Is it Just Luck?

As racing and betting enthusiasts we are regularly bombarded with pitches for the latest wonder system/tipster/method. Among those pitches will be a mixture of good and bad – some that are absolute ‘must haves’…. and some that are outright scams.

The only real way we can assess whether something is any good, is by a detailed analysis of past results (assuming of course that the past results claimed can be verified – we should never consider buying anything that doesn’t have a checkable, verifiable set of past results!)

Once you are sure that the results are true (recommended by someone you trust, proofed to a reputable proofing agency, freely available during live testing etc etc) there are a number of tools we can use to assess how good those results are and how likely they are to be replicated in the future.

One such tool that I use myself is the Archie test.

The Archie test gives us an assessment of whether the results we’re looking at demonstrate a genuine edge or have been achieved by pure chance (ie a flukey run of results)

To carry out an Archie test we first need to extract certain figures from the results data:

1) The number of bets.

2) The actual number of winners from those bets

3) The expected number of winners from those bets

The first two of those are acquired simply by counting them! (or best of all, letting excel do the counting if you have the results in a spreadsheet.)

The 3rd one requires a little more work because we need to calculate the average odds across all the selections backed before we can calculate the expected number of winners(again if the results are in a spreadsheet, all you need to do is highlight the odds column and the average odds will be displayed at the bottom of the sheet!)

Let’s look at an example of finding average odds and then the number of expected winners those odds imply:

We have 10 bets, the odds of which are 4/1, 5/1, 5/1, 3/1, 3/1, 2/1, 4/1, 6/1,5/1, 2/1

To calculate the mean average we add up all those odds and divide them by the number of bets:

4+5+5+3+3+2+4+6+5+2=39

39 divided by 10 (the number of bets) = 3.9.

So we’ve discovered that the average odds across those 10 bets was 3.9/1 Now we need to convert that into a percentage to calculate the expected number of winners backing at those average odds.

To turn the odds into a percentage we must divide 1 by the odds:

1/3.9 = 0.256 – or an expected win rate of 25.6%

so across those ten bets we would expect 2.56 winners.

I’ve used a small number of bets in that example to make it easy to demonstrate how to calculate expected winners – I wouldn’t advise using any system/tipster/method with only a 10 bet history!

Now we have the data required to carry out the Archie Test.

I’ll use a further example to demonstrate:

We’re looking at a system with the following verifiable results data:

Number of Bets: 1000

Number of Winners: 290

Expected Winners: 250

To get the Archie Score we perform the following calculation: We have an Archie score for this method of 8.53. What does that mean?

Well, it’s pretty blooming good. Archie scores can be converted into a probability number (through a lot of complex finagling which I don’t completely understand!)The probability number predicts the percentage chance that the results were achieved by pure luck.

To save us the hassle of learning how to calculate the probabilities,below is a conversion table showing the relevant probabilities for a range of Archie scores (please remember these are just approximate values!): As you can see, our hypothetical system has  a 0.4% chance of achieving those results by luck – or looked at the opposite way, there is a 99.6% chance that the method has a genuine edge and that similar results will be seen over the next 1000 bets.

I’d purchase such a method!

There are further analyses that can be carried out using an Archie score but I will look at those another time. This one is a very good starting point when trying to get a handle on any method you are considering.

****************************

In my ceaseless quest for good free betting resources, I’ve just discovered that my friends over at Secret Betting Club have come up with a football system based on a free football ratings site. The system can be downloaded by anybody who signs up for their free newsletter.