Lay Kelly Staking Plan
This plan is based on so called Kelly criterion. Its main idea is to determine the chance of winning a bet based on history of previous bets placed on selections chosen on certain conditions.
The pure mathematical odds of a specific selection to win the game are almost always represented by the selection's price on BetFair Games.
The probability of Hand 2 winning with these cards is roughly 100/1.96 = 51%.
By combining these odds with your own statistics of betting in this game, you can work out an efficient staking plan.
This trigger example is based on the article in The Staking Machine. Do not be scared off by the formulae given there. The two important parameters you need to pay attention to are:
Average Odds - this is the average price of your lay bet in the channel. Suppose you lay on the favourite hand in Poker at preflop. In that case the average odds will be around 2.3 - 2.8, as that is normally the average price of the favourite in those circumstances. But if you choose to apply Kelly plan to other channels and selections, the average odds will be different. So the first part of your homework will be to calculate or roughly weigh up the average price at which you will lay.
Strike Rate - this is the ratio of winning bets to the total number of bets placed. Say, if you placed 100 bets and 85 of them have won, the Strike Rate will be 85/100 = 0.85 or 85%. The more often you win, the bigger will be the strike rate. Usually it also means that the average odds are quite high as well. Figuring out your initial strike rate is the second part of the homework. The easiest way to do it is to run X-Feeder in the Test Mode with a trigger that lays according to your strategy, for around 100 games. Stop the channel. Then click on the "Load" link and load the last 100 games.
Then open the "View Variables" window and type in the following formula:
Then press the manual refresh button to let X-Feeder calculate that variable. It will show the strike rate that you can use for these triggers.
Multiply 0.6593 by 100 to get the strike rate of 65.93%.
Now you are ready to open the trigger file. There are five parameters that you can change (two of them should be changed if you will use a different trigger for laying). They can be found in the top of the trigger window, among constants.
divisor - it is used to adjust the size of each bet. The higher is the divisor, the less will be the stake size. You need to find out the optimal value by a series of experiments (in Test Mode of course).
avg_odds - the average odds. You need to key in the initial value for this parameter. Later on, after 30 games with bets (the number of games is determined by the variable g_played) the program will start calculating the average odds based on real-time bet history of the last 30 betting games (a betting game is a game where a bet has been placed).
s_rate - the strike rate. As with the average odds, type in the initial value of the strike rate. After 30 games with bets have passed, the program will keep calculating the latest strike rate based on the last 30 betting games.
g_played - this is the "depth" of the history based on which the average odds and strike rate will be calculated. It should be in the range of 1 to 30. It is advised to keep it between 10 and 20 for credible results.
There is also the first trigger that sets a user variable base. Change its value if you want bets to be based on a fixed amount of money and not on the current bank value.
How to apply Kelly Plan to your betting system.
The trigger in the example lays on the favourite in the second round of Poker, if that hand does not hold a pair. But you certainly can replace it with your own laying trigger. To do that, modify the parameters of the last trigger in the block, called "laying on some selection". You can change any field of this trigger and its conditions, but remember to keep the "Amount" field unchanged:
Please take into account that at some point the trigger may stop placing bets due to the Expectation becoming negative. It will mean that either the average odds are too high or the strike rate too low. In that case, if you want to keep betting (though it actually means you shouldn't), you can clear the W/L/Z history and restart the triggers.