frugi
1
- Messages
- 1,843
- Likes
- 133
Just for a laugh I backtested the rather poor system posted on my website and it actually turned out a half reasonable result, although in the months following (Feb 2004 - now) it has taken a hiding.
Very simple rules:
60 minute ES chart 2.30-9.15 GMT
Buy (and cover if necessary) if 7 EMA crosses up through 25 EMA and is still above at close of bar
Sell (and short if necessary) if 7 EMA crosses down through 25 EMA and is still below at close of bar
Use a 19 pt initial stop loss. At EOD move stop to 19 pts away from close if close is in a profitable direction otherwise leave it.
Take profits if 62 points profit attained at close of any bar
Obviously as this is a nearly always in system the buy and sell signals are usually used to stop and reverse.
I would also consider increasing stake after 10 consecutive losses until next win and decreasing it after 5 winners until next loss. Losses are caused by consolidation therefore after an extended period of consolidation you'd expect a nice big trend.
Tested using 2 contracts allowing for $35 slippage at entry and exit and no commission. This is because the deal4free spread on SPX futures is 0.7 therefore you will pay spread of 0.35 a point from the mid price a side. Trading 2 real contracts this would equal $35. Trading direct thru IB u could expect $29.80 a side ($2.40 * 2 commission and $12.50 * 2 for 1 tick slippage).
This is my first attempt a programming so there may well be errors. I have a feeling the drawdown is wrong but I haven't worked out why yet!
In the real world, using a £18k account at £5 per point, it risks about 3% per trade (based on average loss cause the catastrophe stop is rarely hit) and if you had 10 consecutive average losers you'd lose 30% of your capital. Average return about 36% a year on 18k. This is why I'm not so keen to use it! But maybe this ain't so bad?
I've calculated the average loss and return by testing 1,2,3,4,5 and 6 yrs and taking an average of the average loss and return per test period. Clear?
The screenshot that follows is for the longest test, over 6 yrs. The return for this at £5 per point is more like 50% but we don't want to be too optimistic do we.
Also bear in mind the risk involved in holding over the weekend when SBs and Globex are shut, unless u are using SB guaranteed stops (more expensive, more slippage, grrr). not sure what to do about this; I may have to incorporate a market on close if a position is open at 9.14 on Friday night and then re-enter at the open on Monday. Time for more tests then.
Very simple rules:
60 minute ES chart 2.30-9.15 GMT
Buy (and cover if necessary) if 7 EMA crosses up through 25 EMA and is still above at close of bar
Sell (and short if necessary) if 7 EMA crosses down through 25 EMA and is still below at close of bar
Use a 19 pt initial stop loss. At EOD move stop to 19 pts away from close if close is in a profitable direction otherwise leave it.
Take profits if 62 points profit attained at close of any bar
Obviously as this is a nearly always in system the buy and sell signals are usually used to stop and reverse.
I would also consider increasing stake after 10 consecutive losses until next win and decreasing it after 5 winners until next loss. Losses are caused by consolidation therefore after an extended period of consolidation you'd expect a nice big trend.
Tested using 2 contracts allowing for $35 slippage at entry and exit and no commission. This is because the deal4free spread on SPX futures is 0.7 therefore you will pay spread of 0.35 a point from the mid price a side. Trading 2 real contracts this would equal $35. Trading direct thru IB u could expect $29.80 a side ($2.40 * 2 commission and $12.50 * 2 for 1 tick slippage).
This is my first attempt a programming so there may well be errors. I have a feeling the drawdown is wrong but I haven't worked out why yet!
In the real world, using a £18k account at £5 per point, it risks about 3% per trade (based on average loss cause the catastrophe stop is rarely hit) and if you had 10 consecutive average losers you'd lose 30% of your capital. Average return about 36% a year on 18k. This is why I'm not so keen to use it! But maybe this ain't so bad?
I've calculated the average loss and return by testing 1,2,3,4,5 and 6 yrs and taking an average of the average loss and return per test period. Clear?
The screenshot that follows is for the longest test, over 6 yrs. The return for this at £5 per point is more like 50% but we don't want to be too optimistic do we.
Also bear in mind the risk involved in holding over the weekend when SBs and Globex are shut, unless u are using SB guaranteed stops (more expensive, more slippage, grrr). not sure what to do about this; I may have to incorporate a market on close if a position is open at 9.14 on Friday night and then re-enter at the open on Monday. Time for more tests then.
Last edited:
