MacKenzie Lung Swam the 2026 NCAA Championships While Pregnant: Breaking Down the 0.97 and 2.76 Seconds

On the official results sheet of the 2026 NCAA Swimming and Diving Championships, MacKenzie Lung's times read like a routine loss of form. She arrived as the 8th seed in the 100 breaststroke with a breakout 57.92 from the mid-season Trailblazer Invitational, and the 10th seed in the 200 breaststroke at 2:07.03. She left with 15th in the 100 breaststroke (58.89) and 29th in the 200 breaststroke (2:09.79) — 0.97 seconds added in the short race, 2.76 seconds in the long one. To anyone who reads results for a living, this is familiar data: a swimmer with a strong seed slips at the biggest meet of the year. My models would tag it with familiar labels — championship pressure, a crowded schedule, a mid-season peak that did not repeat. This week, on her TikTok channel, Lung supplied a variable no model of mine carries: she was pregnant. "And I was confused why I swam so bad," she wrote in the post caption. It is a short line, and to me it is worth more than a long analysis. It shows that even the athlete herself lacked the tools to read her own body. To understand why 0.97 and 2.76 seconds deserve serious analysis, they need context. Fresno State has never been an NCAA swimming powerhouse. In program history, Lung is only the second swimmer ever named a CSCAA All-American, after Aliz Kalmar earned Honorable Mention in the 200 breaststroke in 2026. That detail says a great deal: her 58.89 in the 100 breaststroke, nearly a second slower than her seed, still earned Honorable Mention All-America honors. At a small program, a slot like that is not a small thing. Her path was not linear either. She once competed as MacKenzie Miller, swimming three years at BYU before marrying Tanner Lung in 2026, right after her freshman season. For the 2026-2026 season she transferred to Fresno State for her senior year. On September 16 she announced her pregnancy: "we are so excited to welcome our sweet baby girl into our family in December!" So when she stepped onto the blocks at the 2026 NCAA Championships, she had been pregnant for months. That is the central fact, and it is the fact the results sheet never records. One structural detail matters for what follows. At NCAAs, events are split into morning prelims and evening finals. Finals qualification is decided by prelim times, while final placement depends on the evening swim. A swimmer can qualify well in the morning and lose places at night. That structure makes fitness and recovery as important as peak speed. A note on the Trailblazer Invitational is also needed. It is a mid-season meet, usually held late in the year, when swimmers are partly rested. A time like 57.92 in mid-season is a positive signal, but it does not always predict a result at the end-of-season championship. This is a common blind spot in swimming prediction models: a seed time is not the same as real ability on race day. Now, the numbers. In the 100 breaststroke, 0.97 seconds is not small. At the elite level, a hundredth of a second separates places, and nearly a full second is the gap between a finals slot and a mid-pack finish. A 57.92 seed put her among the eight fastest; 58.89 dropped her to 15th. But — and this is the point I want to stress — 15th still sits inside the All-America zone. She lost position, but not the honor. In the 200 breaststroke the drop was larger: 2.76 seconds, from the 10th seed to 29th. This is the kind of fall my models usually attribute to fitness or poor pacing. Over 200 meters, losing nearly three seconds suggests the body could not hold its rhythm through the second half. In breaststroke, rhythm and leg drive decide everything, and both are sensitive to physiological change. Before applying any label, look at her long-course season, where the data is cleaner. At the US Open in December, Lung finished 3rd in the 50 breaststroke (31.00) and 7th in the 100 breaststroke (1:08.70). More telling: in prelims she was faster — 30.73 and 1:08.48 — just missing a spot on the US National Team. This is a pattern I call the prelim paradox: a swimmer goes faster in the morning and slower at night. Normally I would read that as a recovery problem between swims, or as final-night nerves. For Lung, both explanations are plausible, and both may be partly true. But there is a third explanation that sports data rarely carries as a variable: pregnancy. Here I have to state what I repeat in every piece. Numbers have no gender, but the people who read them do. The NCAA results sheet has no column for pregnancy. It has a column for time. Read only the time column and you quietly turn a story of biology into a slump. If my betting data had such a column, everything would change. I would adjust expected times by week of pregnancy, by fatigue, by reduced training load. But we do not have that column. Neither does she. That is why she wrote that she was confused. Look at precedent, and Lung is not alone. Dana Vollmer, an Olympic gold medalist and former world record holder, raced at a Pro Swim Series meet in 2026 while 26 weeks pregnant. Olympedia, the leading database of Olympic history, has found at least 28 women who competed at the Olympic Games while pregnant — though none of them were swimmers. That makes Lung's case rare, but not unique. And here is the most interesting part, data-wise: there is some evidence that women can become better athletes after giving birth. The most famous swimming example is Dara Torres. She gave birth to her