Almost every player scores worse deep into a session. Some of that is tiredness, some of it is tilt, and a surprising amount is simply who you end up playing.
Look at a large game history split by position in the session and a pattern usually appears: the first few games score better than the last few. It is tempting to call that tilt and move on. The real picture has at least three causes, and they call for different responses.
This is the one almost nobody accounts for. If you win your early games your rating goes up, and matchmaking pairs you with stronger opponents. By game fifteen of a good session you are playing better players than you were at game one, so your win rate falls even if you are playing exactly as well.
The effect is mechanical rather than psychological, and it means a modest late-session drop can be entirely normal. Before concluding anything about fatigue, check whether the opponents' average rating rose across the session. If it did, part of your drop is already explained.
Real tiredness shows up differently from stronger opposition. It looks like more time trouble, more one-move oversights, and a fall in accuracy rather than just a fall in results. If your late-session accuracy is several points below your early-session accuracy against similar opposition, that is fatigue rather than matchmaking.
Fatigue also arrives earlier in faster formats. An hour of bullet asks for far more decisions than an hour of rapid, and the quality curve bends sooner.
Tilt is not a synonym for playing badly. It is the particular spiral where a loss changes how you play the next game: faster, more aggressively, less willing to accept a draw or a quiet position, and often against a rematch opponent who just beat you. It shows up as a cluster of quick losses in a short window rather than a gentle decline.
The signature is time, not just results. If your losing games late in a session are also your shortest games, you are probably not tired, you are annoyed.
You need three numbers, all easy to get from a game list:
Beware small samples. A twenty-game difference proves nothing; several hundred games in each bucket is where the comparison starts to mean something. Most "I always tilt after two losses" beliefs do not survive counting.
If the data says your late games are worse, the fix is a rule you set in advance, because the decision to stop is exactly the decision a tilted player cannot make well. Useful ones:
It helps to know the size of the problem before fixing it. Take the games played late in your sessions, add up the rating you actually gained or lost across them, and compare it with the same figure for your early games. That number is the real cost, and it is often smaller than people fear, which is itself useful: if your late sessions cost you twenty points across a year, a strict stopping rule is not where your improvement will come from.
Statmate groups your games into sessions automatically and shows win rate by position in the session and by session length, along with your early against late comparison. The Insights tab reports it with the sample sizes attached, so a drop based on thirty games is not presented as a finding.
The rating leak detector goes further and measures the rating you actually won or lost in those late games, rather than estimating it, so you can see what a stopping rule would really be worth.
Late-session drops are normal and only partly about you. Check whether your opponents got stronger, use accuracy to separate fatigue from matchmaking, treat tilt as the specific fast-losses pattern it is, and if the numbers justify it, set a stopping rule before you start rather than during the session.