A look at how fast a scalp really needs to be

[Shadow Trader]

Scalping is often described in seconds, but the real range of holding times is wide. Some traders hold for 20 seconds. Others hold for 15 minutes. Both call it scalping.

The difference matters. Ultra-short holding times, under a minute, depend heavily on execution speed and spread. A half-pip delay changes the whole result. Traders in this bracket are often playing order flow and liquidity tricks that rely on professional-grade infrastructure.

Holding periods of 3 to 15 minutes depend more on structure, levels and momentum. Spread matters less, since targets are larger. Execution speed is less critical.

For most retail traders, the second bracket is more realistic. It leaves room to think, accepts normal broker latency and reduces the share of costs in the result.

Tracking average holding time in the journal, split by winners and losers, shows which bracket the strategy actually occupies.

What is the typical holding time here, and does it differ between wins and losses?

[Amanhasnoname]

I split mine out last quarter and the gap between winners and losers was bigger than I expected. Winners averaged a little over six minutes, losers about two and a half. That sounds healthy until you notice that a lot of those short losers were me bailing at the first wiggle, not the stop being hit. Some of them would have turned into winners if I'd left them alone.

I've ended up in your second bracket for exactly the reasons you give. My broker's fills aren't fast enough to make sub-minute holds pay, and I spent a few months proving that to myself the expensive way. Once I started trading off fifteen-minute levels with a one-minute trigger, spread stopped being the main thing I worried about.

One addition: track holding time by session too. My London holds are much shorter than my New York ones on the same setup, simply because the moves come faster. Without separating them the average was misleading and I kept tweaking time stops for the wrong reasons.