Your scalping backtest is mostly an assumption
Scalping is where TradingView's backtest is least able to tell you what you want to know, and the reason is specific rather than general.
The emulator has to guess, and on a scalp it guesses constantly
When a bar's range contains both your stop and your target, the backtest has to decide which was reached first. It does so by inferring the path price took through the bar from where the open sits relative to the high and the low [2]:
If the opening price of a bar is closer to the high than it is to the low, the emulator assumes that the market price moved in this order: open → high → low → close. If the opening price of a bar is closer to the low than it is to the high, the emulator assumes that the market price moved in this order: open → low → high → close. [2]
Now notice what scalping does. A tight stop and a tight target on a short timeframe means both levels usually sit inside one bar's range. So the case the emulator resolves by assumption is not an edge case in a scalping strategy — it is the normal case.
You can measure your own exposure to this. Open the List of Trades, and for each exit check whether that bar's range contained both your levels. On a scalping strategy the fraction is often most of them. That fraction is how much of your equity curve is the rule above rather than the market.
This is also exactly where MetaTrader differs: its tester can replay real broker-accumulated ticks with no simulation [3], so it does not have to infer the path. For a scalping idea specifically, that is not a minor difference — the engine comparison sets out why the two cannot agree.
Two platform limits that bind here first
The order cap. A script can place a maximum of 9,000 orders in a backtest, rising to 1,000,000 with Deep Backtesting [1]. A scalping strategy on a short timeframe over a long range reaches 9,000 and simply stops adding trades — the curve flattens and the statistics stay plausible. Always check the trade count against the cap before concluding the strategy stopped working.
The execution-time limit. 20 seconds for basic accounts, 40 for others [1]. Short timeframes mean more bars per unit of history, so this binds sooner here than on a daily strategy.
The arithmetic that decides scalping
Costs are not a detail at this horizon, they are the strategy. If your target is a small number of ticks and your commission plus slippage is a meaningful fraction of it, you need a large directional edge just to break even.
So run the doubling test before anything else: double commission and slippage and re-run. On a scalping strategy this frequently removes the entire edge, and finding that out in an afternoon is the best possible outcome — the cost sensitivity test.
If you still want to build it
| Field | Value | Note |
|---|---|---|
| Timeframe | 1–15 minutes | more bars, so more exposure to every limit above |
| Entry | one condition, not three | fewer parameters at the horizon with the least data per trade |
| Stop and target | separated enough not to share a bar | reduces how much the emulator decides |
| Costs | commission and slippage set honestly | the single most important field |
| Session filter | yes | spreads widen outside liquid hours |
The third row is the actionable one. Widening the gap between stop and target is not about better risk-reward — it is about leaving the emulator less to guess about, which makes your backtest mean more.
Tactix AI — Studio vs Guide
Tactix AI is AlfaTactix’s product assistant brand (open Tactix AI).
- Tactix Studio turns a one-sentence strategy description into a draft across Timeframe, Signals, Filters, and Risk in the visual Strategy Builder. You review and edit every field before Code Generator writes MQL5 or Pine Script.
- Tactix Guide explains the step you are on — what to fill, what a control means, or how to phrase a rule — without dumping untested source code.
That is form-first automation: the LLM never replaces Code Generator, and you keep plan limits and real-time validation.