Help Required: Calculating 'equity curve' from R:R & Win-Rate

[RedGreenBen]

I took the easy approach (for me, not the spreadsheet) when calculating a running balance over N trades with a given % win-rate and constant risk:reward and constant %risked... I used N lines of individual trades and randomised the wins/losses to be consistent with the win-rate.

I was surprised by the variability in the outputs when I re-randomise the win/loss distribution, although this in itself was a very useful lesson.

My question though is that there must(?) be some relatively simple way to create a curve with a mean and (say) an envelope of 2 standard deviations. Rather than trying to re-invent the wheel then any guidance that I could be pointed towards would be appreciated.

I may need to include more detail to give you a fighting chance at this but I didn't want to bore you all with 3 pages when someone immediately knew both the question and the answer.

Thanks in advance,
Ben

 

[Rhody Trader]

My question though is that there must(?) be some relatively simple way to create a curve with a mean and (say) an envelope of 2 standard deviations.

A mean and StDev of what?

 

[RedGreenBen]

A mean and StDev of what?

Capital after trade N.
(thanks for taking time to read/consider).

Ben

 

[shadowninja]

Random Equity Curve Simulator of a trading system. Learn it before you trade

 

[Rhody Trader]

I don't think mean/stdev are measures you can reasonably use here. You wouldn't really come up with a distribution of final values. You'd have a curve or line. Maybe it would make sense to measure the slope of that line.

 

[RedGreenBen]

I don't think mean/stdev are measures you can reasonably use here. You wouldn't really come up with a distribution of final values. You'd have a curve or line. Maybe it would make sense to measure the slope of that line.

Shadowninja - Thanks for link, this applet does exactly the same as my spreadsheet (wish I knew about it first!).

If you press the 'generate' button then you get a new curve each time. If you had no life then you could sit there and press it a million times and get a distribution of y values for each point on the x-axis. With that you could generate a mean/median curve and some confidence intervals (which is what I am trying to do).

I agree that standard deviation would be a bit of an approximation (but a tolerable one for the back of my beer mat). If there is no shortcut then I will have to replace the 'press button a million times' with a looped bit of code but I am trying to avoid anything outside the spreadsheet.

I *think* this is valid, although you can of course (constructively ) tell me I am speaking out of my a**e

Ben

 

[shadowninja]

Erm. Set it to produce 100 lines, then!!

 

[Vaco]

what are you hoping to get out of this and how will it help your trading?

 

[Vaco]

Random Equity Curve Simulator of a trading system. Learn it before you trade

I got my 3 year old to draw a line accross a page a few times and it looked just about the same as wht that generator produces

 

[RedGreenBen]

I got my 3 year old to draw a line accross a page a few times and it looked just about the same as wht that generator produces

The first thing I get out of this is that is does look exactly like a 3yo has drawn a line across a page a few times, that is actually a very interesting observation (trust me, it is ).

Secondly, after a bit of coding, I now have a curve that projects my account over the next 12 months together with confidence intervals. For anally retentive people like me that is of some use for planning (yep, I know the approximations/idealisations) and lots of use to amuse my depressingly analytical brain.

Where was Spanish89 when I needed someone with a similalrly rigorous approach to planning?

Ben