Invalid simulation of fills - occasionally getting 1 tick better profit

[misova]

I found problem = getting invalid simulation of fills of limit orders.

I am testing simple strategy with several trades I get 1 tick better profits, then would be possible in the market.
All trades are the same, but somehow, filling is not consistent and a few trades have 1 tick bigger profit.

I debugged and found out, that there is no problem with my order and prices.
They are correct. For a few trades, filling engine evaluates,
that profit will be filled for 1 tick better price, than was set my profit order.

That is problem, because it results in unrealistic better profits.

I cannot explain, how I could get bigger profit with limit order, when it is clear (when looking at bars), that is should be filled at the exact limit price - not 1 tick better.

It is consitently reproducible.
The only problem is, that I cannot debug it more, because problem lies in the
FillType, that is not accesible and cannot debug it.

I can send you my working strategy, with exact steps, how to reproduce it
and spot the problem easily:

• you will spot, that profit order was send for exact limit price
• then filled and see, that execution was 1 tick better, then limit price in the order
• no magic - easy to reproduce everytime

I offer quick Teamviewer session, where I can demonstrate quickly in 2 minutes,
so it does not take your precious time and then you can find the fix / solution...

If you leave me contact / email, I will write you back and we can arrange very efficient quick Teamviewer session, to demonstrate and confirm real problem, which you can fix then.

[misova]

Another problem with filling engine.
I see, is that sometimes limit orders are filled on touch,
despite Fill Limit on touch is not checked.




Settings - Fill Limit on touch is really disabled:

[NinjaTrader_ZacharyG]

Hello misova,

Thank you for writing in.

Can you please provide a sample strategy script that demonstrates this behavior that you are seeing?

Additionally, please provide the steps you are using to reproduce this behavior so I may take a look on my end.

If you would rather not post the strategy on the forum, please email us with your strategy attached at platformsupport [at] ninjatrader [dot] com and reference both my name and a link to this thread in the body of the message.

[misova]

Hi,

I have prepared very precise scenario, that shows you the invalid fill.
The only thing you need to do is to go over filling-algorithm, and you will spot the problem.

I found out, that this problem does not happen, if I submit same order into
1-min dataseries. It happens when I submit to the 1-range dataseries.

My expectation is that it relates with the gaps in the 1-range data,
It is really simple. I hope, it will be easy to find problem now.

[misova]

Hi,

I have prepared very precise scenario, that shows you the invalid fill.
The only thing you need to do is:

• read the screenshot
• go over filling-algorithm (that is not accesible now for developers),
• you will spot the problem easily

I found out, that this problem does not happen, if I submit same order into
1-min dataseries. It happens when I submit to the 1-range dataseries.

My expectation is that it is related with the gaps in the 1-range data and how filling algorithm processes this specific situation.

It is really very simple scenario. So I hope, it will be easy for you to find the reason,
why filling algorithm evaluates fill at better price, that would probably not happen in reality.

One small note:

• The profit order in the screenshot was submitted at 14:00 and then waiting to be filled for around 40 minutes...
• This is important, so you don't think it was submitted quickly somewhere between the 1-range bars in the screenshot. It waited very long time to be filled... and then, after 40 minutes, it was filled at wrong price = 1 tick better.

[NinjaTrader_Ray]

Hi,

I have prepared very precise scenario, that shows you the invalid fill.
The only thing you need to do is to go over filling-algorithm, and you will spot the problem.

I found out, that this problem does not happen, if I submit same order into
1-min dataseries. It happens when I submit to the 1-range dataseries.

My expectation is that it relates with the gaps in the 1-range data,
It is really simple. I hope, it will be easy to find problem now.

Hi misova,

Thanks very much for the effort put forth in isolating an issue. Unfortunately, I am unable to reproduce with the instructions provided. Limit orders are filling correctly in my tests.

The easiest and most efficient way for us to replicate what you are seeing is for you to provide us a sample strategy that reproduces the issue along with a screen shot of your backtest settings.

Thanks in advance.

[NinjaTrader_Ray]

Hi misova,

Thanks very much for the effort put forth in isolating an issue. Unfortunately, I am unable to reproduce with the instructions provided. Limit orders are filling correctly in my tests.

The easiest and most efficient way for us to replicate what you are seeing is for you to provide us a sample strategy that reproduces the issue along with a screen shot of your backtest settings.

Thanks in advance.

Misova,
Disregard this post. I may have found something. Once I investigate further, we will report back.

[misova]

One small note, on top of what was said:

• This problem occurs in my strategy, that uses Unmanaged mode.
• I explicitly submit order into 1-Range dataseries .
• Strategy is set up with Order Fill Resolution to value Standard (Fastest) = I want to say, that I do not use that new High Order fill resolution, that would redirect orders to another dataseries, set up from the UI.

[NinjaTrader_Ray]

The problem has been identified and fixed for the next release. Thanks again for reporting this and for providing us with the information needed to resolve.

[misova]

Thank you, great work !