I used the real-time auto-trading paper account and back test to test my trading strategies and found the position profit on the paper account was different from the positions profit on the back test. Here are my questions:

1. Will the live account auto-trading position profit be same as the real-time auto-trading paper account position profit?

2. How can I verify the real-time auto-trading position price of the paper account? I could not find the auto-trading paper account position price from the live trading records based on the transaction time. In my auto-trading paper account records on 04/20/2018 below: the 1st row shows that TQQQ was bought at 10:25am for $150.88 (at 5 minutes scale), but in the live trading records, the price range was $151.30-$152.07 at that time (10:25am). The 2nd row shows that TQQ was sold at 10:57am for $152.49, but in the live trading records, the prices from 10:45am to11:00am were between $149.36-$151.65. Both of the buy and sell prices, $150.88 and $152.49 can not be identified in the live trading record. The live trading records came from Yahoo.com.

Action Account Date/Time Quantity Symbol Price Trade Type Strategy Scale

Buy PaperAccount1 4/20/2018 10:25 AM 660 TQQQ 150.80 DRAFTS-1 5 Minute

Sell PaperAccount1 4/20/2018 10:57 AM 660 TQQQ 152.49 DRAFTS-1 5 Minute

3. In real-time paper account 5-minutes trading, the increment of transaction times does not follow the 5-minute rule as shown in back test. It seems that it can be executed at any minute, such as at 10:57 am, instead of 11:00 am. Does it follow a different set of rule.

4. In the paper account, is the auto-trading transaction time based on my computer clock or based on the Wealth Lab's system time?

Hi Eugene, Dave, KCbars, DWitzel, mdsag125, and ALL

Like to share my experiences of applying Auto-trade at Fidelity and also ask some questions concerning the actual auto trades.

I talked to a Fidelity Rep for the application. He asked me to close down my WL on my PC so that he could reinstall another version of WL with the application form. After he finished the installation, I still cannot see the application form. Then he realized that I had only 488 trades for the past 12 months. So right now I am anxiously awaiting to complete my 500 trades and then to see if I am qualified for Auto trade.

In the meantime, I have been doing some auto trades with paper accounts. I run into some problems similar to the questions posted by mdsagf125 concerning the different results from the back-testing and the auto paper trade as well as the timing issues. Because of the differences, it leads me to question whether the auto trade results of the live account is the same as the paper account.

It sounds that KCbars had some good experiences with the Auto-Trade, would you please elaborate your experience on hoe to deal with this issue or any intricacy involved. And also Eugene, from reading the posted questions and answers, I know your the EXPERT of it all. Please chip in your expertise knowledge.Thanks.

@mdsagf125
1. No. The User Guide > Orders chapter explains differences in trading paper accounts and live trading executions.

2. Yahoo.com - you're working with 15-minute delayed data.

3. We don't know anything about your strategy. Limit and stop orders, for example, can be filled at any time during the interval.

4. Wealth-Lab doesn't keep time. Times shown depend on where you're looking, and should be identified in the User Guide. You're going to either see the Alert's time (End of bar) based on the bar data, or an execution time, which in the case of the Accounts Local Trade history is based on your system clock.

Look, paper trading is not likely to track live trading very well - especially if you're using delayed Y! data. There are other limitations with the way paper trades are executed. The idea of paper trading is to give you an idea how auto-trading works, so that you can become comfortable with the cues and other issues that may occur.

@claifu66

QUOTE:
leads me to question whether the auto trade results of the live account is the same as the paper account.

Definitely not!

I would expect backtesting to be more representative than paper trading. Why? User Guide > Orders chapter.

Thanks Cone, appreciate your fast response. Will study the Orders chapter.

Cone, thanks for your quick response!

As for Y! data, I also compared my paper account transection price with the Fidelity data, but it was out of the high-low range too (based on the transaction time).

Cone,

What if the strategy only use Market Order? Does that imply all restrictions caused by stop or limit order should have no impact on the trading? For example, if I apply the 5-min scale and the strategy issues a signal at bar i to buy, the actual transaction time should be at the opening of bar i+1, but not earlier or later. More specifically, if the strategy issues a buy at 10:30, WL would execute the buy transaction at 10:35 with a market order and not in between, right? And the transaction only depends on the time not on the price.

As you have indicated that the auto trade of live account is definitely different from paper account. But which one provides a better results? I guess that's an unfair question. I should ask which is more true to the reality. I have also tested that the results of auto trade of paper account is different from the back-testing. You expect that the results of back-testing is more representative of the live account than the paper trade. Could you enlighten me as to why, because I couldn't quite figure out the reason from reading the user guide. Is it because paper trade use the limit or stop order in a simulated environment as described in the user guide or there are other reasons?

