LineBreak historical tick and MRD based chart discrepancies

[nicknamed]

Hi,
I got a stupid question (as usually) I need to clarify but first I need to write an introduction:

I am sorry, for the length of this question (though, I got a multibroker license ), but I need to solve this problem as I see that tick based charts are different while they should not be and I need to test on data I will be using for real trading, right? I have read the part of help about backtest discrepancies now and before, I have done many tests, so I am aware of many discrepancies, but this is the newest one I was not aware of as this is the first time I test on tick or volume (that come from downloaded tick data) based charts.

I am running strategy on MRD (market replay data) and the chart looks different than from chart drawn from tick data I downloaded from CQG, I do not filter ticks and I guess CQG as well as you (the MRD source) also do not filter ticks...and I guess that both of you get the data directly from the market so the chart should be the same. But it is not if I compare LineBreak chart based on tick data downloaded from CQG and from MRD, those charts are based on Volume, not time...however those charts are extremely similar, but sometimes lows and highs differ even by 10 ticks and by a few seconds or even minutes...if I put them next to each other, the proportionality of bars is about the same...I compared the newest CL 5-11 contract. I consider futures markets as centralized one(s), unlike currency market, so I can trust the volumes and tick data as those should be the same, if obtained from data supplier and not brokerage house...unless I encounter strange things that even if I select last price data for chart in NT7 then TWS of IB supplies median price and not last price...so the chart looks different than the one based on IQFeed data even if that comes from centralized markets as AMEX/NYSE are - but that is something different as broker can alter market data the way IB do by supplying median price instead of last price data to NT7....but here with tick by tick futures it should not be altered in any way, when comapred CQG as data source and MRD as well, right?

I have also noticed, that If I run MRD test, it will save tick data into tick directory to let me use those tick data for backtest.

So here comes the question: If I can create tick data from your MRD source and backtest on those, am I going to get the same results from MRD test if I select calculate on bar close? The chart will be based on Line Break / Volume.

Second question: Do you think that If I used CQG as my data provider for tick data, it would make the same trades on their data as on backtest on their data (considering zero latency delay and calculate on bar close = true)?

Thank you.
N.

[NinjaTrader_RyanM1]

Hello N,

Thank you for your post. Yes, there can definitely be differences in data across data providers, and strategies function differently when running historical or real time. Unfortunately we can't say whether your test will produce identical results. If you want your market replay files to match the incoming real time stream by CQG, then turn on the data recorder through Tools > Options > Data tab > Check Record for market replay.

[nicknamed]

Well Ryan,
that does not answers my first question, I mean if I get the volume and tick by tick data from MRD, then will I receive the same bars (NOT trades) when the chart will be drawn on backtest and on MRD?

Please see my second question I just added.
N.

[nicknamed]

One more thing (which adds to the severity of the question), I just check on tick chart, and they are the same, NT7 jsut draws the bars differently for MRD replay and for backtest on tick data (Volume Line Break chart).

[NinjaTrader_RyanM1]

No, we do not expect trades in a backtest to match real time results. We also don't expect data from varying sources to be the same. If you have run into any issues, can you please send a couple screenshots showing and steps needed to see.

[nicknamed]

Ryan, ok, I will simplify it - how differently are bars build with tick data from the same provider in Line break chart on backtest vs. realtime?...if you do not expect backtest results to comply with real time results?
Consider ideal trading conditions, like zero latency, infinite liquidity, maybe even zero spread and other aspects, then the bars (I repeat Bars, NOT trades) should look the same, right? I just need to know if I supply the same tick data from backtest to MRD (theoretically, I know that is not possible), then I should get the same bars, right? If this is right then that is what I need (to know), I know market conditions are far from static historical even tick data, but I need this assumption to be true to continue with my mathematical analysis. Otherwise I guess the charts are build different way under backtest even from tick data and in MRD/realtime and that is something I cannot influence with my system, data feed, or anything else...just to use maybe different program. Please tell me if my assumption is correct.
Thank you
N.

[NinjaTrader_RyanM1]

Market replay data is recorded from live incoming data, so comparing it against it against any historical source may not be the same. Your source for real time data is also different than the NinjaTrader-provided market replay files.

If you want to view market replay data and want it as similar as possible to data available in a backtest:
These options are set from Tools > Options > Data tab:
Turn on the market replay recorder.
Turn on Save chart data as historical.
Turn off historical data from your provider: This is set in the tools > Account connections for the specific connection.

The same data is recorded for market replay, and saved historically in a backtest. Yes, we can reasonably expect that the bars will be similar with this setup.

[nicknamed]

I am already recording those data from CQG, so tomorrow I will see, I hope those data will be at least 99,9% the same...so thank you for that "Yes".
N.