In recent articles, I have begun to include A/E figures when looking at potentially profitable systems. It relates to the potential value of bets and is based on looking at the actual number of wins and dividing it by expected wins (A/E = Actual/Expected).

If the A/E figure is above 1.00, it means the horse won more often than the odds expected it to; and if the figure is below 1.00, it means the horse won **less often **than what the odds expected. In other words, you want an A/E value of over 1.00 because it is a sign that your horses are outperforming what is expected of them; you are receiving **value **which is key to earning a long-term profit.

If you have what looks like a winning system with an A/E figure of 0.85 for example, it means the value you’re getting is 15% lower than what the odds say it should be. In other words, eventually, this ‘winning’ system will turn and begin losing as lack of value catches up to it.

**A/E in Practice**

Let’s say you have a system which involves backing horses at odds of 3/1 (4.00). The overall ‘expected’ chance of winning is 25% [(1/4.0) x 100.] In a hypothetical scenario, your system has backed 100 horses and 28 of them have won so far.

When using A/E, we use these figures (28/25 = 1.12) because there are 28 actual winners against 25 expected winners. The A/E value is 1.12 which means you are getting value for money.

Here is data from Richard Johnson’s performances when riding at Newbury. Since 2012, punters who backed all 233 of his rides would have earned over 297 units of profit on Betfair, and an ROI of almost 128%. The overall A/E figure of 1.16 shows that you are also getting tremendous **value **from his entries.

As he has won 46 times and the A/E is 1.16, it means Johnson **should **have won around 40 of his races although you won’t get a nice round number due to the nature of how A/E is calculated. Remember, it is based on the odds of each Johnson entry. It is also based on the SP odds.

Johnson had a bad 2018 at Newbury with just 6 wins from 41. The A/E value of 0.85 shows that horse and jockey underperformed throughout the year. Here is a screenshot highlighting the 41 entries so you can see the odds of each horse.

The average odds of all 41 Johnson rides should be approximately14.16 which would equate to 13.16/1. 100/14.16 = 7.06 ‘expected’ wins against 6 ‘actual’ wins. 6/7.06 = 0.85. When I added up the figures, the average odds were slightly different due to variations when I rounded odds up and down.

You can also plainly see that if you used the Betfair Exchange odds instead, Johnson’s A/E figure would have been higher due to fewer expected winners. In fairness to Johnson, he wasn’t exactly riding favoured horses each week. Over a quarter of his 41 runners were available at odds of 10/1 or longer.

**Using it to ‘Frank’ Systems**

The A/E figures are often fascinating when looking at the records of prolific and successful trainers. The above data is Paul Nicholls’ record in UK National Hunt races since the beginning of 2015. A win rate of 23.04% is pretty good, and you would even have earned a tiny profit from blindly backing his horses, but the A/E figure of 0.92 shows that you are usually NOT getting **value **when selecting Nicholls’ horses.

We can use this data to determine how many winners Nicholls **should **have had based on the odds of each horse.

*570 Actual Winners / 620 Expected Winners = A/E figure of 0.92*

We now know that Nicholls should have trained 620 winners instead of 570 during the above timeframe.

*2464 runners / 620 expected winners = Expected win rate of 25.16%*

His **actual **win rate was just 23.04%, so punters are giving up a lot every time they back a Nicholls horse without performing due diligence. For the sake of completeness, the average odds of a Nicholls horse should be (100/23.04) = 4.34 but according to A/E figures; it is actually (100/25.16) = 3.97. Imagine giving up 0.37 every time you place a bet!

Experienced punters will tell you that there is a hell of a difference between an Evens bet and one at 11/8 or an 8/11 bet and one at 11/10; that is roughly the difference between the odds you **should **get on a Nicholls horse and what you end up receiving.

The **ideal **betting system will offer steady long-term profit with an A/E value well above 1.00 on average.

The above is a prime example of a betting system with real promise. It relates to the performance of jockey Nico De Boinville in non-handicap National Hunt hurdles races. It is a system that ticks every box:

- Profit over a long period
- Consistent annual profit
- A decent number of bets each year
- An A/E value well above average

It seems as if bookmakers continue to underestimate horses with De Boinville on board. A 32%-win rate is pretty special in its own right, as is a profit of 44% across a period of over four years, and a minimum profit of almost 40% each year since 2016.

The A/E figure hasn’t come close to falling below 1.0 with a low of 1.12 back in 2016. The total A/E figure of 1.22 means De Boinville wins 22% more races than he should according to the oddsmakers. The one possible issue is the fast start in 2019; it is probable that his win rate will fall, but the A/E seems likely to remain high.

**Final Thoughts on A/E**

In summary, if punters latch on to systems with a good A/E figure, the bank balance of bookies will send them to the A&E! It is a useful measure of value. If you have a betting system, don’t focus solely on ROI or strike rate. Make sure the A/E figure is above 1.0 or else you are not receiving value, and your system will not win in the long-term.

You might have noticed that the A/E stat is not something you see regularly; either online or in the media. All the focus seems to be on strike rate and ROI. While such data is useful information, it does not offer an indication of value for money; A/E does. It is also important to calculate A/E based on a fairly large sample size, or else the figure is meaningless.

The average odds of all 41 Johnson rides should be approximately14.16 which would equate to 13.16/1. 100/14.16 = 7.06 ‘expected’ wins against 6 ‘actual’ wins. 6/7.06 = 0.85. This is not explained very well in layman’s terms. How you get from the 41 race odds averaged to the 0.85 number should be explained better. It is not very clear.