Finally, if auto trade in a live account where the limit or stop order is totally avoided, would it make the trading events more adapted to the real dynamic market if the strategy is right?

Again thanks!

QUOTE:
What if the strategy only use Market Order?

In backtesting market orders always executed at the opening price of the bar, unless you enable slippage, in which case the hypothetical trade will almost always be executed at a worse price than the open. In trading, sometimes you'll get the opening price, sometimes a worse price (slippage), but also you may actually get a better price!

QUOTE:
More specifically, if the strategy issues a buy at 10:30, WL would execute the buy transaction at 10:35 with a market order and not in between, right?

Yes, but it's the trade's bar timestamp that's labeled 10:35. The contents of the bar are all the ticks between 10:30:00.001 to 10:35:00.000. A market order issued at 10:30.00 (10:30 bar timestamp) will most likely be filled just a few seconds later, say 10:30:03 (which will appear on the bar timestamped 10:35 for 5-min bars).

QUOTE:
And the transaction only depends on the time not on the price.

What do you mean? A fill will always depend on a price as explained above.

QUOTE:
You expect that the results of back-testing is more representative of the live account than the paper trade. Could you enlighten me as to why,

Because prices may have moved significantly between the polling interval of the paper trading engine - I think that interval is 20 seconds. This means that limit orders can often be filled at better prices than in live trading. In backtesting, you'll [almost] never get a better price than the limit (except for opening gaps, which are only practically applicable to EOD trading), and you're also unlikely to get a better price in live intraday trading.

QUOTE:
Finally, if auto trade in a live account where the limit or stop order is totally avoided, would it make the trading events more adapted to the real dynamic market if the strategy is right?

With market orders, you'll almost surely avoid the potential "strategy is out-of-sync" condition (if that's what you're referring to), but that's a strategy design and trading decision you have to make. Market orders can be risky in a volatile market.

Thanks so much Cone! Since I am still new in dealing with WL environment, please be patient with me. I am trying to recap whether my understanding is correct or not. (Following the order of the five quotes above:)

Quote 1, So it's random and depends on your luck if you get a good price or not. Essentially, it's a wash.

Quote 2, Once the order is issued, it might be filled sooner than waiting for the next cycle of the scale as in your example. So WL is very sensitive to the order issued timestamp. Does that imply that the execution of the trade and the scale are not aligned on the same time increment? What confuses me is in the code using bar i or i+1. if the condition is triggered at bar i (10:30), we have to use i+1 (10:35) to trade and it will be filled at the timestamp of i+1 plus some seconds (10:3505), but appears on the bar timestamped as i+2 (10:40).

Quote 3, It's a mute question. What I meant to ask was under the limit or stop order, the price will also be the criteria for a fill not only time.

Quote 4, The answer at this point is beyond my comprehension. I will try to study it further. For now, suppose I'll be contended with the trade at or near the opening price of what the strategy dictates, not worrying the best or worst price, can WL accommodate that for me?

Quote 5, Same problem I have with Quote 2, confusion of i and i+1. Using i will be very responsive to the strategy, i+1 has a legging time. So please disregard the question.

Again thank you Cone for your patience. I had made 6 more trades during the last few days or 6 more trades nearer. I am counting on the day I try my hands on Auto-Trade. By then, question or no question, the results will tell.

1. I don't think "random" is the right word. Market orders hit the ask.. it's all about timing - where the ask price is with respect to the actual open of the bar at the time your order is placed by the broker and filled. There are delays involved... getting the order to the broker, broker to the market, etc.

2. Oh gosh, it's a lot simpler.
WL executes the Strategy when a new complete bar enters the chart. That is "bar". It takes several milliseconds to run the script and a few more to send the order out. That "Alert" is generated practically immediately, but clearly this order will be active for the next bar, i.e., bar + 1.

4. Paper trading is fake. Every so often (like every 20 seconds), WL's paper broker looks at what orders are active and sends out requests (polls) for prices to see if it should execute an order. Here's an example.

10:30:01 Limit order entered: Buy 100 ABC @ 22.34
10:30:02 WL Paper Trader poll for price: it's 22.50
10:30:05 Price is now 22.30
10:30:10 Price is now 22.20
10:30:15 Price is now 22.27
10:30:19 Price is now 22.35
10:30:20 WL Paper Trader poll for price: it's 22.36

So what happened? The paper trader did not execute that trade [yet] because it didn't know the price dropped below the limit. Obviously live trading isn't work that way.

Cone,

Don't know how to thank you! I am sure there will be tons of questions down the road to ask you.