Market replay data ....

[maxpi]

I've been recording data for the YM for a few days, running a chart or two all day with IB as the data provider. Today I'm running replay at 50x speed as the data source and I notice that the data does not look right. The volume for the first bar of Friday's regular session is about a third of what it is if I connect OpenTick and pull up the same chart and the prices seem to be jumping several points every tick. Does this have something to do with the way IB sends data? If I connected OpenTick along with IB would I record a better set of data?

[NinjaTrader_JoshP]

When viewing replays at 50x you might see some stutters and race conditions in price fluctuations simply because of the sheer amount of information you are trying to display at high speeds.

In regards to the data itself, different providers will have differences in their data. Running both IB and OpenTick at the same time will not affect your recordings. It will record from your primary connection.

[maxpi]

I'm more interested in doing forward tests of strategies on speeded up data after the market than anything. I can play with the speed settings to see what my computer can keep up with.

If I set up an Opentick connection with IB as secondary source will I be able to trade live with IB during the day and have an Opentick recording? The Opentick data might be better since they don't aggregate it...

[NinjaTrader_JoshP]

Even if your computer can handle the sped up speeds you will notice differences in fills due to race conditions. You can try it for yourself. Find a spot where you strategy enters a market position. Go through that time at 1x and note the fill price then go through it again at 50x and sometimes it will fill differently.

I believe you should be able to do that OpenTick/IB combo.

[maxpi]

Thanks much Josh

[zoltran]

If I set up an Opentick connection with IB as secondary source will I be able to trade live with IB during the day and have an Opentick recording? The Opentick data might be better since they don't aggregate it...

My experience is .. it depends which feed you connect to 1st, and if you have paid for R/Time exchange fees at OpenTick.

A) IF you connect NT to IB 1st, . then IB is your R/T feed, and this is what gets recorded for replay. Opentick is only used for backfill data.

B) IF you connect Opentick 1st and then to IB... IB is only used for Brokerage/trading, not quotes.
Opentick data is then recorded for Replay.

Option B requires you to have paid the appropriate real-time exchange fees's on Opentick. (money goes to the exchange, not Opentick)

Option A does not require this, and only use Opentick for backfill

[maxpi]

Nice thing about opentick, YM data is free... I set up the opentick connection to have IB as backup, maybe that will be good, if not I will have to connect them in order...

[zoltran]

Yup

Just to be clear .. it's really ECBOT that make YM and the other 'Y' mini's free.

YG for example.

[maxpi]

Uhh huh, and they are being bought by the CBOE so the free part could likely change at some point.

[maxpi]

Does the market replay aggregate data and send it to the screen? That is what it appears to do... I was thinking I could forward test strategies on replay data at high speeds in lieu of backtesting, maybe I can do that but not at such high speeds... and maybe what the stragegy does and what the screen does is two different things....

[NinjaTrader_Ray]

Its is like any real-time feed and can be forwarded tested. The only limitation is that the simulation engine runs on real-time and not accelerated time, therefore if you run at 50x, the simulator is still running at 1x and your fill could be based on data in the future. You can eliminate some of this delay by setting some of the time delay parameters in the Tools > Options > Simulator to values of zero.

[maxpi]

Ten times speed would be ok. A day in 40 minutes with testing to the tick level is fine, I'm looking to map out a general idea that is backtested fairly well and see it on the screen. There's just something about seeing the product's first iteration working in a real environment, it always clears the way for more questions

[koalboy]

Testing Results Market Replay Speed

I'm not able to code a tick-based-strategy which could generate "close-to-reality" backtesting performance in the strategy analyzer. So my only alternative is to test my strategies on market replay what only makes sense if I can take advantage of the high-speed-function (up to 500x - fantastic!!!!)

But unfortunately I read that the filling-simulator cannot cope with highspeed of the replay. So I've tested the market replay function due to its results on different speed levels (500x, 50x, 1x).

I took Zenfire-recorded data of the FDAX of a whole week. The results were better than expected:

There was no (!!) difference in filling-prices (default-fillingtype) over about 250 Orders at all.

Im on Version NT 6.5.0.8.

So I'd like to know, if this speed-restriction-problem of the simulator is already solved?

thx

[NinjaTrader_JoshP]

Yes, it has been resolved.

[koalboy]

@ Josh

If I run e.g. 10 strategies on the same instrument at the same time in the market replay environment @ speed 500, is there any possibilty, that they affect each other in a way?

When I was doing that yesterday, some of the strategies stopped during the session without any notice (no "order rejection" or "termination of the strategy" window opened; no message in the log).

thx