How much should you risk?

Visaria said:
kut2k2, don't you just divide the ev by the profit factor to obtain the kelly fraction?
More...

News to me. In any event, it doesn't work for the six-outcome scenario in the OP. As I said earlier, GAT's answer is the best so far.
 
botpro said:
I think kut2k2 means his new formula he had posted here:
http://www.elitetrader.com/et/index.php?threads/a-new-kelly-formula.291307/
More...

Not really. Thanks for the mention and the link but the new Kelly formula gives an answer of 84%, which is farther off the mark than GAT's answer.

The new Kelly formula is the best estimate of the Kelly fraction in the public domain, which makes it fairly valuable. But there is a proprietary formula which is even better. Then there's GAT's method: using Excel's Solver routine to numerically solve the optimal geometric growth equation.
 
12,205.63% but actually I'd bet max 20% (Depending on the base capital) since 10 successive losses isn't improbable, even if it's a 3&more-sigma event. I know I am conservative but I'd still get big returns out of this and able to stay in the game, whatever happens.

Guys that would bet more than 20% are just pre-broke gamblers. #Drawdowns.
 
Last edited:
K-Pia said:
12,205.63% but actually I'd say 5% since a drawdown a 10 could shows up, even if it's a 3&more-sigma event. I know I am conservative but I'd still get big returns out of this and able to stay in the game, whatever happens.
More...

Big returns? Hardly. Here's what your 5% gets you:

((1+.05*.20)^.35)*((1-.05*.15)^.25)*((1+.05*.12)^.2)*((1-.05*.06)^.15)*((1+.05*.5)^.04)*((1-.05)^.01)

== 1.00282504

In other words, your account grows on average 0.2825% per bet.

And here's what GAT's 75% gets him:

((1+.75*.20)^.35)*((1-.75*.15)^.25)*((1+.75*.12)^.20)*((1-.75*.06)^.15)*((1+.75*.5)^.04)*((1-.75)^.01)

== 1.02868794

In other words, his account grows on average 2.87% per bet..
 
kut2k2 said:
Big returns? Hardly. Here's what your 5% gets you:

((1+.05*.20)^.35)*((1-.05*.15)^.25)*((1+.05*.12)^.2)*((1-.05*.06)^.15)*((1+.05*.5)^.04)*((1-.05)^.01)

== 1.00282504

In other words, your account grows on average 0.2825% per bet.

And here's what GAT's 75% gets him:

((1+.75*.20)^.35)*((1-.75*.15)^.25)*((1+.75*.12)^.20)*((1-.75*.06)^.15)*((1+.75*.5)^.04)*((1-.75)^.01)

== 1.02868794

In other words, his account grows on average 2.87% per bet..
More...

No way.
GAT goes broke.
And I retire right now.

Ps: I've changed my mind (Edited my post) while you made your calculations. Sorry.

I'd make 1,4% returns for every % risked.
Knowing I'd invest 13%.

And full kelly gives me 36%.
We know full kelly's insane.
 
Last edited:
Ok, it's an optimization problem.
How was the value 75% found?
Iteratively (ie. trying 0..100, step 1 or smaller), or does there exist a math equation?
 
No rational risk taker would invest 75%.
P(Loss)= 40%
So you almost go broke once every 3 trade.
P(2 Consecutive Losses) = 16%
So you go broke once every 6 trades.
But only once.

Trading isn't the same as math olympiads.
 
You must log in or register to reply here.
Back
Top