[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re : Same System, Same Datafeed, different result



PureBytes Links

Trading Reference Links

Dans un courrier daté du 03/10/98 05:43:10  , vous avez écrit :

<< 
 Fichier :  report.txt (2394 octets)
 Durée du téléchargement (28800 bits/s) : < 1 minute
 
 Here's the same system , with the same datafeed type, same time frame, same
 dates and totally diffrent results.
 
 Same system does the exact same thing on the S&P, and Bonds. >>

Pendragon Ver.10.1  NDX-60 min   05/09/96 - 10/02/98			 

	Performance Summary:  All Trades		 

Total net profit	$ 167581.00	Open position P/L	$      0.00
Gross profit    	$ 230059.00	Gross loss      	$ -62478.00

Total # of trades	     173	Percent profitable	      58%
Number winning trades	     101	Number losing trades	      72

Largest winning trade	$   7221.00	Largest losing trade	$  -1657.00
Average winning trade	$   2277.81	Average losing trade	$   -867.75
Ratio avg win/avg loss	       2.62	Avg trade(win & loss)	$    968.68

Max consec. winners	       9	Max consec. losers	       6
Avg # bars in winners	      17	Avg # bars in losers	       7

Max intraday drawdown	$  -5069.00		 
Profit factor   	       3.68	Max # contracts held	       1
Account size required	$  13069.00	Return on account	    1282%
===============

OK.

Some clues for you:

1-Seems that your system has exits to be able to feature a 1:33 Net profit /
maxDD ratio.
If fyou used the include system, do not forget that the information are not
transmitted from the master system to the included one, unless you use global
variables.
2-Check if it produces the same results on the same machine when realtime  and
historical.
3-Check again by closing the charting an reopening in the morning before the
market
4-Check also if the results are different when changing the Bouncing ticks
setting.
5-Build a new systm that generate more trades (change some parameters) and
redo the above testing. (only to have more cases to study)
6-Try to identify the bars where differences appears, and the variables that
are involved in the errors.
7-Focus onte guilty variable and how it's calculated.
8-Print anything that is useful to understand the error appearing
(intermediate variables from the raw data to the faulty variable value).

The you will find...

For the moment, it's the only thing that I suspect from this document.

Rgds

Pierre Orphelin