OmniVest User Forums
Questions on Using OmniVest
Discrepancy between simulated and TP trades
15 replies, 1336 viewings
|
OmniVest User Forums
Questions on Using OmniVest
Discrepancy between simulated and TP trades
|
Last Activity 6/10/2024 12:04 PM 15 replies, 1336 viewings |
|
|
|
| Printer friendly version |
calmger Member Posts: 18 Joined: 1/1/2016  User Profile |
Now I see discrepancy between simulated trades and trades that has been done by Trade processor. The differences are in stocks that were selected for trades or in number of stocks. I did not change any settings in my only traded account. I would like to get a meaningful explanation why it happens. I attach my simulated and TP historical trades for the last 2 weeks. [Edited by calmger on 9/18/2016 11:37 AM]  Attached file : Comparison Simulation and Trade Processor.xlsx (15KB - 346 downloads) | |||
| BrianD Legend  Posts: 302 Joined: 2/23/2013 Location: Grand Rapids, MI User Profile |
Different Values in Trading, Current and Simulated account balances often cause this. Were all balances set with the same $$ value on Sept 6th? Net, if OV/TP finds it has 16K more to spend, it will. If OV/Simulation has ~16K less, it will Simulate different allocations and even symbols. Especially when using Share Rounding - one NFLX lot was worth 10K. Note too that a broker may cancel an order, overnight prices do change, etc. These all create discrepancies, especially when the account is running fully allocated. However, for me in my experience and based on my settings, my accounts balance out quickly and Sim/Historical get back on track again. I too like to validate my Strategy/Portfolio/Account performance expectations by seeing Sim and Historical track 1:1. I watch closely and see a discrepancy once in a while, but overall my account Simulations to Historical track closely, within +-2%, with a high TPM rate. I use IB too. [Edited by BrianD on 9/18/2016 10:18 PM] | |||
| SteveL Veteran  Posts: 189 Joined: 10/11/2012 Location: Boulder, CO User Profile |
| |||
| Barry Cohen Icon   Posts: 1844 Joined: 10/11/2012 User Profile |
On your Accounts page, looking at the simulated trades, we see that NFLX does indeed show 200 shares. But if we look at the actual trade on the Trades page, we see that it was 300 shares at the time it was tracked/saved on 9-7. So why are they different? The Trades page (which is what the TP uses) tracks or saves the trades each day. Once trades are Historical Positions they remain there forever (unless you reset the account). Simulated trades on the other hand are not saved on a day to day basis. You change anything in the account (setting, portfolio, etc) & the Simulated trades will update to those changes & show them instantly. As a result, it's very difficult to get them to match especially over time. The closest you can get is to make sure the Simulation Starting Date/Balance both match your Trading Start Date/Balance & never change anything after the first day. The slightest change in the account, will cause them to never match again. If we look at your Simulation settings, it shows that it's calculating all your Simulated Trades using a starting balance of $60,000. But your Trading Start Balance shows $100,000. Not only that, but your balance at IB wasn't $100,000, so after you started that account, at some point afterwards you connected your TP to it & your actual broker balance updated your Trading Current Balance. (That isn't taken into account in Simulations as Simulations just use one starting balance) And I also see some activity where you made a couple changes in your account after you connected the TP to it, but also all the settings changes from before the TP was connected. All the changes made in setting up the account would be fine, but the important piece is that there was no account reset performed after everything in the account was finished being set up. This is something we plan to improve. Ideally we want OV to prompt an account reset after connecting a TP so that it starts the account with your broker's exact balance. So for someone that wants to get Simulated Trades to match (which isn't really necessary, but some may choose to do so), you would need to set your OV account up exactly how you want it, connect your TP to it, then reset the account using your current broker balance as the starting balance. Going forward never change any settings or portfolios in that account. Any deviation from that will cause them to not match. Attached file : Simulation Settings.jpg (34KB - 582 downloads) Attached file : Account Settings.jpg (15KB - 589 downloads) Attached file : Simulated Trade.jpg (10KB - 586 downloads) Attached file : Trade Tracked.jpg (10KB - 589 downloads) | |||
| BrianD Legend Posts: 302 Joined: 2/23/2013 Location: Grand Rapids, MI User Profile |
I also felt I was missing OV strategy initiated opportunities in a position in for instance Amazon, Google, so...Obviously, commissions in smaller $$ accounts with small lots and higher TPM is another issue, but Simulation does give you a feel for that impact. But ultimately, I'd like calmger to find comfort with OV between Sim and Historical. If you take into consideration what Barry laid out above, you should. | |||
| calmger Member Posts: 18 Joined: 1/1/2016 User Profile |
Thank you for your efforts to help me. Brian and Steve, I appreciate your notice with regard to the ability to trade uneven lots with IB not suffering of bigger slippage. Now I am thinking to test this opportunity. However a rounding to 100 stocks cannot explain why in simulations and real trades different stocks are used. So it is not only discrepancy in number of stocks in position, but different from simulation stocks are sent to trade. You can see it in My Excel file. There are 4 examples, when stock can be found in simulation (EWZ,SIRI) and was not traded by TP. Also conversely stocks were sent to TP and traded (PCAR, ARCO), but never appeared in simulation. Barry, it seems to me that since I started trading in real account I did not touch any settings. I paid a special attention to this because you warned me in February 7 months ago. I had a long test for 7 months (Feb 2 - Sep 2) trading in IB paper account. I started having $35K in both simulation account and in IB paper account. I was watching how a discrepancy between simulation and real trades were growing, and could not