In part 1 of the Horse Spy horse racing basics we looked at using real probabilities to determine each horse’s percentage chances of winning. Now we’ll focus on creating a staking strategy to get the most from your race analysis, odds and probability.
Horse Racing Betting Basics
There are many different ways to place a bet on a horse. When people bet on a horse some will have a dedicated betting strategy; most will not. In essence, staking strategies range from wreckless to ultra-strict.
In the Horse Spy post what type of person are you, we highlighted those that are most likely to profit from horse racing. Those that regularly profit from horse racing are also the most likely types that have a dedicated staking strategy.
A staking strategy is a system to get the most out of a betting bank. The easiest way of describing this is by comparing a sum of money that a City Trader will use to make money:
The City Trader has £1000 and needs to invest some of it to make it grow. He does this by speculating and trading. He won’t plump all £1000 on a trade to make another £1000, there’ll be a strategy behind it to grow it gradually. If the nominal % increase over a year is 10%, he can expect the total to be £1100 at the end of the year. As the balance grows, the size of the trade grows in direct proportion to the balance. His commission also rises proportionally to the accumulated growth.
Successful horse racing betting works in the same way. You wouldn’t stake an entire £1000 betting bank on a horse at evens. This would be utterly wreckless. The wise would stake a proportion of the bank based on its probability of winning in order to preserve the balance.
Three examples of common staking strategies:
- Level stakes: This is the easiest and perhaps the most common of horse racing staking plans. This is also a common strategy for those that don’t have a set betting bank. Betting £2 on every win bet, or £1 each-way, is a strict system. It’s main drawback is that is does not include any focus on the probability of a horse winning and does not scale up or down depending on the bank balance.
- Percentage stakes: This is a highly effective horse racing staking strategy that depends on the size of the betting bank. Professionals will use this system as will semi-pros and intermediaries. If the betting bank is £1000, setting a fixed percentage stake at, say 2%, will mean that the first bet will be £20. If that selection won the next stake would be proportionate to the new balance. If the £20 won at 2-1 (£60 payout), the balance becomes £1060; so 2% for the next bet will be £21.20. If the first bet lost and the balance was £980, the next bet would be £19.60. This strategy’s advantages are:
- It will take a very long time to get to zero;
- With success, bets are raised proportionally to raise more profit; and
- It reduces the urge to increase bets when on losing runs.
Common percentages for this type of system can range between 2% and 4%.
- Percentage stakes based on the Kelly criterion: This is a very professional method and utilises your probabilities and prices that you produced in part 1. Similar to number 2 above, this strategy uses a percentage staking strategy, but it is based on the probability and the odds of the horse rather than the total of the betting bank.
Using the Kelly Criterion
The Kelly Criterion was developed by John Kelly in 1956 whilst at AT&T. His calculations were primarily employed to reduce signal noise on long distance telephone signals. It wasn’t long before gamblers, horse racing professional gamblers, sports and financial traders found they could manipulate and use the formulae to their advantage.
The most straightforward formula for calculating a percentage stake based on the Kelly criterion is:
(pb-q) / b = %
p = probability of winning
b = decimal odds
q = probability of losing
If in your analysis you have a horse with a 16% chance of winning and its current available price is 6-1; the probability (p) becomes 0.16; and 6-1 in decimal odds is 7.0.
The losing percentage is 100% – 16% = 84% = 0.84
Substituting these figures into the formula will provide a percentage stake:
((0.16 x 7.0) – 0.84) / 7.0 = 0.04 ( x 100 for %) = 4% of betting bank
The reason this percentage stake is slightly higher is because the odds are higher than it should be for a horse that has a 16% chance of winning.
If the final percentage is a minus, this represents a no bet scenario as the horse has a price lower than its chances.
Similarly to the percentage stakes, the balance will take a very long time to reach zero, but it’s main advantage is that it will give you a reading of just how much of an edge you hold over the market.
To put this into perspective, using the above example a horse priced at 6-1 should have a probability of 14%. Transposing this into the formula:
((0.14 x 7.0) – 0.86) / 7.0 = 0.017 = 1.7%
Spotting the fact that there’s an additional 2.3% in the market is a big opportunity to profit and represents a significant value bet.
There are examples where some people will divide the Kelly % in half to preserve the balance further. This can be a good tactic when initially building the bank and establishing competence in creating odds and prices.
Another good use of Kelly can be used when dutching selections instead of pulling one selection from a card. This is especially useful is you have three horses that all have a very realistic chance of winning the race and you can’t split as to which to back. Between them they have a 50% chance of winning the race:
Horse 1: Price available 4-1, your probability of 18% chance
Horse 2: Price available 9-2, your probability of 16% chance
Horse 3: Price 5-1, your probability of 15% chance
Horse 1: ((0.18 x 5.0) – 0.82) / 5.0 = 0.016 = 1.6%
Horse 2: ((0.16 x 5.5) – 0.84) / 5.5 = 0.007 = 0.7%
Horse 3: ((0.15 x 6.0) – 0.85) / 6.0 = 0.008 = 0.8%
Sometimes it is worth dividing the stakes by three as there are three from the same race being bet. This means the total stake % for the race becomes 1.02%:
Horse 1: 0.53%
Horse 2: 0.23%
Horse 3: 0.26%
Kelly criterion for Laying
The Kelly criterion can be used for laying, although it seldom is. As stated above, if the chances of a horse are over-rated and the price is substantially lower than it should be, there is an opportunity here to lay the selection.
Let’s say there’s a selection at 2-1, or 3.00 in decimal. This represents a 33% chance that the horse will win. But in your analysis it realistically has a 20% chance because of under-rated competitors.
Your analysis: ((0.20 x 3.00) – 0.80) / 3.00 = -0.067 = -6.7%
Here, it’s worth using the 6.7% as the liability for the bet.
If the betting balance was £1000, and 6.7% represents £67, at 1-2 on the lay market on BF there’s a profit of £33.50 on the selection; which is the real difference per £1000 that it has been overbet.
Useful learning point:
When selections are bigged-up by the media and tipsters and subsequently overbet, using Kelly to determine how much is an opportunity to either:
- Search out the under-rated selections at higher prices; or
- Lay the selection with the negative figures that have been over-rated.
Hopefully using parts 1 and 2 of this mini-series will forge a greater understanding of how to extract value from the betting market. More importantly, it also provides an even deeper insight into horse racing as the analysis required makes you become the bookmaker. You will also be seeing through the prices to a point you will see selections over-rated and overbet, and selections that are under-rated and under-bet. To make substantial gains from horse racing it is these under-rated and under-bet selections that will raise your betting bank and give you the biggest edge in horse racing.