Hi, All,

In a system with 1 asset and 1 algo, and which occasionally closes trades with Take Profit limit orders, setting TICKS makes no difference to the result.

However, setting FAST in addition to TICKS - does.
(results are better).

So, what does FAST actually do - in this case of 1 asset/1 algo system?

Comparing trade exit prices between these 2 scenaria with prices in the database,I see that:
- this only affects Trade exits, when exitLong/Short() is used (stops in this case)
- FAST exits trades at the OPEN of the M1-tick of the exit signal.
e.g. with BarPeriod=5 and exit generated at 7:16, FAST will exit at the Open of 7:16 M1-tick, i.e. at 7:15. - which guarantees peeking.

IS this really how FAST supposed to work?
What's the use case of this?
(I tested 2 assets/2algos and see neither memory savings nor calculation speed up)

The difference is that intrabar management is per-trade in FAST mode, but per-tick without FAST mode.

In FAST mode the time moves from bar begin to bar end for any trade separately, and then jumps back. In normal mode, the time moves continuously.

FAST mode can strongly improve backtest speed with many open trades, but is normally not recommended. It will probably be removed in a future version, because it adds a lot complexity to the engine.

Thanks for clarifying.
So, the normal per-tick mode simulates portfolio level interactions realistically, and FAST is a shortcut, useful during development. Clear.

But why such fill prices?
Time supposedly jumps back AFTER processing a trade, not before...