This article is a continuation of “Solutions for Sale: Inside the Discord Marketplaces Undermining America’s Top Math Competitions,” published earlier this year in February.
Within hours of scores releasing the Monday after the American Mathematics Competitions, the Art of Problem Solving (AoPS) forums — the closest thing the competitive mathematics world has to a town square — had collapsed into open warfare. Students who had spent years grinding through problem sets, sacrificing their school breaks and weekends, attending math circles, and paying for tutoring had opened their competition results and found that their efforts weren’t enough. They had been outscored by people who had barely studied for the test, paid a couple hundred dollars for its solutions, and qualified for the American Invitational Mathematics Examination (AIME), the first selective test in a long exam series that leads to the selection of the team representing the U.S. at the International Mathematical Olympiad.
The cutoff score to advance from the AMC 12, a version of the American Mathematics Competitions (AMCs), to the AIME, had jumped by nearly 20 points. It was the largest single-year increase in the history of the competition, and on platforms across the internet, furious students turned to public AoPS forum channels to look for someone to blame. Several blamed the Mathematical Association of America (MAA), who creates the AMC and AIME annually and certifies teachers and other officials to proctor it. “Now, it’s a deranged, bottomless pit,” one anonymous user wrote, watching the argument spiral. A February Tatler investigation exposed how nearly two dozen cheaters “leaked” the solutions to the AMC 10 and 12, versions of the national competition for students through grades 10 and 12, respectively. Sellers were able to register as proctors and bypass the MAA’s verification system, and they were also able to obtain exam questions from photographs of the exams covertly taken during the test. Cheating on the AMC has evolved from isolated leaks to an organized Discord-based marketplace where sellers distribute early test materials and solutions at large scales, which in turn increase the prevalence of high-performing competitors, thus raising cutoff scores.
In the eyes of major academic prep organizations, the American Mathematics Competitions are a gate. These prep organizations market to parents nationwide that exemplary performance on these exams garners the attention of college admissions officers, and serves as a stepping stone to careers in finance and research. According to the Ivy League Education Center, a tutoring firm specializing in competition mathematics, “Good AMC scores will greatly enhance admission opportunities for students to elite colleges.” In a 2023 open letter to the MAA, the Areteem Institute, a nationally accredited math education organization, wrote that qualifying for the AIME, “can be what sets one college applicant apart from another.” The center also claims that the AMC is more important than the SAT, writing on their website that if students are serious about getting into prestigious colleges such as Caltech or MIT, “you should be focused as much on the AMCs as you are on the SATs. In fact, we think the AMC is more important.”
Clearly, the AMC series plays an important role in the academic lives of many high school students, including Lakesiders. Each year, approximately 50 Lakeside students participate in the American Mathematics Competitions series through the school’s math team, according to the team’s faculty coach Dean Ballard. In recent years, roughly 30 of those students have qualified annually for the AIME. Lakesiders have seen international success in these competitions in past years, notably Alex Zhao ’24’s gold medal at the International Mathematical Olympiad, the most renowned competitive mathematics competition for precollege students. According to competition results published by the MAA, Lakeside consistently sends five or more students to the USAJMO and the USAMO, both of which are the next step in the AMC series after the AIME.
Mr. Ballard, who has proctored the AMC and AIME for decades, said, “If it wasn’t for the over 250 competitors who received perfect scores, then those who didn’t quite achieve the cutoff this year likely would have.” As Tatler covered in the first article in this series, when cutoff scores increase as a result of cheating, it becomes increasingly difficult for tests to accurately measure the ability of hard-working and honest students in the sciences and mathematics, thus damaging the chances of hundreds of teenage students of getting accepted into the elite universities they are seeking to enter. One competitor stated on Reddit that “If [the AMC is] a flawed signal then [colleges] absolutely shouldn’t want to be associated with it,” while another wrote, “the test completely lost any meaning.”
Tatler reviewed more than 300 messages and dozens of images posted by exam sellers across multiple Discord servers dating from the spring of 2025, and interviewed three individuals involved in operating cheating “communities” with membership counts ranging from several dozen to several hundred users. Details from those investigations can be found in the first article of this series, published by Tatler in February 2026. During this investigation, Tatler found irrefutable evidence that exam solutions and answer keys were being sold for multiple nationally-administered academic tests across at least 30 different Discord servers, with an estimated market volume of $300,000 or higher. This figure is based on the amount of published “reviews” on these servers who post photo proof of successful test scores, multiplied by the unit prices of the products they’re vouching for. Tatler recorded 3,233 unique members across 27 servers dedicated towards cheating at the time of this article’s writing, with leaks of the USAMO and JMO problems posted on at least eight platforms with a combined membership of 1,511 within 30 minutes of the test beginning. This means that any competitor with access to an electronic device, such as an Apple Watch or an iPhone, would be able to access the olympiad problems as early as several hours prior to the exam.
