Question about slippage/sim strategies

[crimsonite]

Thanks allot for the input and pointing me to the links... I'll dive into this info when I get home.

I'll undoubtedly have something to ask later, today or tomorrow, so... until next time!

[crimsonite]

How much slippage should I expect on NQ in the "live automated strategy" in the strategy tab (during real time, not market replay) with market orders?

Also, just to clarify, should I expect the exact same results in market replay as live automated trading, given the strategy is the identical one with the same input parameters?

Your continued support is extraordinary... thank you for the hand holding in Ninja ...

[NinjaTrader_RyanM1]

Thanks for your kind words, Crimsonite.

Unfortunately we do not have an expected slippage number in the simulator. You can expect buy market orders to fill at the best ask, and sell market orders to fill at the best bid.

We provide no guarantee on the exact same results since there are so many factors to consider. We expect that the strategy behaves as it is designed based on its settings, data, and the exact code used.

To help line things up you should record your own market replay stream from Tools > Options > Data tab. Then you will have the same data coming in live as you play back in market replay. The strategy should behave the same mechanically in market replay compared to live. If you are able to line up every potential factor identically between your two test runs, then they should be similar or same.

Ideally you should be able to follow your strategy behavior through Print() statements and TraceOrders output, so that you can identify how it should do things within the context of one particular run. This will provide you more insight on your strategy behavior than attempting to chase identical results between multiple tests.

[crimsonite]

Thanks for the insights. I'm not too worried about it being guaranteed, I'm just trying to pinpoint the problem with my system. I'm 70% certain I made a code error (1 tick stop loss!) on my strategy and that was the cause for it not functioning properly live.

I'll take your suggestions and do (or redo) them again, thanks.

Is there a 'rough estimate' of what the slippage you would expect on NQ? The spread is usually 2-3 ticks all the time, so would it be typically 2 ticks of slippage based on the last price? 5 ticks? My system is good for about 6-7 ticks, so I want to be sure that my original conclusion that it was blowing up due to market orders (and not my new one that I made a code error with stop loss) wasn't completely true. Either way, I can always get it to print on the screen and trade it live with limit orders discretionary style.

Thanks for your kind words, Crimsonite.

Unfortunately we do not have an expected slippage number in the simulator. You can expect buy market orders to fill at the best ask, and sell market orders to fill at the best bid.

We provide no guarantee on the exact same results since there are so many factors to consider. We expect that the strategy behaves as it is designed based on its settings, data, and the exact code used.

To help line things up you should record your own market replay stream from Tools > Options > Data tab. Then you will have the same data coming in live as you play back in market replay. The strategy should behave the same mechanically in market replay compared to live. If you are able to line up every potential factor identically between your two test runs, then they should be similar or same.

Ideally you should be able to follow your strategy behavior through Print() statements and TraceOrders output, so that you can identify how it should do things within the context of one particular run. This will provide you more insight on your strategy behavior than attempting to chase identical results between multiple tests.

[NinjaTrader_RyanM1]

Yes, a narrow stop loss like 1 tick could cause problems for a strategy running live, as there will be high chance of rejected orders.

It's difficult to provide you a number of ticks expected slippage. We are the trading platform only, so once you submit your order its out of NTs hands, and will be handled by constantly changing market dynamics. We hope to estimate live trading conditions as best we can with the simulator but don't want it construed that expected slippage in simulator would be identical to expected slippage in a live account.

If you are interested in seeing how the fill price on the simulator are handled, you don't need an automated strategy for this. Just practice submitting various orders on the Superdom. You can place them near the market price for a fast trigger, or even use the simulated data feed to control the market direction towards your working orders.

[crimsonite]

Is it not common practice to run NInja automated strategy on a live account? IOrders could help solve this, or my broker could guarantee a market order to be filled within a certain number of ticks of the last price or not filled at all.

I would probably trade live discretionary following signals from the strategy, but I was also considering running Ninja in the cloud with the automated strategy tab running a live account. I was also planning on telling my broker to set a daily loss limit of 4% or more, in case something went wrong with NT7. Is this a bad idea? Would you trust the integrity of NT to run a live account in automated mode?

[NinjaTrader_RyanM1]

If your goal is to identify expected slippage in the simulator, you don't need an automated strategy. You can just place stop orders and evaluate stop price compared to fill price. I mention this as a potentially simpler approach but you're welcome to use strategies to evaluate as well.

Unfortunately we are not in a position to offer trading advice here, but hopefully other members viewing the thread can share their feedback on setups and experience.