The hot hand in scalping: real streaks versus lucky clustering
Traders talk about being in the zone, a stretch when everything works. Some of it may be real: certain market conditions suit the strategy, and in those periods results cluster. Some is just randomness mistaken for skill.
The distinction matters because the response differs. If streaks reflect favorable conditions, then identifying the conditions allows more deliberate trading. If they reflect luck, increasing size during a streak merely raises exposure to its end.
A test: tag each trade with market conditions such as volatility level, session, and trend state. Check whether good streaks coincide with specific conditions. If so, the edge is conditional, and the filter can be formalized.
If streaks appear random with respect to conditions, treat them as noise. Size should stay constant.
Also compare the length of streaks with what chance would produce, using a simple simulation.
Feeling hot is not evidence. The data can say whether the feeling reflects anything real.
Do winning streaks in the log line up with identifiable market conditions?
The hot hand in scalping: real streaks versus lucky clustering
-
Amanhasnoname
- Posts: 2060
- Joined: Fri Oct 30, 2026 9:45 am
Re: The hot hand in scalping: real streaks versus lucky clustering
Feeling hot used to make me add size, which is how I found out the streaks were not what I thought. I tagged four hundred and twelve trades over nine weeks with session, a simple trend flag on the fifteen-minute chart, and whether realized volatility sat above its twenty-day median. There were six runs of seven or more winners. Four of those six landed on London mornings that were both lively and trending.
A shuffle of the same win-loss sequence produced a seven-winner run about once every seventy trades, so six such runs in four hundred and twelve trades was close to chance. What was not chance was the overlap with those two conditions. After I required both filters, the next ninety-four trades won sixty-one percent of the time against fifty-one percent outside them.
I no longer raise size during a streak. Size stays fixed, and the only change is whether the conditions that actually clustered the winners are present. Feeling in the zone still happens; I just do not treat it as information on its own.
A shuffle of the same win-loss sequence produced a seven-winner run about once every seventy trades, so six such runs in four hundred and twelve trades was close to chance. What was not chance was the overlap with those two conditions. After I required both filters, the next ninety-four trades won sixty-one percent of the time against fifty-one percent outside them.
I no longer raise size during a streak. Size stays fixed, and the only change is whether the conditions that actually clustered the winners are present. Feeling in the zone still happens; I just do not treat it as information on its own.
-
FXS London Scalper
- Posts: 976
- Joined: Sun Oct 18, 2026 12:25 pm
Re: The hot hand in scalping: real streaks versus lucky clustering
Lucky clustering and real streaks look the same over short periods. Comparing the number of streaks in the trade log with what random sequences of the same win rate produce is a simple test.
FXS Official Team | London Session Desk