Brokers do notice scalpers, and the response is rarely a ban

[Fairman]

Brokers do notice scalpers, and the response is rarely a ban

Accounts rarely get closed for scalping. What happens instead is softer. A little extra delay on fills. A slightly wider spread on the account. A maximum-trades-per-day rule that appears in the terms. A shift from instant execution to a mode that enjoys last look.

None of this is announced. It shows up as drift. Backtests that looked great on demo start to disappoint on live, and the explanation always seems to be "market conditions".

Some patterns that point toward throttling:

- Demo fills noticeably better than live on identical logic
- Execution speed degrading after a run of profitable weeks
- Orders held for a second or two before filling, only on certain entry types

The way to check is a control: a small second account at a different broker, same strategy, same hours. If the second account holds up and the first fades, the difference is the broker.

It is uncomfortable to consider, but the alternative is spending months tuning a strategy that was never the problem.

Has anyone seen execution quality change after a streak of profitable trades, or was it all coincidence?

[Fairman]

Brokers monitor client flow, and scalping accounts with very short holding times, many trades and consistent profits stand out. How a broker responds depends largely on its business model. A broker that hedges client trades with liquidity providers mainly cares whether those providers are happy to take the flow, and providers may complain if a client's trades are consistently followed by adverse moves for them. A broker that internalises flow, taking the other side itself, has a direct interest in whether clients win or lose. In both cases, a profitable scalper may face changes such as wider spreads, slower execution or a request to move to a different account type.

[Amanhasnoname]

I saw the drift version of this at a broker I used for about eight months. The first six weeks were great, with fills almost exactly at the quote and spreads that matched the advertised numbers. As my profits built, I noticed orders taking about half a second longer and a pip or so more slippage on entries, but only on my fast scalps.

I did not trust my memory, so I opened a small account at another broker and ran the same logic during the same hours for three weeks. The second account averaged roughly 0.4 pips better per trade, and its demo and live results lined up much more closely.

The most telling check was comparing fill times by entry type. Market orders at the first account were delayed far more than pending orders, which suggested a last look style execution. I moved most of my volume away and kept the old account small. Run the control account before you decide anything, because the numbers are hard to argue with.