Foreign Exchange trading problem - risk reward calculation

BillH

New member
Joined
Nov 22, 2011
Messages
5
I trade the Forex markets and record all my trades with the amount risked, amount gained etc. There appears to be a correlation between the success rate of trades (how many are positive out of 10), the amount gained on a sucessful trade (the risk reward ratio) and the amount of the capital risked, and I dont understand why!
Here is the example using hypothetical amounts:-

A trader has a trading account with £1000 in it. He is a consistent trader and for each sucessful trade he makes he doubles his investment. He wins 4 out of every 10 trades he makes. So if he was to risk only 2% (£20) of his capital on each trade he would lose 6*2% = £120 and win 4*4% = £160, so his net gain is £40 for every 10 trades he takes. OK, he is keen to see his capital grow faster so he compounds his wins, but as a measure of safety he also compounds his losses (not sure if the terminology is right here so I'll explain). He starts with £1000 and he knows he is only going to win 4 out of every 10 trades, to keep it simple lets assume he losses 6 in a row then wins 4 in a row. First trade risk is 2% £20 and its a loss so the capital reduces to £980, second trade risk is still 2% but now 2% of £980 i.e. £19.60 and its also a loss so the capital reduces to £960.40 and so on for the first 6 trades. At the end of his 6th trade the capital has reduced to £885.84. He now hits his winning streak of 4 consecutive wins and compounding those wins (this time he makes 4% on each one because his risk reward is always 2 times) at the end of his 10th trade he has increased his capital to £1036.31.

So far so good, now here's the bit I dont understand.

If you use the same calculations but increase the amount risked I would expect the size of the captital at the end of 10 trades to just keep increasing but it doesnt. It peaks at 10% and then falls off again as follows :- risking 9% on each trade givesa capital size after 10 trades of £1100.97, risking 10% gives £1102.00 and risking 11% gives £1100.98.
Interestingly if his sucess rate increases and he wins 5 out of 10 trades the optimum peak percentage risk increases to 25%, for 6 out of 10 its 40%, 7 out of 10 its 55%.

I have modelled this in excel so I'm fairly sure its accurate and the calculations are correct but I just dont understand why the is a peak and then a drop off. I'd be very grateful is someone can explain.

Many thanks

Bill
 
Jeff, thanks for the reply. I don’t think there is anything wrong with my logic, perhaps I have not explained the problem particularly well. Let me have another go.
The trader does not know what sequence the wins and losses will come in, all he knows is that over time he has a level of proficiency that ensure he will win ‘x’ trades out of 10. In the example I gave he wins 4 out of 10. He also has a level of proficiency that will mean that his winners produce twice the profit on any trade than the loss resulting from a loser. In Forex trading this is possible because you can set a limit on your losses (in this case 2% of capital) and also set a point at which to close the trade ( in this case 4% of capital or 2 x the amount risked), the real skill is being able to predict which trades will win as well as win twice the amount risked. Because he realises that he could get all of his loses in a row he adopts a money management strategy that will protect his capital. That strategy is based on only risking a fixed percentage of his capital on each trade, in the example this is 2%. This gives him protection in the event he has a string of losers as in the example below (using this method even having 50 losses in a row the capital only reduces to £364.17)
trade no
outcome
capital before
capital after
1
loss
1,000
980.00
2
loss
980
960.40
3
loss
960
941.19
4
loss
941
922.37
5
loss
922
903.92
6
loss
904
885.84
7
win
886
921.28
8
win
921
958.13
9
win
958
996.45
10
win
996
1,036.31
But it doesn’t matter which sequence the winners and losers come in, the overall result is still the same as in this example (same starting capital, same no of winners and losers, same end result)
trade no
outcome
capital before
capital after
1
loss
1,000
980.00
2
loss
980
960.40
3
win
960
998.82
4
loss
999
978.84
5
loss
979
959.26
6
win
959
997.63
7
win
998
1,037.54
8
loss
1,038
1,016.79
9
loss
1,017
996.45
10
win
996
1,036.31
The important thing is that the win rate is 4/10 or 40%. I agree if you knew when they were going to occur then you wouldn’t risk anything on them, you would only trade the winners but as you say that would be absurd.
OK, so we have established that with a consistent win rate of 4/10 and good money management you can both protect and grow your capital over time. Where it gets interesting and reaches the bit I don’t understand is when you start to increase the % of capital risked on each trade.
Again, being consistent and increasing the amount risked on each trade to the same amount (if you risk 4% its 4% on every trade), over 10 trades you end up with a capital sum of £1064.93. So far so good, this is what I would expect – if you are successful and risk more the return would increase. And this proves to be correct as you increase the percentage risked per trade the end capital after 10 trades grows. BUT when you reach 10% risked the capital peaks and if you risk 11% and onwards it decreases. So there is an optimum amount to risk per trade – in this case 10% as shown below.
4 out of 10 trades
% risked
final balance
2
1036.31
4
1064.93
6
1085.52
8
1097.89
10
1102.00
12
1097.95
13
1092.94
14
1086.00

