Winning by not losing....

When examining a trading system, a lot of emphasis is placed on the "positive expectation" of the system. Lets say you have a system where you bet $100, and 50% of the time you make $50, the other 50% of the time you lose $40. Sounds spretty good, you have a positive expectancy of $5 every time you play (0.5*50 - 0.5*40).

The big surprise is if you play this game, each round betting your full purse, you are headed for ruin.

The reason is that by betting this way, you are going to achieve the Geometric average ($0.949), NOT the Arithmetic average (1.05).

The 2 returns are related by the following formula

Arithmetic Average - (Vol*Vol)/2 = Geometric Average

The key takeaway is that in any system where you want to compound, the Volatility (vol drag...) is crucial - you need it as low as possible.

I was reminded of the concept when I came across this thread in twitter.

This guy also does an amazing job of breaking down the concept using games

I am not a maths guy, so dont shoot me down on the maths front.....but I think the concept is important for any investor...

 

Kelly Criterion Bet Calculator: Optimizing Bet Sizes

 

thats it....

 

Hey @Gav

You have sort of scrambled things up a bit here.

Position size properly and this absolute $ payoff ratio is an absolute winner.....I would (have) trade it with appropriate position size until the cows come home.

But if what you are trying to get at is that +50% win and -40% loss creates a geometric average below 100% (90% I think, not 94.9%) then you are correct. A negative geometric return will guarantee ruin given enough bets are taken. (law of large numbers)

But winning $50 or losing $40 on $100 is +50% and -40% only when your bank is $100. if your bank is say 10K and you are making many $100 bets with +$50 or - $40 payoffs than you can get your geometric average high enough to keep you safe from a realistic probability of a sequence that will ruin you - this is exactly the math behind a casinos' business. Small edge, bet limits sized to casino capital and large numbers of bets.

 

Yep, that was the point I was trying to make, its a certain road to ruin if you go all in on every bet - appropriate position sizing will work.

Apologies for the confusion...