find any explanation. Finally real trading in paper account brought me somewhat 45% annualy, while simulation did 95% annualy. I would say it is huge discrepancy for 7 months. I did not change any settings in account within this time period. With regard to "Trading Starting Balance" once the account is created with $100,000 later the web site does not allow to change it when you decide to trade this account. The web site allows to type "Trading Current Balance" only. So I changed to $60,000 and started trading. However my IB account had $67,144. I did it with a purpose to have some additional money on real account to avoid the situation when IB can reject the order. I have a delay to open orders 30 seconds, but a couple of times orders to open were rejected, so I open manually and selected to go with Omnivest. Barry, your explanation is based completely on the fact that the discrepancy can be caused by following account settings changes. While it is reasonable in my case it cannot be explanation because I restrained to make any changes. So I have a concern that Omnivest has some deficiency in program code or in the logic. Brian and Steve do not confirm though... | |||
| BrianD Legend Posts: 302 Joined: 2/23/2013 Location: Grand Rapids, MI User Profile |
I always found IB data for paper trades to be questionable. IB could not give me a solid answer why, at a minimum, it would not publish Open and Close numbers (which is what OV uses) accurately, except to say it was a paper account. In real trading, I (and others) have tracked slippage for a long enough periods to believe, on average, at least using IB Smart, I do better than published OV Open and Close values over time. So you may be wary of using paper in drawing empirical conclusions. You said you had 60K in Simulation and 67K in account. I'd defer to Barry's followup on this, but I believe this is where they separated; Simulation was looking to allocate 60K, TP was looking to allocate 67K. Maybe that's what they did on the 7th. And they diverge from there...For me, any time I have not reset an account to have both exactly the same value in all fields, Live and Simulated have gone their separate ways. However, I still find Sim and Live will have divergences. Nature of the trading beast - slippage, lack of availability, market price changes, (as well as account settings) etc. I overcome some of this by favoring large volume, highly liquid stocks. But I hate to see different symbols between Sim and TP live and have been down your path of concern. Happens once in a while, usually at full allocation, and I can usually find a root cause. But as I said earlier, it always levels out and gets back on track. But as they say, stay vigilant! | |||
| Jim Dean Elite Posts: 1059 Joined: 10/11/2012 Location: L'ville, GA User Profile |
Simulation is - simulation. Not real world. Similar to real world. So, don't expect the two to match - doing so creates headaches not profits. Simulation is intended for testing and development and making design choices - doing A vs B comparisons - where A and B are both simulations. If you have a (healthy) concern for the how's and whys and reliability etc of the live-trading side of OV, then focus your time and study and questions on "why did OV make such and such a decision" without trying to compare it to simulation runs that are NOT supposed to track with it, for a large number of (unavoidable) mathematical/logical reasons. The same goes for OT/VT or any other trading platform that does both historical simulation and also live trading. The two will not be precisely the same. Shorter and more pithy: Study the apple. Don't compare to the banana. | |||
| Barry Cohen Icon Posts: 1844 Joined: 10/11/2012 User Profile |
Here is the thread you mention. The important text is this: "Your Account Log shows that you've made several changes in the account as far as settings & portfolios being enabled/disabled. So your simulated trades & trade record will never match, which is very normal. If you absolutely want them to be equal you have to reset your account & never make any changes to that account going forward." I do not see that a reset account was performed using your broker's balance as the starting balance. That reset is absolutely necessary if you have any chance in matching simulated trades. Just making sure you don't make changes in the account is not enough. You have to reset AND not make changes. This is something that we need to fix & it is on the list of things to change. Again, what ideally should happen is that you create your account, set it up, then connect your TP. During that first TP connection it should ask you if you want to reset the account or not. If yes, then it resets it for you using your broker's balance as the starting balance. This is not implemented yet, so we have to manually reset it for now. If you don't mind your Historical Positions on your Trades page being different from the trades taken at your broker, you could reset your account now if you like. Then your Trades page will match your Simulated trades. If you decide this, just set the starting balance as the exact current balance at your broker. | |||
| calmger Member Posts: 18 Joined: 1/1/2016 User Profile |
| |||
| BrianD Legend Posts: 302 Joined: 2/23/2013 Location: Grand Rapids, MI User Profile |
Better to close all positions manually, or stopping OV/TP from entering new trades (disable "automatic entries") so exiting positions are eventually closed by the Strategies, etc. Once there are no open positions, then reset and start on a level (equity value) playing field. | |||
| Barry Cohen Icon Posts: 1844 Joined: 10/11/2012 User Profile |
| |||
| BrianD Legend Posts: 302 Joined: 2/23/2013 Location: Grand Rapids, MI User Profile |
My bad. Thanks for the correction. [Edited by BrianD on 9/21/2016 5:52 PM] | |||
| calmger Member Posts: 18 Joined: 1/1/2016 User Profile |
| |||
| calmger Member Posts: 18 Joined: 1/1/2016 User Profile |
The check mark in the page is grey read only. | |||
| Barry Cohen Icon Posts: 1844 Joined: 10/11/2012 User Profile |
If you don't want a certain account trading, you can disconnect your TP if you haven't already. Then connect a new TP to your other account. |
|
|
|
|
Administrator
Forum Moderator |
Registered User
Unregistered User |
 E-Mail this thread to a friend(un)/Freeze thread | |
Toggle e-mail notification |
