Backtest on MNQ is exceptionally slow compared to others?

[Skifree]

I just ran a 6 month backtest on MGC with consistent pacing (printouts every week). Trying this on MNQ slows to a crawl. I expect MNQ will have more trades, but when its all grouped into bar series (execute on bar close) why is it taking forever? Is there an optimization to make?

[NinjaTrader_Jesse]

Hello Skifree,

The backtest could differ for a variety of reasons depending on what your strategy is doing. If there was a substantial amount of trades difference that could be one reason if there was more to process.

As a test try the SampleMaCrossOver using the same time range and whatever other settings you may have edited. Do the tests still take longer or is it similar?

[Skifree]

its really difficult to compare something featherweight to something complex. But the point remains that with all the same parameters, MNQ is incredibly slow. I know it has more volume than many other signals, how does that impact something like a tick series (20,000 ticks per bar)? Is there a backend process that might be doing aggregation of ticks and i'm seeing the impact?

[NinjaTrader_Jesse]

Hello Skifree,

As I have no details on what your strategy does testing something which we have the source for and know how it works would be the best way to check if something is happening with the platform. That eliminates any custom code that may be not working efficiently.

In regard to volume, if you are using a tick based chart and an instrument has much more volume/ticks than another that will result in a longer test. The more data that needs processed will result in a longer test. That will create more bars for a period of time by having more ticks so there is more for the script to process compared to the other instrument with less ticks/bars.