Backtesting assumptions problem

Questions about MultiCharts and user contributed studies.
turbofib
Posts: 213
Joined: 11 May 2013
Has thanked: 67 times
Been thanked: 1 time

Backtesting assumptions problem

Postby turbofib » 16 Apr 2018

hi, look this pic

i set in backtesting assumption Fill Limit order 1 point

But in the chart buy order is entry at high candle (2716.25 )

Why? it must only entry at 2716,25 if high candle was 2716,25+1 point
Attachments
Immagine.png
(121.13 KiB) Downloaded 740 times

User avatar
Henry MultiСharts
Posts: 9165
Joined: 25 Aug 2011
Has thanked: 1264 times
Been thanked: 2957 times

Re: Backtesting assumptions problem

Postby Henry MultiСharts » 17 Apr 2018

Hello turbofib,

A Limit order will execute at the specified price or better. A better price is a lower price for Buy order.
According to your assumption settings the order can be filled at limit price, if there is 2726.25 - 1 point improvement.
This condition is satisfied at this bar, therefore the order is filled.

AdrianP
Posts: 46
Joined: 02 Sep 2014
Has thanked: 2 times
Been thanked: 1 time

Re: Backtesting assumptions problem

Postby AdrianP » 10 Jan 2019

Turbofib is correct Henry. Its clear after emailing Anna recently on this issue, that both she and software engineers are applying this selection box incorrectly or just don't understand it. And your response indicates you are also seeing it the wrong way.
The dialog box is very specific in what it states. Only fill the order IF price exceeds the limit order price by the specified amount. With a setting of 1 therefore, it should be 100% impossible to get a fill on the high of the bar. Just as Turbofib explained.
There are other deeper problems with this checkbox choice which I am posting to the PM site. Silly really, as anything there is mostly a waste of time, and this particular section of code is FLAWED, so its not about getting votes on whether to fix a flaw. That's insane. It should just be fixed...period.
Please read my PM post.

wilkinsw
Posts: 662
Joined: 21 Apr 2013
Has thanked: 154 times
Been thanked: 104 times

Re: Backtesting assumptions problem

Postby wilkinsw » 11 Jan 2019

Turbofib and AdrianP,

Can either of you give a clearer explanation of the issue you are experiencing?

Ideally can you share some code please?

A couple of glaring issues:

1) Please share you code: in your pics it looks like all your fills are on stop entries, NOT limit entries. So am a little confused as to why you are highlighting a limit fill assumption issue. I'm guessing Henry missed this when he answered. I might be wrong; please share your code and create a repeatable test case.

2) You are using an older version of MC, by the looks of things (MC12r4 doesn't have the Time in Force dropdown). I've spent much of this week going through MC12r4 and am yet to find any real issue with fills assumptions. In fact it looks like MC12r4 is the best Multicharts yet on that front.

AdrianP; I'm fascinated to know what issues you are posting to PM. Again please share links etc. I've put a lot of effort into working through fill assumptions with MC over the last couple of years. I'm happy with all the fill assumptions that I've tested so far in MC12r4. If you both are using the forums to raise issues please please please put in a little extra effort and provide all the info needed to allow other users to reproduce and advise.

I'm speculating you are both wrong in this case. Put forward 100% reproducible examples on MC12r4 and we can get somewhere.

AdrianP
Posts: 46
Joined: 02 Sep 2014
Has thanked: 2 times
Been thanked: 1 time

Re: Backtesting assumptions problem

Postby AdrianP » 11 Jan 2019

Hi Wilkinsw,

Please read this post and let me know what you think.
https://www.multicharts.com/pm/public/m ... es/MC-2545
Its only relevant to Limit orders, and particularly those executed in the Forex cash markets, where spreads are extremely important.
If I am misunderstanding something about the issue, I need and desire to know.

wilkinsw
Posts: 662
Joined: 21 Apr 2013
Has thanked: 154 times
Been thanked: 104 times

Re: Backtesting assumptions problem

Postby wilkinsw » 12 Jan 2019

Hi AdrianP,

I can see from your posts that you are really worked up about inaccuracies you are suffering when modelling fills in FX markets using just Bid data.

I think your post in this particular thread is misplaced. Turbofib was trying to demo an issue in the ES futures contract. Stating "Turbofib is correct Henry" and then ranting, i feel, was not helpful at all.

I will post a reply with suggestions to your FX fill modelling issue in PM.


Return to “MultiCharts”