Staking System

valueman · Oct 13, 2004

i've used the kelly staking plan/4 multiplied by a class of race factor which can bring the bet up to kelly/2.5 for many years and have been very successful. I use this for fixed odds Horse race betting mainly Uk racing but have noticed I have become more specialist and am doing fewer bets hence less turnover which leads to less profit.

i am considering moving to a base of kelly/3 which for certain bets will be 0.6 multiplied by kelly - is this too aggressive?- i have about 10 bets a week for 6 months a year and show a profit of 20% to stake year in year out.

my fear is a losing streak on this plan will waste my bank away too quickly.

Fishhead · Oct 13, 2004

Are you using strictly a gambling bankroll for the kelly/2.5 wagers?

valueman · Oct 13, 2004

Yes - strictly gambling bankroll for all bets. At present it is kelly/4 for a class D race which is multiplied by 1.25,1.5,1.75 for grades C,B & A respectively. i have found I have more success in higher grade races over the years and was considering moving to the base of kelly/3 with the same grade factor multiplication.

Fishhead · Oct 13, 2004

You should be fine then, no?

Faceless · Oct 13, 2004

I made some research (read post by guys with higher IQ than mine) about Kelly and what I found was that optimal Kelly divider is /2, but for most people using /4 will save money if estimation was wrong.

Here is another Kelly calculator, which which also take care of commision, if you are using Betfair or some sites like that. scroll down

http://koti.mbnet.fi/bets/tools/

valueman · Oct 13, 2004

I think Kelly/3 with my grade of race multiplicator will be fine. It refines bets so I am staking more on a grade A race at sandown than a grade D at fontwell.

With the respective differences in commission you'd think more people would use sporting options & betdaq than betfair but I end up almost entirely using betfair as the odds/volume are almost always better.
Be very interesting if Mansion put up big liquidity on UK horse racing - 0% commission - yummy!

Faceless · Oct 14, 2004

Valueman,

What I have read also is, that bettor should use bigger divider if betting mainly big underdogs and vice versa. If you estimate that some horse has a 55% chance to win, but a true % was 50, mistake is not so bad with high Kelly/3. Giving 15% to a 10% horse with Kelly/3 might be a mistake in a long run, if you a betting just one horse in a race. Of course if you bet 5 horses in one race with each having 15% chance to win, Kelly/3 would be fine.

I haven't seen a mathematical proof about this, but it makes sense to me.

valueman · Oct 14, 2004

I mostly do one horse in a race, occassionally 2. I bet very few favourites, generally basing bets around weak favourites and if I had to pick an average price of a horse I bet it would be in the 6/1 range with an approx 20% chance of winning in my eyes.

Hence there is the possibility I can long runs without winners - last year went 5 weeks without a winner so the importance of the right staking system is paramount.

My kelly/3 will be almost equivalent to kelly/2 with my grade of race multiplier so any thoughts on whether this is too aggressive or not would be appreciated.