Why different results on different replay results?

[timmbbo]

Folks, it's unclear to me why I get different strategy results from different replays setups.

1) Replay at 1000X, Calculate = OnPriceChange, Result = ($1112.50)
2) Replay at 500X. Calculate = OnPriceChange, Result = ($1287.50)
3) Replay at 1000X, Calculate = OnEachTick, Result = $13,675 (I wish!)

This is running the ES replay for 3/25. I ran this same strategy today on the live market data, and the result was $175.

Therefor, it's unclear to my why different replay formats are giving different results. Most importantly, how do I set up a replay to make the same results I get when the strategy is running on live market data? Thanks for your help,

[NinjaTrader_ChelseaB]

Hello timmbbo,

The Calculate setting will change when the logic is called, when, and how often, orders are placed. Orders placed at different times will have different fill prices. More or less orders would results in different performance.

Are there any values that can change such as a random number generator or logic that can change when orders are placed?

Are you able to reproduce the behavior using the SampleMACrossover included with NinjaTrader?


Regarding the speed, I would not expect this to cause differences. Below is a link to a video of running a test on this twice using ES 06-21 on a 1 minute chart, once at 500x and once at 1000x This is using the latest version of NinjaTrader 8, 8.0.24.1. The profit and number of trades is the same.

https://drive.google.com/file/d/1fWsLHpK-_Mi5Mf83evYdEHOjiRBJ6Mgh/view?usp=drivesdk

[timmbbo]

> The Calculate setting will change when the logic is called, when, and how often, orders are placed.

OK, then how do I set up a simulated backtest which accurately mimics live data? For example, if a strategy gains 5 points during a trading day, then I want my simulated backtest to show that same result. Thanks,

[NinjaTrader_ChelseaB]

Hello timmbbo,

1 tick Intra-bar granularity and Tick Replay will allow the strategy to processes as similar as possible to real-time.

https://ninjatrader.com/support/forum/forum/ninjatrader-8/strategy-development/94098-isfirsttickofbar-vs-onbarclose-for-backtest-live?p=773377#post773377

https://ninjatrader.com/support/forum/forum/ninjatrader-8/strategy-development/100192-comparing-real-time-historical-and-replay-performance#post100192