I am seeing something odd in the Enhanced System Tester and curious if anyone else is seeing this...

I have developed a model and fed the Enhanced System Tester a series of stocks to run and compare. I told it I want each investment to default to a $50,000 position. Well, when I look at the ACTUAL trades in the output, they are NOT $50,000 orders! In most cases I am getting something like $45,000-$47,000. But never $50,000. Weird! Why?

I thought at first it might be defaulting to some rounded "block" size, but , no, the position details clearly show the simulator is calculating an exact number of shares (e.g. 723 shares, or 437 shares, etc.). Then I thought maybe it was using some price other than what I told it to use (which is the Open price), but, no, it is correctly selecting the open price on the day of interest. Lastly, I thought it was running out of money in my simulated account, but I inflated that to $1M and still get the same thing!

So I am baffled.... Why is $50,000 not $50,000 in this simulator? Any one have any ideas?

Thanks!

OK, thank you for the quick response.

That is indeed what is happening. I have the commissions set to zero, so that was not a problem. But... I am using a trick to buy at the Open, and the simulator is picking up the price of the PRIOR day's open when it calculates the number of shares. A bit strange, because the output report CORRECTLY lists the price of entry as today's open, but the number of shares calculated is clearly based off yesterday's open. Odd but that's what is happening.

Thanks for the help, Justin.

I am fine to share the formulas here; I'll just leave out portions that I want to keep private. But I can give the basic architecture:

I am playing with a simple "buy on gap down" model. I want to buy at the Open if the stock has gapped down and met a series of other criteria. Then I want to sell that same day at the close unless a stop has been reached. My first problem was getting the System to allow me to buy and sell on the same day, as it kept wanting to buy one day (at open) and sell the next (at close). Searching this forum over the weekend, I found the following trick works:

I defined a function that triggers a buy signal under my specific conditions. It looks at the prior closing price, sees if the open has gapped down by an amount I plug in there:

Function "Enter":
(OPEN<=Ref(C,-1)*(1-{gap down})) AND
(other conditions are met; these I'll keep private...)

Then I set up a System Tester as follows:

System Buy Order
Ref(Fml( "Enter"),+1)
with Strategic Delay set at +1 Day

System Sell Order
Set to sell at the close unless a Stop is reached; no Strategic delay

So as you probably recognize, it is looking ahead a day at the Open to decide whether to buy, then delaying a day to make the execution. Whereas the sell order is happening without a Strategic delay. And bingo, both the buy and sell happen on the same day.

I am guessing the reason the Cost Basis is off is because I am looking ahead a day, so part of the simulation is picking up the Open price on the day I look, rather than the target day. But as I mentioned previously, it still seems odd that the output correctly identifies the true open (and buy) price for the execution, but the math calculating the number of shares reverts back to the incorrect open price the prior day.

I'm game for a different approach if it will solve the problem of buying at the open and selling at the close the same day *and* has the number of shares in the resulting profit calculations.

Thanks!