daughter in 2026. In 2026, at 41, she won three Olympic silver medals. More recently, Sweden's Sarah Sjostrom won gold in the 50 butterfly and silver in the 50 freestyle at the 2026 European Championships, less than a year after giving birth to her first child. This is where I want to pause and analyze carefully, because it is where data is most easily misread. When an athlete like Torres or Sjostrom returns and wins, the media often frames it as childbirth made her stronger. But correlation is not causation. A woman performing well after birth does not prove that pregnancy improves performance. There are at least three other explanations for the pattern. The first is selection. Athletes who return after birth and attract media attention are the ones who succeed. Those who never return, or return and fail, rarely make the news. This is classic survivorship bias. Count only the winners and you will believe childbirth is a lucky charm. Another explanation lies in baseline. Torres and Sjostrom were already world-class before giving birth. Their post-birth results reflect that baseline, plus training, not necessarily any effect of pregnancy. The remaining explanation is physiological. There is a hypothesis that pregnancy increases blood volume, raises lung capacity, and shifts hormones — factors that could support performance, temporarily or long-term. It is a hypothesis worth studying, but there is not yet enough data to conclude. In other words, Lung's case and Torres's case sit at opposite ends of the same story, and both teach the same lesson: the female athlete's body is a variable that sports data has not yet modeled. Zoom out. As women's professional sport grows globally, renewed attention is going to what happens to female athletes' bodies after pregnancy. Leagues, sponsors, and teams are starting to account for what they once ignored: maternity leave, the right to keep a roster spot, and training conditions during pregnancy. This is a structural shift, not a trend. And in sports betting, where I work, it is an unfilled gap. No book lists odds on the pregnancy variable. Which means that when an athlete like Lung steps onto the blocks, the market prices her as if her body never changed. That is not a flaw in the model. It is a flaw in the input data. Now the counterintuitive part. The laziest way to tell Lung's story is to turn her into a symbol: pregnant athlete still competes at NCAAs. I do not buy it. It sounds inspiring, but it erases the data. What the data actually shows is a swimmer who went slower in both of her signature events, at the biggest meet of the season, while pregnant. She still earned All-America honors, but at the edge of the zone, not the top. Say only that she still competed and you skip the fact that she lost nearly a second in the short race and nearly three in the long one. For a swimmer with a 57.92 seed, nearly a second is a lot. The second blind spot sits on the other side: betting and prediction models. I used to work in that industry. Our models carry no pregnancy variable. They carry recent time, age, injury history, number of races. When a swimmer like Lung steps onto the blocks, the model ranks her by her 57.92 seed and expects a result near that threshold. A 58.89 will be read as below expectation. No link in that chain of calculation knows her body is carrying a child. This is what I learned in Kazan: a model can be arithmetically right and humanly wrong. On the day Germany collapsed in Kazan, every possession metric pointed one way and the result pointed the other. The lesson: numbers describe what happened, not what is happening inside a person. There is a third temptation I want to flag, and it is close to me. As a woman working in a male-dominated sports analytics field, I am tempted to turn this story into a symbol of feminism. But a symbol is not analysis. Use Lung to prove a point about gender and I am doing exactly what I criticize in others: using one individual to illustrate a preconception. I do not trust emotion. I trust a string of data longer than your emotion. But I also know some strings of data have never been recorded, and staying silent about them does not make them disappear. I have to be clear about my own limits here. I do not know how many weeks pregnant Lung was when she swam at the 2026 NCAAs. I do not know whether a doctor cleared her to compete, or under what conditions. I do not know whether she adjusted her training load. I do not know whether she had morning sickness, fatigue, or a changed heart rate. None of that appears in any public dataset. And one thing I know for certain: every pregnancy is different. You cannot use Vollmer's or Sjostrom's case to infer Lung's. Nor can you use Lung's to infer anyone else's. Kazan is the day I learned that a 99% probability can still die on the betting table. This is the swimming version of that lesson: a perfect model of time can still fail to describe a person. So what is the signal for the next round? Next season is the real test. If there is any evidence of pregnancy's effect on performance, it will sit in Lung's own data, after birth, when she returns to the pool. She is expecting her daughter in December. In parallel, watch how leagues and bookmakers adjust. As women's professional sport grows, ignoring the pregnancy variable will become a data gap, not just an ethical oversight. Whoever fills that gap first gains an edge. And the next time you read any results sheet, remember: numbers have no gender, but the people who read them do.