A 2024 Change.org petition calling on the MAA to conduct a thorough investigation of leaks and to address test security issues garnered 1,279 signatures. Two 2025 online petitions have since circulated on the website proposing changes to the ways in which the exams are administered. One calls for additional test security measures, while the other calls on the MAA to lower the AIME cutoff. In one of the petitions, a supporter wrote, “The abnormal bimodal distribution [of scores] and the year by year … trend in cheating … has made learners question their own efforts and gave birth to an entire market of cheating and selling answers.” In the recent AMC 12A alone, more than 250 students received perfect scores, according to competition data published by the MAA. In previous years, that number rarely exceeded 30, with the highest number of perfect scores being 46. Until 2020, the AMC 12 cutoff for AIME qualification was capped at 100 points. The cutoff for the distinguished honor roll, which recognizes the top 1% of scorers, required a perfect 150 points. Students who had scored in the 100s by their own hard work — scores that, in any prior year, would guarantee an honor placement — were cut, their hard work rendered invisible by a flood of purchased answers.
On one of the petitions, a student at Allen High School in Allen, Texas, wrote, “I’m furious. More than I’ve been in years. I’m talking to people crying. Like actual crying. People who studied for months, people who did it honestly, people who were finally feeling hopeful, and then the cutoffs drop and it’s just… unreal.” He added, “Keeping inflated cutoffs is not an anti-cheating strategy. It’s just moving the goalposts while the field is full of people playing unfairly … Once trust is gone, the contest turns into a leaderboard for who had access, and frankly money with connections.”
The MAA, which has administered the AMC since 1950, responded to Mr. Ballard’s inquiries about the 2025 score distributions with a statement it also sent to other community members who wrote in about cheating and a possible redress of the AIME cutoff scores. The email was received by both Mr. Ballard and two anonymous Reddit users who both inquired about cheating and a redress. They separately posted the identical text, addressing both of them as “Community Member.” A third Reddit user wrote in asking about cutoff scores, only to receive an email saying “Thank you for contacting MAA. We appreciate your feedback.” The MAA wrote in their response that protecting the integrity of the competitions is a “high priority,” and the association “has strengthened security practices across exam development, administration, and review.” In their email, the MAA also added that they “actively investigate and address cases that raise concerns using both established and digital review processes.” It did not describe what those processes or practices were, nor did it say how many investigations had been opened. There was no information about whether any seller had ever been identified, referred to authorities or law enforcement, or disqualified. Kristin Morris, director of marketing & communications of the MAA, stated, “We did receive reports of cheating, including reports related to Discord specifically, and our review processes identified confirmed violations. Action was taken, including disqualifications and, in some cases, life bans. We do not share the number of cases, the identities of those involved, or the criteria used to identify violations. We do not share details of these violations to protect the privacy of competitors and our security practices.” Morris added that the “AIME was released to competition managers just 30 minutes before the exam window opened.” Tatler was able to independently verify that the second of two administrations of the 2026 AIME was posted online at 5:46 a.m. PST the day of the contest, nearly five hours before the exam window opened at 11 a.m. PST.
Tatler spoke with Danny Shi, an assistant professor of mathematics at the University of Washington and the Washington State site director for this year’s administration of the USAMO and JMO. According to Shi, there was a large push for stricter anti-cheating measures this year. He credited these efforts to “many tips and pieces of information from the broader community,” which included parents of contestants and former members of the U.S. International Mathematical Olympiad team. These measures included purchasing high-grade metal detectors and roughly a 1:4 proctor-to-student ratio for 36 students. “The proctors were responsible for checking photo IDs, making sure the bathrooms were phone-free, and using metal detectors to make sure students were not smuggling phones and smart watches into the secured area,” says Shi, who adds that they were “highly effective.”
These concerns about cheating were also not unfounded. According to Shi, there was evidence nationwide of efforts to leak the exams. In his words, “Whatever you can think of, someone will have tried it.” 20 minutes into the administration of the 2026 USAMO and JMO, both tests’ problems were leaked on at least two Discord servers. Tatler was able to verify the timestamps and validity of the Discord messages containing the leaks with Shi, who noted that the leaks were likely reprinted versions of the actual problems, based on differences in the ways the problems were typeset on the actual version and the reprint. Tatler was able to verify that at least one of the leaks was generated by Claude. In Shi’s words, “some other sites that did not have metal detectors ran into problems, including students planting phones in toilet seat cover boxes, students hiding smartwatches in their sleeves, and students copying test problems, bringing them into the bathroom, and taking pictures with their phones to send to someone else.”
According to Shi, at the Washington site, the security measures worked as a deterrent against some students who “got scared and did not even show up because they realized there was no way for them to cheat anymore, and they did not want to sit in the contest for 4.5 hours each day.” He added that to his knowledge, and based on the scores from the Washington site, no cheating happened at the USAMO and JMO administered in Washington.
A 2024 Wall Street Journal investigation documented the 2023 leak of AMC materials in China. A 2026 article in The Free Press observed that “in the past, tests were stolen by unscrupulous proctors and sold online. Now with AI tools, if an exam is stolen, it can be instantly solved — and sold along with the answers.” According to Mr. Ballard, the process used to determine the cutoff score has changed several times over the decades. He described his experience as a participant in the American High School Mathematics Examination, the predecessor to the AMC, and later as a proctor of the exams. According to Mr. Ballard, in earlier years the MAA used to set a fixed cutoff score prior to the administration of the exam, where any number of students who scored above 100 points out of 150 would advance to the AIME. The organization later shifted to a model targeting a fixed number of qualifying students, and today, they use a fixed percentile of competitors to set the cutoff score for the AIME.
