Marrit Steenbergen and the 25.95 Back Half: A European Record Built on the Way Home

In the middle lane, Marrit Steenbergen touched the wall at the 50-metre mark in 24.37 seconds. Slower than her own swim at the 2026 European Championships in Poland, where she went 24.33. No one in the stands shouted for that opening length. Then she turned and swam home, and the scoreboard stopped at 50.32 seconds — a new European Record in the women's 100m freestyle short course (25m), 0.10 seconds faster than her own previous record. What made me pause on this result was not the 50.32 figure. It was where the 0.10 seconds came from. Had she gone out faster, I would call it a raw speed surge. Instead she went out slower, then came home faster. For someone who works with data, those two kinds of improvement mean entirely different things. Before going deeper, the technical context of this event needs to be set correctly. The 100m freestyle short course is a race with three turns. Unlike a 50m long-course pool, which has only one flip, a 25m pool forces the swimmer to handle three wall contacts — and every turn is a chance to gain or lose time. In an event where the gap between the leading swimmers is measured in hundredths of a second, turn quality and finishing technique become the decisive variables. I began covering swimming in 2026 as a swimming reporter for a newspaper, before moving into sports data analysis. That period taught me that swimming is a sport where the smallest detail is decisive. A hundredth of a second is not a figure rounded off for convenience; it is the distance between a record and second place. In short course that is even more true, because every turn is a chance to create or lose such a hundredth. I have followed short-course races across many seasons, and what caught my attention with Steenbergen was her split structure. The results show she divided the 100 metres into two near-even halves: 24.37 going out, 25.95 coming home. That is a positive split — front half faster than back half — but the gap is only 1.58 seconds. For the 100m freestyle short course, such a near-even structure signals calculated energy distribution, not a reckless fly-and-die. Across years of analysis I have learned that short course is a different sport in nature, not merely a shorter version of long course. The presence of three turns turns each race into a sequence of technical events. A swimmer can lose in open water but win at the wall. So when I read a short-course record, I always separate two questions: how fast is the swimmer's pure speed, and how much do turn mechanics contribute to the final figure. The source analysis gives me a partial answer: the back half is faster than the front half, which usually points to strong turn mechanics and closing speed. One point about the data source needs to be stated clearly. The original analysis I am working from carries no named outlet and no specific publication date, only two meet references. That means I mark every figure here as pending cross-verification against the World Aquatics results database and the Omega timing system. I believe in the number, but only after it clears three rounds of checks. A European Record is something that can be looked up, so verification is mandatory, not optional. Following international swimming from Vietnam carries its own challenge: data often arrives late and incomplete. So I have built the habit of logging every data point with its source and confidence level the moment I receive it, so that when cross-checking is needed I do not have to rebuild from scratch. Steenbergen's record is a textbook case of the kind of information that must be handled this way: an impressive figure, but lacking the context to speak for itself. Now to the core: the chain of evidence showing this record was built from the back half. Point one: the 24.37 opening is slower than her own 24.33 in Poland. This is the most important and also the easiest detail to miss. If someone only looks at the final 50.32 against 50.42, they will assume Steenbergen swam faster in every segment. The split data says the opposite. She lost 0.04 seconds on the way out and made up more than that on the way home. Point two: the 25.95 closing 50 — under the 26-second mark. This is the standout technical marker in the entire analysis. In a 25m pool, the final 50 metres includes a turn plus the finishing touch. Swimming the back half under 26 seconds means her turn mechanics are in good shape and the finish was executed cleanly. Sprint swimmers are usually measured by raw speed, but in short course, the turn is where the time hides. Point three: the 0.10-second improvement does not come from raw speed. It comes from back-half speed and finishing execution. That is the pattern of a swimmer who has matured tactically, oriented toward efficiency rather than simply moving the arms faster. To place this record on the right scale, a benchmark is needed. Gretchen Walsh — exactly 0.01 seconds faster than Steenbergen on the all-time list — is the swimmer directly above her. In other words, Steenbergen sits fourth on the list of the fastest performers ever in this event, and she trails the one above her by one hundredth of a second. That gap is small enough to fall inside the noise of measurement. A small GPS deviation was enough to teach me: verification is everything. And here, 0.01 seconds is exactly the kind of error that stops me from ever declaring who is better in absolute terms. The all-time list in the 100m freestyle short course is a very crowded place. When the top four performers in history sit within a few hundredths of a second, everything becomes fragile. A better start, a cleaner turn, a tighter finish — any one of these could reorder the standings. That is why I read Steenbergen's fourth place as a sign of world-class level, but not as proof of absolute superiority. The world record also needs to be placed in context. The original analysis does not identify the world record in this event, and I do not want to fill that blank with guesswork. The reasonable statement is: the world record is presumed faster than 50.31, and the exact figure is pending verification. In my work, an empty data cell is more honest than one filled with a weakly grounded estimate. Data does not tell stories; it records everything so that I can tell them myself. The story the split data tells here is this: Steenbergen won not because she started better, but because she finished better. At this point I have to argue against myself. The first thing to state plainly: this is a short-course record, and short course does not transfer proportionally to long course. In a 25m pool, three turns amplify the value of flip technique and underwater distance. In a 50m pool there is only one turn, and weight shifts more heavily toward pure swim speed. A swimmer can shine in short course through excellent turn mechanics without holding equivalent form in long course. So I do not read 50.32 as proof that Steenbergen will dominate the 100m freestyle long course. That would be an inference beyond the data. Second, the original analysis lacks the baseline metrics to fully confirm the claim of leading technique. There is no reaction time, no turn split data, no stroke rate, no distance-per-stroke figure. We have the two-half split, but we lack the pieces to rebuild the full mechanical picture. That means: the conclusion that the back half is the key is grounded, but the conclusion that Steenbergen owns the best technique is a leap the available data cannot yet support. Third, I want to raise a hypothesis I hold at medium confidence. The improvement focused on back-half speed suggests a deliberate training block targeting turn and finishing conditioning rather than start reactivity. If true, that is a sound training choice for a short-course swimmer. But it is an inference from data, not confirmed information. I leave it there as a hypothesis to verify, not a conclusion. The limitations of the model need to be listed plainly, as I do in every article. The sample here is just two swims by the same athlete about a year apart — far too small to conclude a trend. We are comparing 50.42 in Poland with the current 50.32, but the two swims took place at different meets, possibly at different points in form, under different pool and timing conditions. The 0.10-second gap may reflect real improvement, but it may also contain some environmental error. I do not have enough data to separate the two, and I refuse to pretend otherwise. I also have to recall an old lesson of mine. Croatia 2026 was not a miracle — it was xG written into history. Here too: 50.32 is not a sudden moment of transcendence. It is the result of a specific split structure, which may or may not repeat. Distinguishing what can repeat from what is mere random noise is the whole of my job. The signal I am tracking in the next round is not whether Steenbergen breaks another European Record. The interesting signal is whether the near-even split structure — 24.37 out, 25.95 home — repeats in her following swims. If the sub-26 back half becomes a stable pattern, we are looking at a swimmer who has built a reliable home-stretch engine. If it appears only once, it is a beautiful but isolated data point. And the question I cannot yet answer, and do not want to answer hastily: can this back-half speed transfer to the 50m pool, where only one turn remains and everything depends on pure swim speed? That is the real test. The short-course record is one chapter; it is not the whole book.