UNBELIEVABLE p/l

[revlej]

First time with automated trading... using TradeStation w/ EasyLanguage coding our own set of "trading rules" and the returns it shows backtesting is amazing; the returns are too amazing in fact...

The backtesting doesn't currently take into account slippage, commission, or exchange fees, but is there anything else we are missing?

The skeptic in me refuses to believe infinite wealth is this easily obtained... hopefully some people on here can spot some errors we are making before we decide to go live and potentially lose real money.

Thanks in advance!!

 

[scose-no-doubt]

but is there anything else we are missing?

reality?

 

[the hare]

postdictive errors are very common. These types of errors are caused when you use data that you wouldnt actually know at the time a trade was placed. For example you might use the value of an indicator based on the daily close, or place a stop under the daily low. These errors are very common if you are using some kind of mult timeframe approach.

Its also worth remembering that its extremely easy to curve fit a decent backtest over a given period of time. If there is an element of curve fitting (optimising indicator settings, stop levels, targets etc) at least split your historic data into two sections. Optimise on one half, and then ssee how it performs out of sample.

To be brutally honest, its usually slippage, commission and spreads that cause systems to fail. If you eliminate them, or dont account for them things generally look a great deal better than they are in reality. They have a far bigger impact than mosst people realise.

 

[Masquerade]

I love when people backtest things which worked over the time period of 2000-2006 then go live with it and wonder why it doesn't work

 

[DionysusToast]

Tradestation has a few quirks.

First thing to check - is it giving you entries at the exact open or exact close price of each bar?

That's the normal thing with TS strategies, they tend to have a skew - it's been quite a few years since I did any of that but I remember a similar issue making me believe I was going to own the world.

 

[Trader333]

I also managed to turn a $10K account into over $4M using tradestation backtesting and I had accounted for slippage and commission in doing it. However, the reality was that it was untradeable from a psychological standpoint.


Paul

 

[NVP]

First time with automated trading... using TradeStation w/ EasyLanguage coding our own set of "trading rules" and the returns it shows backtesting is amazing; the returns are too amazing in fact...

The backtesting doesn't currently take into account slippage, commission, or exchange fees, but is there anything else we are missing?

The skeptic in me refuses to believe infinite wealth is this easily obtained... hopefully some people on here can spot some errors we are making before we decide to go live and potentially lose real money.

Thanks in advance!!

leave any trade that is negative pips open......

 

[pboyles]

First time with automated trading... using TradeStation w/ EasyLanguage coding our own set of "trading rules" and the returns it shows backtesting is amazing; the returns are too amazing in fact...

The backtesting doesn't currently take into account slippage, commission, or exchange fees, but is there anything else we are missing?

The skeptic in me refuses to believe infinite wealth is this easily obtained... hopefully some people on here can spot some errors we are making before we decide to go live and potentially lose real money.

Thanks in advance!!

So why dont you put in an amount for commission? Its very simple to do.

The results are probably curve fitted by adjusting TP and SL levels or removing non profitable instruments. Like others have said pretty much anything can be made to look profitable in a backtest if you curve fit.

 

[Hotch]

I have probably backtesting more than most, and have never come across such unbelievable strategies. I must be doing something wrong/right.

 

[DionysusToast]

Hotch

Try writing a system on Tradestation and then applying it to their "3 line break" bars.

You'll sh1t your pants when you see the results.

DT

 

[Hotch]

That would require using their code I guess?

 

[DionysusToast]

no - just any TS strategy & switch the chart type to 3 line break.

 

[Hotch]

Yeah, but they've coded how to plot the 3 line break, I'm guessing I know what sort of problem it is, silly people, next they'll be using boost.

 

[DionysusToast]

Hotch - the issue with their 3 line break is that if you run a strategy on it backtesting it will act as if the high & low of the three line break was the high & low price.

So - actual price is ignored.

I emailed them about this and they told me "if you dont trade those prices, you aren't trading the 3 line break". I then asked them if their brokerage would allow me to trade those prices. Didn't hear much back from them.

 

[eminiglobex]

no - just any TS strategy & switch the chart type to 3 line break.

Can't say I've ever noticed this. How do you change chart type to 3 line break?

 

[eminiglobex]

First time with automated trading... using TradeStation w/ EasyLanguage coding our own set of "trading rules" and the returns it shows backtesting is amazing; the returns are too amazing in fact...

The backtesting doesn't currently take into account slippage, commission, or exchange fees, but is there anything else we are missing?

The skeptic in me refuses to believe infinite wealth is this easily obtained... hopefully some people on here can spot some errors we are making before we decide to go live and potentially lose real money.

Thanks in advance!!

unrealistic fill assumptions if using limits. ie price will hit the target. ts will assume fill, but in reality you will not get it. what happens when you use "fill entire order when trade price exceeds limit price"