How do I maximize profits from being right 80%+ of the time?

[anley]

The problem with not using stops is that for most people it means they haven't answered 2 critical questions -

1. Where will I be wrong if price moves against me, and
2. How much will I lose if wrong

So when price goes against a position (for those not using stops), fear, greed and of course hope enters the equation and those together are a lethal combination.

BY all means don't use stops, but at least have the sense to work out before the trade at what sort of area you'll throw the towel in.

Remember everyone, successful speculation is first and foremost a game of managing risk, never managing profit - that can come only after the risk angle is sorted. Do it the other way around and you're a gambler but you've convinced yourself you're a speculator/trader, and therefore your time will come..............

 

[the hare]

Just out of curiosity, does your fluctuation remain stable within say 1SD of the long term average?

If I look at the win rate, then its reasonably stable (generaly within +/- 1 standard deviation). If I break it down into long and short trades, the varience is much greater.

Thats probably down to the methodology I trade. Entries are basically random, so over a short period of time such as a month, I tend not to get a 50/50 distribution of long and short entries.

There are odd months when you see extremes, so maybe 80% of all entries just happen to be long, and if that coincides with a rising market, then obviously I do well, or if I'm unlucky, I find myself predominantly long in a falling market and performance is not so good.

Clearly that sort of nonsense leads to greater short term varience. When performance falls outside statistical limits its generaly due to a skew in the disstribution of longs and shorts. The problems actually getting worse for me, because over the last couple of years I've placed more emphasis on trend bias for a couple of the systems that I trade. Initially the trend bia was used to modify position size, and more recently, to actually determine trade direction so I'm tending to generate far more skewed distributions of longs v shorts. The upside is the trend bias improves performance, but the varience in win rate, average win, average loss etc is probably greater (I havnt really tracked it that closely cos I'm busy with other stuff)

I know this thread is purely 4 d lulz, but the question of how to maximise returns for a given win rate is rather interesting int it ?

Off the top of my head, a high win rate is going to result in high numbers of consecutive winners, interspersed with low numbers of consecutive losers. Perhaps some reseach into staking plans that capitalise on those sorts of distribution ?

The problem is, whatever he does, he'll increase the drawdown. If I'm being cynical, but honest then the best strategy, is to reduce the 3% per day by a factor of 10, reduce the corresponding drawdowns by the same amount, and then get decent amount of capital under management.

 

[robster970]

If I look at the win rate......

Very interesting - will take this up with you elsewhere. More questions later.

 

[the hare]

Taking just your win rate, over any time range you choose to use, (all time, weekly, monthly etc.) it can only at any one given time have a single value. Your win rate so far this week is a single value based on all the trades you have taken. It isn’t a range of values. Your win rate over this current month currently has a single value – not a range of values.

I agree. However we are deriving statistics from an underlying entity which is non stationary. If I choose to break down stats into months, and see the win rate percentages below over a period of 6 months, then its pretty obvious that the results over the first 3 months are signifcanty different to the results over the last 3 months

70,65,68,52,48,51

I dont need any clever stats to tell me that, its bleedin obvious. If I look on a chart, and see that the first 3 months are nicely trending and the last 3 months are choppy then that posssibly indicates that my sytem does well in trending markets. If I've designed the sytem to exploit trending markets I can take some comfort from that.

If the markets been going up over the first 3 months, and the win rate stats for long trades are higher than for shorts over that period, that gives me even more confirmation. Hey it could just be random chance, but the more confluence I get the happier I am

You are however right, the average is still the average, but the temporal changes are equally important.

I track statistical limits to every single element of each system that I trade, and if something falls outside that limit, thats the time to stop. Each of my sytems have probably over 20,000 trades each over the lat decade, so waiting for a shift in long term average to tell me something is wrong isnt really practical is it ? or am I missing something ?

 

[TheBramble]

Bunny, is someone asked you what your win rate was from the beginning of time right up to this point in time (now), would you give them a range or a single value?

 

[the hare]

Bunny, is someone asked you what your win rate was from the beginning of time right up to this point in time (now), would you give them a range or a single value?

I would probably give them a single value.