Invented in 1956, the Kelly Staking Plan seems to have many admirers. I am not one of them.
The Kelly Staking Plan calculates what percentage of your current betting bank should be staked on your next bet. Simplified, the formula is:
(Your assessment of true probability x odds including stake, eg. 2/1 = 3.00)
(odds including stake - 1)
Multiply the answer by 100 to give percentage of betting bank to bet. This system has flaws and in my opinion is basically unsound as the denominator here should be a factor of the true odds not the available odds. Secondly, it massively overstakes, as the table below shows:
|
|
TRUE |
|
Best odds |
Kelly Stake |
|
|
price |
|
offered |
|
as % of bank |
|
|
|
|
|
|
|
|
|
A |
3/1 |
|
8/1 |
|
15.63% |
|
|
B |
2/1 |
|
3/1 |
|
9.96% |
|
|
C |
1/1 |
|
5/4 |
|
9% |
|
|
D |
1/2 |
|
1/1 |
|
33.33% |
|
Anyone staking 9% of their bank on a 5/4 chance with a 50% chance of winning or one third of their bank on an even money chance with a 66.67% chance of winning needs the telephone number of my psychiatrist urgently.
Most professional and serious gamblers will have many bets running concurrently and therefore have no idea what the size of their bank will be after their live bets are known. In the case of ante-post bets they may not know for months. If I have a betting bank of 300 points and place 20 ante-post bets totalling 100 points should I be reducing the size of each bet of similar value as my remaining bank reduces? Of course not. These systems would only be of value if we knew the result of each bet before deciding on our stake on our next bet.
Confused? Worried? Don't be - help is at hand for people who have only a few bets a week and wait for each result before re-investing. Welcome to the ANSELL STAKING PLAN. (ASP)
The percentage of your betting bank to stake on the next bet is:
(% Expectation - 100 ) x True Probability
5 (Average Risk Factor)
Expectation = true probability x odds including stake (for 3/1 chance offered at 8/1 = 25 x 9 = 225)
Using the Kelly examples above the ASP stakes would be as follows
|
|
TRUE |
|
Best odds |
|
ASP |
|
|
price |
|
offered |
|
% of bank |
|
|
|
|
|
|
|
|
A |
3/1 |
|
8/1 |
|
6.25% |
|
B |
2/1 |
|
3/1 |
|
2.22% |
|
C |
1/1 |
|
5/4 |
|
0.63% |
|
D |
1/2 |
|
1/1 |
|
4.44% |
It gets even better. As in the case of investment, business people have a different attitude to risk to others.. When choosing your investments your independent financial advisor must ask you this when completing his ‘Know Your Client’ form. If you like mining shares your risk factor is likely to be ten (maximum). If you can only tolerate safe fixed interest investments your risk factor is likely to be one (minimum).
Obviously most people fall somewhere between the two. In my formula (I am thinking of applying for a patent right now!) I have chosen a denominator of five. Most daring punters could choose three or 2.5 and those with a weaker constitution could choose 7.5 or more.
Incidentally I would not back the 5/4 chance that I gave a 50% probability of winning as the expectation is only 112.5% and to allow for a margin of error and a reasonable minimum level of expected profit I do not bet at below 115% expectation except on Asian handicap markets where 110% would be excellent.
I will not be using the Ansell Staking Plan myself but it contains all of the criteria that I use when deciding the size of my stake on any particular bet. If, like me, you have many bets a day you should use the following guidelines.
1. A separate betting bank is essential. I recommend a bank of at least 200 points with 300 being ideal. You should NEVER have to fund this bank again.
2. For me bet A would be a 12-point bet, bet B would be six points (slightly bigger than ASP), bet C would be no bet and bet D would be nine points.
3. Your minimum bet should be one point and your maximum bet should be 20 points. A maximum each-way bet should be ten points each way.
4. Maximum bets will be extremely rare - maybe three or four a year. People who advise double maximums should be castrated, as should those who invent staking plans after the results of their tips are known to produce illusory inflated profits.
5. Every three months you should look at the size of your betting bank and reassess the amount that you bet per point. If your betting bank has increased by 50% then your stake per point should now increase by 50% for the next three months. After a further three months recalculate again in the same way.
6. You can have as many bets running concurrently as you please within the constraints of your bank size.
On another matter, to date only one person has got last month’s lateral thinking problem correct. I find it hard to understate the gravity of this situation and am leaving the problem open for another month. See www.inside-edge-mag.co.uk for the question and email your answers to: progambler@winonsports.com
*The above article first appeared in the November issue of Inside Edge magazine and has been reproduced with their kind permission. Readers wishing to subscribe to the magazine can do so by visiting Inside Edge subscriptions. For further information on Tony Ansell's own website, visit www.winonsports.com.