koshimoto https://orcid.org/0000-0003-2302-9562 Yutaka Matsubara https://orcid.org/0000-0002-9629-4810 Shota Miyazaki https://orcid.org/0000-0001-9818-1513 Yasushi Muraki https://orcid.org/0000-0003-1978-2092 Kansuke Nunota https://orcid.org/0009-0005-3414-455X Greg Olmschenk https://orcid.org/0000-0001-8472-2219 Yuki Satoh https://orcid.org/0000-0002-1228-4122 Daisuke Suzuki https://orcid.org/0000-0002-5843-9433 References Alard, C., & Lupton, R. H. 1998, ApJ, 503, 325 Albrow, M. D., Horne, K., Bramich, D. M., et al. 2009, MNRAS, 397, 2099 Bennett, D. P., Bond, I. A., Udalski, A., et al. 2008, ApJ, 684, 663 Bond, I. A., Abe, F., Dodd, R. J., et al. 2001, MNRAS, 327, 868 Bond, I. A., Bennett, D. P., Sumi, T., et al. 2017, MNRAS, 469, 2434 Calchi Novati, S., Suzuki, D., Udalski, A., et al. 2019, AJ, 157, 121 Chang, K., & Refsdal, S. 1979, Natur, 282, 561 Clanton, C., & Gaudi, B. S. 2014a, ApJ, 791, 90 Clanton, C., & Gaudi, B. S. 2014b, ApJ, 791, 91 Gaudi, B. S. 1998, ApJ, 506, 533 Gould, A. 1992, ApJ, 392, 442 Gould, A. 2022, arXiv:2209.12501 Gould, A., Han, C., Zang, W., et al. 2022, A&A, 664, A13 Gould, A., Shvartzvald, Y., Zhang, J., et al. 2023, AJ, 166, 145 Griest, K., & Safizadeh, N. 1998, ApJ, 500, 37 Han, C., Bond, I. A., Gould, A., et al. 2018, AJ, 156, 226 Han, C., Udalsk, A., Gould, A., et al. 2020a, AJ, 159, 91 Han, C., Udalski, A., Gould, A., et al. 2017a, AJ, 154, 133 Han, C., Udalski, A., Gould, A., et al. 2017b, AJ, 154, 223 Han, C., Udalski, A., Kim, D., et al. 2020b, A&A, 642, A110 Holtzman, J. A., Watson, A. M., Baum, W. A., et al. 1998, AJ, 115, 1946 Hwang, K.-H., Ryu, Y.-H., Kim, H.-W., et al. 2019, AJ, 157, 23 Hwang, K.-H., Udalski, A., Shvartzvald, Y., et al. 2018, AJ, 155, 20 Hwang, K.-H., Zang, W., Gould, A., et al. 2022, AJ, 163, 43 Ida, S., & Lin, D. N. C. 2005, ApJ, 626, 1045 Jiang, G., DePoy, D. L., Gal-Yam, A., et al. 2004, ApJ, 617, 1307 Jung, Y. K., Hwang, K.-H., Ryu, Y.-H., et al. 2018, AJ, 156, 208 Jung, Y. K., Zang, W., Han, C., et al. 2022, AJ, 164, 262 Jung, Y. K., Zang, W., Wang, H., et al. 2023, AJ, 165, 226 Kennedy, G. M., & Kenyon, S. J. 2008, ApJ, 673, 502 Kim, D.-J., Kim, H.-W., Hwang, K.-H., et al. 2018, AJ, 155, 76 Kim, S.-L., Lee, C.-U., Park, B.-G., et al. 2016, JKAS, 49, 37 Koshimoto, N., Shvartzvald, Y., Bennett, D. P., et al. 2017, AJ, 154, 3 Mróz, P., Han, C., Udalski, A., et al. 2017, AJ, 153, 143 Poindexter, S., Afonso, C., Bennett, D. P., et al. 2005, ApJ, 633, 914 Ryu, Y.-H., Hwang, K.-H., Gould, A., et al. 2021, AJ, 162, 96 Ryu, Y.-H., Kil Jung, Y., Yang, H., et al. 2022, AJ, 164, 180 Ryu, Y.-H., Udalski, A., Yee, J. C., et al. 2024, AJ, 167, 88 Ryu, Y.-H., Yee, J. C., Udalski, A., et al. 2018, AJ, 155, 40 Shin, I.-G., Ryu, Y.-H., Yee, J. C., et al. 2019, AJ, 157, 146 Shin, I.-G., Yee, J. C., Gould, A., et al. 2023a, AJ, 165, 8 Shin, I.-G., Yee, J. C., Hwang, K.-H., et al. 2022, AJ, 163, 254 Shin, I.-G., Yee, J. C., Zang, W., et al. 2023b, AJ, 166, 104 Shvartzvald, Y., Maoz, D., Udalski, A., et al. 2016, MNRAS, 457, 4089 Shvartzvald, Y., Yee, J. C., Calchi Novati, S., et al. 2017, ApJL, 840, L3 Smith, M. C., Mao, S., & Paczyński, B. 2003, MNRAS, 339, 925 Sumi, T., Abe, F., Bond, I. A., et al. 2003, ApJ, 591, 204 Szymański, M. K., Udalski, A., Soszyński, I., et al. 2011, AcA, 61, 83 Tomaney, A. B., & Crotts, A. P. S. 1996, AJ, 112, 2872 Udalski, A., Szymanski, M., Kaluzny, J., et al. 1994, AcA, 44, 227 Udalski, A., Szymański, M. K., & Szymański, G. 2015, AcA, 65, 1 Wang, H., Zang, W., Zhu, W., et al. 2022, MNRAS, 510, 1778 Wozniak, P. R. 2000, AcA, 50, 421 Yang, H., Yee, J. C., Hwang, K.-H., et al. 2024, MNRAS, 528, 11 Yang, H., Zhang, X., Hwang, K.-H., et al. 2020, AJ, 159, 98 Yee, J. C., Shvartzvald, Y., Gal-Yam, A., et al. 2012, ApJ, 755, 102 Yoo, J., DePoy, D. L., Gal-Yam, A., et al. 2004, ApJ, 603, 139 Zang, W., Hwang, K.-H., Kim, H.-W., et al. 2018, AJ, 156, 236 Zang, W., Hwang, K.-H., Udalski, A., et al. 2021, AJ, 162, 163 Zang, W., Jung, Y. K., Yang, H., et al. 2023, AJ, 165, 103 Zang, W., Jung, Y. K., Yee, J. C., et al. 2024, Sci, submitted Zang, W., Yang, H., Han, C., et al. 2022, MNRAS, 515, 928 30 The Astronomical Journal, 167:269 (30pp), 2024 June Shin et al. https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0002-4355-9838 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-9481-7123 https://orcid.org/0000-0001-6000-3463 https://orcid.org/0000-0001-6000-3463 https://orcid.org/0000-0001-6000-3463 https://orcid.org/0000-0001-6000-3463 https://or