For students who score near the margin of qualification, Mr. Ballard noted that even small changes in cutoff scores have prevented some Lakeside students from advancing to the next stages of competition. Lakeside competitors who discussed their test results with Tatler said that the unusually high cutoffs this year left them uncertain about whether the competition still accurately reflects the abilities of the participants. Gabriel R. ’28, who missed the 2025 AMC 10 cutoff by one question, told Tatler that he “started to see the whole AMC series as pointless, especially for college applications.” As soon as he saw the cutoff of 105 points out of 150, he “knew there was something wrong,” later adding, “My rightful spot in the AIME was stolen from me by someone who may or may not have deserved it, but definitely someone who cheated.” When asked about whether he believed that honest competitors were punished, he responded that he definitely did, adding that “99 [points out of 150] on the 10A would normally get me into AIME, looking at past years’ cutoffs, but it didn’t. I’m sure there were plenty of others like me who were unfortunately pushed out.” In terms of what the MAA should have done, Gabriel said it wasn’t immediately clear whether the organization had done anything to address the issue, adding that “figuring out who and how many is hard to do with low test security at schools.” Shuwen Y. ’29, who has participated in the AMC series since 5th grade, spent three to four hours a week leading up to the AMC 10, and told Tatler that “many of my friends and classmates expected my score to qualify.” She missed the AIME cutoff by 4.5 points, a value smaller than the six points allotted for one correct answer. When asked about her response to the cutoffs, she said she was surprised and that they “changed my view of the MAA organization,” adding that “I haven’t seen this much disregard for the effects of it until now. It does make me question MAA’s integrity.” In response to a request for comment, the MAA stated that it “does not recommend or endorse AMC scores as sole indicators for assessments of mathematical talent or for academic placement or selection decisions.”
In the weeks that followed the announcement of this year’s cutoff scores, a second controversy emerged within the mathematics competition community regarding the process used to determine advancement from the AIME to the USAJMO and USAMO, or the USA(J)MO. Unlike the AMC cutoffs, which are publicly released shortly after exams are graded, the exact thresholds used to determine Olympiad qualification were not disclosed for the 2025-26 competition cycle according to a message posted from the verified account of Dr. M. Tip Phaovibul, who is a member of the MAA USAMO committee. Those thresholds have been released in previous years. The MAA did not respond to a request for comment on why the cutoffs were not publicized, nor how they were determined. These thresholds were published along with the individual’s results on the MAA website. In the absence of official data, competitors — including Lakeside’s Raymond Z. ’27 — began attempting to estimate the qualification thresholds themselves. Raymond, along with other members of the competitive mathematics community, put together a spreadsheet of those who qualified and their self-reported scores, called indexes. An index is measured by taking your AMC 10 score and adding it to ten times your AIME score, and is used to determine candidacy for the USA(J)MO. They estimated cutoff scores by using the individual who qualified for neither exam with the highest index, and they found that individuals who had lower indexes than the estimated cutoff still advanced to the USAMO. For example, two individuals self-reported an index of 378, and yet only one advanced to the USAMO, while the other didn’t advance to either. According to Raymond, who has advanced through several levels of the American mathematics competition system over the past several years, the controversy has intensified frustration among students who believe that the earlier stages of the competition pipeline may no longer accurately reflect the abilities of competitors. “There are a lot of deserving people being cut out of this,” Raymond said in an interview with Tatler.
He pointed out that the structure of the competition system means that many students who might perform well on the proof-based Olympiads that come later in the exam series may never have the opportunity to actually demonstrate their abilities if they don’t advance to the next phase, if widespread cheating drives up a cutoff score. “I think they should just release the cutoff scores,” Raymond said. “At least then people would understand what happened.” This comes as rumors continue to circulate amongst online mathematics communities that some qualifiers for the USAMO and USAJMO were handpicked despite their scores falling below an estimated cutoff, determined by students self-reporting their exam scores and whether they advanced to the Olympiads or not. This data was collected through an informal Google Form made by Raymond, who shared the results with Tatler. The MAA did not comment on several questions from Tatler regarding the selection of AMO- and JMO-qualifying students, writing, “The MAA AMC is a U.S.- and Canada-based mathematics competition, and its deepest value is in the quality of the mathematics itself. Every year, hundreds of thousands of students, classroom teachers, tutors, and math circle organizers around the world engage with AMC problems because they represent interesting, high-quality math problems that inspire young people to continue exploring problem-solving and the joy of mathematics.”
While the MAA has stated that protecting the integrity of its competitions remains a priority, many competitors continue to question whether the current structure of the competition system can effectively address the challenges posed by both a rapidly-growing interest in the competition series, as well as a growing underground industry of selling exam materials and solutions. The answer to that question will determine whether the contests continue to serve as a reliable measure of the intelligence and mathematical skill of Lakeside students — and of students across the country.
