Computer / Strategy Performance in Live Sim

[Learning1]

Hi There,

I am seeking to understand a few things around how the NT simulator works when it is in Live Sim mode with "Enforce Immediate Fills" checked. My current understanding (which may be wrong) is that the simulator always (when enforce immediate fills is checked in the options) provides a fill in the event that a limit order is touched by the price action.

I guess it would be wise to expand a bit on my understanding of the delay settings. Am I right in assuming that the above is true only if all the delay settings are at 0? Meaning that if there is an exchange delay of 150 and a comms delay of 50, the Simulator will only provide a fill if a limit order is at a price point 200ms before the price reaches that price point. Is that the case?

If so ... here is the puzzle I am trying to understand.

I am conducting a NinjaTrader6.5 (Rev3) Live-SIM test running three nearly identical computers with Windows2003 Server, each running the exact same strategy, connected to the same ZenFire feed for live data.

Are there any reasons that I shouldn't expect close to the same trading behavior and realized net profit between these three machines? Note that the following settings are being used in LiveSIM:
-Enforce Partial Fills IS selected
-Enforce Immediate Fills IS selected
-Delay Comm = 30mS
-Delay Exchange = 150mS
-Wait for data = 0 Sec
- All 3 machines are running the exact same strategy

I have run this test for the past 4 trading days.
The results between machines have varied significantly and I am trying to undertand the possible reasons why. All the machines have good processor speed and over 2 GB of RAM, and are located in the same place with the same Network connectivity. However, from time to time, I am seeing differences between them on Disk R/W performance. When real-time market data is being received by NinjaTrader 6.5(Rev3) is this data being written to disk before being interpreted and made available to the Strategy? If so, variances in Disk performance might be contributing to the performance differences between machines. Is this likely or are there other factors at play that I should be considering?

Thanks for your help as I am puzzling through this challenge.

Gratefully,

Learning1

[NinjaTrader_Dierk]

We can't expose all details on how the simulator works but here are some hints:
- having the delay(s) at 0 means than order fill events are triggered right as an order would get filled without any delay
- running the same strategy against the sim feed on multiple machines would yield different results, since the sim feed is random
- running the same strategy against the same replay data (NOT sim feed) with delay(s) = 0 would always yields the same results, even on different machines

[teecee001]

i am so confused - where does a newbie start

[NinjaTrader_Dierk]

Please post issues not related to the original thread in a new thread.

There is a wealth of info available, e.g.:
- free webinars: see link in my signature
- docs and videos: http://www.ninjatrader-support.com/H...helpguide.html

[Learning1]

teecee,

I would echo Dierk's comments.

NinjaTrader has done an excellent job of making a bunch of learning available. I was a newbie 12 months ago. Where to start depends on the type of trading that suits you. For me, (initially) learning how to build a strategy using the Strategy Wizard was where progress really started. There are video tutorials available on this as well as occasional courses.

Good Luck with your journey,

Learning1