I'd like to hear opinions about the results of my first ATS

H

Hiperfly

Hello there guys,

This is my first automated system ever, I use it for intraday futures (the ones you see in the pic below).

I'm not really into programming, so I have made this ATS in ToS and tested it for the last 2-3 months (not backtested, but in real time). Here is a screenshot of the results so far. I'd like to hear your opinions because I dont know which would be the benchmarks for some of these data.

The starting cash field is not that much important (could be 10k in the end), is the capital used for each contract. I just used it to calculate the avg. return. For the total field it would be 25k (or 50k)

5y4og.jpg


I know there is a lot to improve, if you could just give me some free advices I would be glad.

Ty guys.

P.S.: The 'general score' is just a dumb formula i did to compare between the different systems, I did 3 different ones although really similar. This is the one with the highest 'general score' and highest net profits as well. If you are interested in seeing the other 2 ATS results won't be a problem, just say it :D
 
Your first line title is "Total net profits", but it appears you are deducting the comms only after computing all your metrics ?!
=> Start with removing the comms.

How much slippage does your backtest accounts for? In the best market (RBOB), your net per trade after comms is $21.5 ... there isn't much room for slippage, nor for technical issues.

You should track the max drawdown ($$$ and duration).

Have you thoroughly reviewed all trades details, to ensure the reported fills make sense, and that you can trust these figures? (in particular, STPLMT orders are easily mis-simulated)

What is the largest position size used on a given market?

All in all, the average per trade seems pretty slim (2-ticks or less), unless you can prove to yourself that these simulated results are truly reflecting what you would get in real-trading, you don't have enough room to cover for all the uncertainties attached with simulated results.
 
dom993 said:
Your first line title is "Total net profits", but it appears you are deducting the comms only after computing all your metrics ?!
=> Start with removing the comms.
More...
You are completely right, with that title i wanted to mean "Profits without the initial capital".
dom993 said:
How much slippage does your backtest accounts for? In the best market (RBOB), your net per trade after comms is $21.5 ... there isn't much room for slippage, nor for technical issues.

You should track the max drawdown ($$$ and duration).

Have you thoroughly reviewed all trades details, to ensure the reported fills make sense, and that you can trust these figures? (in particular, STPLMT orders are easily mis-simulated)

What is the largest position size used on a given market?

All in all, the average per trade seems pretty slim (2-ticks or less), unless you can prove to yourself that these simulated results are truly reflecting what you would get in real-trading, you don't have enough room to cover for all the uncertainties attached with simulated results.
More...
Yea I have to calculate the DD for sure, but I dont know yet how to do it in excel.

I'm doing this ATS in a really coarse way 'cause I do not have the means (nor all the knowledge/experience needed) to do it properly. Even though, I think it worths to try it.

About the size and the order type, is always for 1 lot and aggresive buy to the ask or sell to the bid.


Thank you, waiting for more answers
 
all of those contracts will experience slippage on actual fills... to my knowledge, TOS sim is not that realistic unless they changed its structure from years past.

so if/when you go live, it becomes a breakeven to slow-bleed loss process in reality. Most "profitable" systems that the average person can write are barely above breakeven at best, and then turn to net-loss in the reality of actual fills.

systematic trading is a long, hard, winding road to success and not one iota easier in any way than discretionary trading for many reasons such as those above, and more.
 
Hiperfly said:
Could you give me some more opinions? Last bump, promise.
More...
=============
Hi-fly;
a] Always remember profitable trading may take longer than exspected;
A] always rememember what Rich Dennis said in Futures magazine interview ''its the job of the markets to WRECK computer systems''

Aa] Asking you , how long is the average trade/time or average bars /candles in trade???????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????Hardly ever see managed money in futures short term, see some , NOT much,, compared to longer/medium trends.[Aa ]Are you aware of that??

B+C] Always[c] compare cash trades to leveaged trades;
its a suprise the difference in cash[spot] profit + leveraged losses.!!!!!!!Palladium is in a great long term[25 year candlechart ] uptrend, but just to show you how tricky even a good uptrend is ; its too close to $1,,000 top for me now. Not a prediction nor is that a countertrend comment.

It may take me 20 or 50 days to reply; not a prediction

Wisdom is profitable to direct:cool:
 
I haven't doodled with commodities, so I am speaking somewhat from a position of ignorance--especially compared to the others who already posted. Regardless, don't you use margins in futures? If so, wouldn't you need some kind of stop-loss to prevent a margin call? I would suppose if you are not factoring that in, the system could be trying to ride out a loss that would have killed your account in reality.

There could also be some devils in unmentioned details. Like, say, I wasn't sure if you really meant you were actually paper trading this forward. If not, even if you were doing something like a forward test, but you previously optimized the settings, then they're curve fitted to that particular data. What you'd want to do regardless is have an out-of-sample test. It could even be older data.
 
You must log in or register to reply here.
Back
Top