Sharpe Ratio

[Adamus]

Very Sharp

 

[windsurfing_stew]

> Basically I'm wondering what sort of graphical representation would best display the Sharpe Ratio of a particular trade history, i.e. what sort of graph would I have to look at to be able to best guess what the Sharpe Ratio is.

If your equity line looked like a straight line sloping diagonally up then the sharpe ratio would be high indeed (ie a very good strategy). Basically the steeper your equity line slops up the higher the sharpe ration, the more wavey and jumpy it is, the closer to zero your sharpe ratio will be. Not suprisingly many choose to optimise their sharpe ratio.

Another useful ratio is the kelly criterion = av / (sd)^2 = sharpe ratio / sd. This basically gives you an idea of how much you may want to place per trade, as well as being another useful choice to optimise against. It will favour similar equity curves to sharpe ratio but applies even more penalty to variation.

Also, in both sharpe and kelly, people tend to set the av term as being the av return of the stock - the risk free return. Eg if your av return is $1000 but similar money in the bank could have earned $300 then you should use $700 not $1000.

 

[Adamus]

hi stew, thanks for the input. I thought Kelly Criterion was only applicable for bets with only 2 possible returns - win or loss. but i found it useful to help understand the situation - it's explained quite well in Ralph Vince - Mathmatics of Money Management, in the Introduction. I haven't actually got much further than that due to sleep deprivation and the book's strong sedative effects.

Actually I have been listening to N.N.Taleb a lot recently and he's right that Black Swan events that don't make it into your sample data will make a mockery of your Sharpe Ratio.

 

[intradaybill]

I thought Kelly Criterion was only applicable for bets with only 2 possible returns - win or loss.

You probably meant just two payoffs for win and loss. This is true but the formula used is an approximation anyway. This is not the main problem I think. It is the fact that effective use of Kelly assumes a stationary win rate and avg. win to average loss ratio. These values change all the time and that makes this technique inapplicable for most. The author in the paper below explains this well and you do not have to pay for a book

http://tinyurl.com/2c7kng2

 

[Adamus]

You probably meant just two payoffs for win and loss. This is true but the formula used is an approximation anyway. This is not the main problem I think. It is the fact that effective use of Kelly assumes a stationary win rate and avg. win to average loss ratio. These values change all the time and that makes this technique inapplicable for most. The author in the paper below explains this well and you do not have to pay for a book

http://tinyurl.com/2c7kng2

Thanks for the link. Looks like another insomnia cure for me. I'll give it a read.

 

[windsurfing_stew]

All good.

Kelly Criterion should definitely be used with a scepticism, however can still be better than nothing. Most people who calculate it, half it before using it. This is termed a 'Half Kelly'. In other words if your systems suggests you bet £1000 then you might try a Half Kelly of £500.

 

[Ambrose Ackroyd]

Haven't heard of this one.

Is it one of Cornwell's new books >?

 

[alexander]

All good.

Kelly Criterion should definitely be used with a scepticism,

Is this a polite way of saying "stay away"? I think so. Stay away of Kelly.

 

[Blaiserboy]

You probably meant just two payoffs for win and loss. This is true but the formula used is an approximation anyway. This is not the main problem I think. It is the fact that effective use of Kelly assumes a stationary win rate and avg. win to average loss ratio. These values change all the time and that makes this technique inapplicable for most. The author in the paper below explains this well and you do not have to pay for a book

http://tinyurl.com/2c7kng2

Thanks for posting this........!!