What time of day do big movers peak? 1,380 measured

Ask a room of day traders when a runner tops out and most will say the first half hour. It is the oldest piece of intraday folklore there is, and it has a mechanism behind it: the opening auction clears the overnight imbalance, volume is at its peak, spreads are at their widest, and the move that brought the name onto the screen is usually already underway by 09:31. The folklore is wrong about the peak, and it is wrong in an interesting direction.

1. The high is a U, and the tall side is the close

We took 46 sessions across nine months and, on each, the 30 most liquid names whose regular-session high stood at least 10% above the open. Selecting on the high rather than the close matters: rank movers by where they finished and you have already selected for names that held their gains, which puts the high late by construction. Selecting on the high itself is neutral as to when it happened. 1,380 symbol-days. Here is the half hour in which the high printed. The opening half hour is the second most likely time for the high. The most likely is the last one, and not by a little: 21.1% against 13.9%. Midday is where highs go to not happen, bottoming at 4.1%. Median time from the open to the high is 213 minutes, which puts the typical peak at 13:03. The quartiles are 67 and 343 minutes, so the spread is enormous and any single answer to the question is the wrong shape.

2. The low, by contrast, is not ambiguous at all

1,093 of 1,380. Median time to the low is 1 minute after the open. That asymmetry is the real structure. These names open at or near their worst price of the day and then spend the session working away from it. The opening half hour is where the low is decided; the high is decided anywhere.

3. Most of the range arrives early, the high does not

Both of those can be true, and they are. Half the day's range is built in the first half hour. But the range expanding and the high printing are different events: the range widens fast because the low is set immediately, then the top edge creeps upward for hours. By the time the last 17.5% of the range is being added, it is late in the day and it is being added at the top. This is why the folklore survives. Traders remember the first thirty minutes as the violent part, and they are right about that. They then assume the violent part contains the extreme, and it usually does not.

4. The bigger the move, the earlier it tops

A name that doubles reaches its high in 133 minutes; one that adds a quarter or less takes 241. That inverts the natural reading of section 1: the late highs are not the big runners finally topping out, they are the modest ones grinding up all day. The violent names do behave the way the folklore says. There are just far fewer of them, and the average is made by the other 989.

5. Almost nobody closes on the high

So the 21.1% of highs printing in the final half hour is not mostly a close at the very top. Some of it is, and 17.5% is not nothing. But the median name gives back 4.16% from its high before the bell, which means a late high is frequently a late spike that fades into the close rather than a clean trend into it. What this does and does not settle It settles that the claim about runners topping in the first thirty minutes is false in general and true of the largest moves. Those are different claims and they get conflated. It does not tell you when to sell anything. Every number here is measured after the fact, on the completed session, which is the only way to locate a high and also the reason the result cannot be traded directly. Knowing the median peak is at 13:03 tells you nothing about which side of the median today is on, and section 1's quartiles, 67 and 343 minutes, are the honest width of that ignorance. Limitations The sample is the 30 most liquid qualifying names per session, taken every fourth session, so it is weighted toward the names a screener would actually surface rather than the full tail. Of 555,276 symbol-days screened, 6,353 qualified and 1,380 were sampled. No symbol-day was lost to missing or unusable minute data. Highs and lows are taken from one-minute bars, so the timestamp is the minute containing the extreme, not the second. Regular hours only: a higher print before 09:30 or after 16:00 is outside this measurement entirely, which matters for names whose real peak was in the pre-market. The 10% threshold is measured from the open, not the previous close, which excludes names that gapped up overnight and traded flat afterwards. That is deliberate, since a gap is not an intraday run, but it does mean this is a study of names that moved during the session rather than of every name a gainers list would show.