bangkoker
Well-known member
- Messages
- 314
- Likes
- 17
Some big moves this week. looks like I would have been:
M) Didn't trade (thought +40, would have actually been b/e), nearly +40
T) B/e
W) +15 (stopped out by 1pt and then went on to hit 40pt level)
T) B/e
F) - 15.
So actually no points for the week.. but M and W were very close to +40.
Clearly meaningless at this stage. Will try to get some back testing done.:whistling
Have a good w.end all :clap:
M) Didn't trade (thought +40, would have actually been b/e), nearly +40
T) B/e
W) +15 (stopped out by 1pt and then went on to hit 40pt level)
T) B/e
F) - 15.
So actually no points for the week.. but M and W were very close to +40.
Clearly meaningless at this stage. Will try to get some back testing done.:whistling
Have a good w.end all :clap:
Last edited:
Thoughts on position sizing.
It seems to me that a trade system that can only generate 1 entry signal per day, with a 60% win rate, on about 4 days out of 5, and whose positions can only last 4hrs, demands a hefty stake to be at all financially interesting.
The standard advice of course is that not more than 2% of an account should be at risk on any single position. For BB trades, following the logic above, that seems to demand a pretty big account or to remain BB trading for pocket money. Fixing risk per trade is very sensible, and 2% risk per trade should ensure that a string of losers does not damage the account to a degree that you can't trade back to b/e without increased risks. So what's the likelihood of a string of losers in BB trading?
Maybe someone can work this out mathematically but I don't know how to do that. What I can do is review trade records and use the past as a guide. Going back to when I started applying a fixed points target and stop (14/02/11) rather than the depth of the BB range the system shows a performance of 39-30 (57%). Over the period, excluding holidays and no-trade days, the maximum consecutive number of stopped out trades is 4: this has occurred twice. The worst periods of 20 trades showed performances of 10-10: there have been no losing periods of 20 trades, assuming 1:1 r:r.
This finding surprises even me, I have not worked this out before. However, I feel it does help justify my raising my position size to 4-5% of account.
It seems to me that a trade system that can only generate 1 entry signal per day, with a 60% win rate, on about 4 days out of 5, and whose positions can only last 4hrs, demands a hefty stake to be at all financially interesting.
The standard advice of course is that not more than 2% of an account should be at risk on any single position. For BB trades, following the logic above, that seems to demand a pretty big account or to remain BB trading for pocket money. Fixing risk per trade is very sensible, and 2% risk per trade should ensure that a string of losers does not damage the account to a degree that you can't trade back to b/e without increased risks. So what's the likelihood of a string of losers in BB trading?
Maybe someone can work this out mathematically but I don't know how to do that. What I can do is review trade records and use the past as a guide. Going back to when I started applying a fixed points target and stop (14/02/11) rather than the depth of the BB range the system shows a performance of 39-30 (57%). Over the period, excluding holidays and no-trade days, the maximum consecutive number of stopped out trades is 4: this has occurred twice. The worst periods of 20 trades showed performances of 10-10: there have been no losing periods of 20 trades, assuming 1:1 r:r.
This finding surprises even me, I have not worked this out before. However, I feel it does help justify my raising my position size to 4-5% of account.
bangkoker
Well-known member
- Messages
- 314
- Likes
- 17
Having a scout around the internet on probability:
It appears that heads and tails (assuming 50% of course) having a run of 10 in a row:
is a probability of 0.976 in a 1000 (or 1 in a 1000 of a 10 in a row run)
using 56% (well 44% as that is chance of losing each time in theory) I work it out to be 0.27ish or circa 1 in 4000 of 10 in a row.
Doing 5 losers in a row, it seems to suggest that you would expect that to happen about 4 times a year (with the assumption that the system is used 250 times in the year).
Interesting to work out
It appears that heads and tails (assuming 50% of course) having a run of 10 in a row:
is a probability of 0.976 in a 1000 (or 1 in a 1000 of a 10 in a row run)
using 56% (well 44% as that is chance of losing each time in theory) I work it out to be 0.27ish or circa 1 in 4000 of 10 in a row.
Doing 5 losers in a row, it seems to suggest that you would expect that to happen about 4 times a year (with the assumption that the system is used 250 times in the year).
Interesting to work out
bangkoker
Well-known member
- Messages
- 314
- Likes
- 17
Should that count though Tom', the system didn't fail as it told you not to trade as nothing had happened by 1:00pm. Of course there was a chance that the long entry would have worked if US market woke the FTSE up.
I am struggling with the automation of my system. I really want to use 'One Cancels Other' Orders and I am very keen to move to break even once 10 pts up. I think I may need to learn MT4 (as I believe it can do such things).
I need to do similar for other styles of trading and the only solution I can get at the moment is for one order to close and a new one to open (which isn't great as I would have to pay a second spread).
Oh well will ponder on. Going to read up on MT4, sure there must be stuff on internet.
bangkoker - you are corect, thank you. In fact, the system says no trade to be opened after 1pm, so this knocks one loser off the perfomance record. I guess this emphasises the psychological aspect of trading - it was not a system error, it was the trader's.