Interestingly if you run the same numbers for a more successful trader who wins 5/10 trades the optimum amount risked is 25% and for 6/10 its 40%, 7/10 its 55%..
So my real question is not so much the process of calculating the returns etc it’s “why is there a peak” and interestingly why does the peak increase in linear amounts for a greater success rate.
There are three variables, Success rate – expressed in no of trades won out of 10, Percentage risked – a percentage of the capital and the Risk Reward – the amount you win as a ratio of the amount risked. There must be some relationship between these that dictates the optimum amount risked per trade but I don’t know why. If you do please let me know.
Many thanks in advance
Bill
 
Denis, Thanks I'll have a read through the thread. Good to see I'm not alone!

Jeff, Thanks also for your patience, I dont think I explained it clearly in the first place. Hey if it takes time but you can come up with the answer I'm OK with that, Its an interesting problem, I'd be very happy just to understand why.

Looking forward to some interesting feedback.

Bill
 
Bill, have a look at this:
http://www.mymathforum.com/viewtopic.php?f=24&t=24897

A bit like yours?

Denis, a bit like mine yes but also not really! It would seem that the other problem will have a solution it just has MANY permutations and would take a long time for me to model, I'm not a mathematician but I suspect someone with a command of complex algebra (if thats the right term) could come up with a formula, maybe not - maybe the only way is to construct a matrix.

My problem is different in that I know that the capital vaule peaks at a certain level dependant upon the amount risked, what I dont know is why it peaks. I was hoping by posting here someone could say "of course the capital value increases up to a point but then falls off because of ................". Compounding your winnings and losses based on a fixed percentage risk makes sense (beginners are always advised to risk no more than 2% as it allows for a longer string of losers without damaging your capital too much), always closing your winning trades for a minimum risk reward makes sense (in this case 2:1) as this means that as long as you win more than 3 out of 10 trades you will make money over time. I am a semi professional trader and I make good money doing this, most people fail because they dont understand the money management, as my performance has improved my ability to risk more per trade has allowed me to increase my capital risked per trade. Being an inquisitive person I decided to see what would happen if I risked greater amounts, I did this in excel (not with real money, never practise with real money especially if its your own) using my performance track record as a basis and I discovered this strange peak that I wasnt expecting.

Now I'm just interested to understand why?

Bill
 
Jeff, Interesting indeed. The maths is beyond me now, its 35 years since I did A level maths at school in England. My son is currently doing Further Maths at A level and off to study maths at university next year so I'll run it by him. I will delve into a bit of revision to see if I can get a handle on it and I'll also look at the references your son gave. If you can find a laymans way of explaining it that would be great. Thanks for giving up time on your thanksgiving its much appreciated.

Bill
 
Top