Value betting
If you want to win in the long term you have to be able to identify value betting opportunities and have the disciplie only to bet where you believe value exists.
What is a value bet?
A value bet is one where you believe the odds available to you on a particular outcome are more than a fair reflection of the probability that the bet will win. So if you think there is a 50% chance that a team will win a match then you should only bet if you can get odds of greater than evens (2.0 in the decimal format)
Why is value so important?
If you don't stick to a discipline of only betting where you believe value exists, you will likely lose money in the long term – at best your bankroll will be slowly eroded away. The reason for this is that the payouts you would receive on the winning bets will not be enough to make up for the stakes that you would lose on the losing bets.
To illustrate this the following table looks at six consecutive, theoretical bets where we believe the chances of a win are 50% on every occasion. We might be betting on the toss of a coin coming up heads for example:
| Bank before | Probability of win | Odds (decimal) | Stake | Outcome | Profit/loss | Bank after |
|---|---|---|---|---|---|---|
| £100 | 50% | 4/5 (1.8) | £20 | Win | + £16 | £116 |
| £116 | 50% | 4/5 (1.8) | £20 | Lose | – £20 | £96 |
| £96 | 50% | 4/5 (1.8) | £20 | Lose | – £20 | £76 |
| £76 | 50% | 4/5 (1.8) | £20 | Win | + £16 | £92 |
| £92 | 50% | 4/5 (1.8) | £20 | Win | + £16 | £108 |
| £108 | 50% | 4/5 (1.8) | £20 | Lose | – £20 | £88 |
The first thing to note is that our prediction of a 50% chance of winning appears to be right with three wins and three losses. But look at how the bank has reduced to £88 after just six bets. This has happened because the odds available were worse than evens. Continuing this example through, we could expect to see our bank dwindle to zero after another 44 bets.
But what if we'd been able to get odds of 6/5 (2.2), i.e. better than evens? In this case, after 50 bets we'd have doubled our bank! If only it was that easy. Sadly bookmakers aren't in the business of giving money away and in the majority of cases the odds will be stacked in their favour so that we would probably be offered worse than evens in a 50% probability scenario. Instead we have to look very hard for occasions where we think the odds are more generous than they should be.
Identifying value
There are various ways of identifying value, some far more sophisticated than others. At the most simple extreme, some people are very knowledgeable about the markets they are bet on and do a good job of instinctively knowing when the odds seem a little better than they should be. At the other extreme, complex ratings systems are used to determine the probability of a certain result and this is then used as a reference point for the odds offered by the various bookmakers and on the betting exchanges.
Ratings systems also vary in complexity and will potentially take in a wide range of quantitative and qualitative inputs. Apologies if you are not familiar with these terms. Quantitative or quant based inputs are statistical data such as the number of corners to each team in a football match or the records of the jockeys and trainers of the horses in a race. Qualitative data is usually far more subjective and might be, for example, what the manager says in a pre–match interview, a weather forecast, or perhaps news that a player has picked up an injury in training.
Experience and knowledge of your chosen sport are of most value when trying to make sense of qualitative data. Generally it helps to look at both quant and qualitative inputs in deciding on the probability of an outcome. For example if a player is out injured (qualitative) you might want to look at the contibution that player has made to the team recently – goals scored, assists, etc. (quant) – and how the team performed last time he was out injured – did they still manage to win as many games? (quant).
To illustrate the next step in identifying value, let's create a very simple rating system for football, based purely on one quant input: league table points. Tottenham are at home to Portsmouth and we want to come up with what we think are will be fair odds for the game.
| Team | Games Played | Points |
|---|---|---|
| Tottenham | 6 | 4 |
| Portsmouth | 6 | 13 |
Let's say the rating system looks at the point ratio and converts this to home win and away win probabilities with a fixed assumption that there is a 30% chance of a draw. The calculation we'll use is as follows:
| Result | % probability | Fair odds (decimal) |
|---|---|---|
| Tottenham win | 4 / 17 x 70% = 16.5% | 100 / 16.5 = 6.0 |
| Portsmouth win | 13 / 17 x 70% = 53.5% | 100 / 53.5 = 1.87 |
| Draw | 30% | 100 / 30 = 3.33 |
Of course there is no way you would want to use this very basic rating system – it just isn't reliable enough – but hopefully it is starting to get you thinking in terms of identifying value and converting probabilities to fair odds. In reality you'd want to work with more reliable indicators such as possession, shots on target, home and away records, quality of opposition when winning and losing, injuries and so on.
The actual odds offered on Tottenham were in fact below 2.0 and on Portsmouth around 4.0, which a clearly more realistic. With a more robust rating sytem the margins of difference between your fair odds and the odds quoted by the bookmakers will be much smaller.
Ratings systems are considered in much greater detail in the section titled Betting Systems.
No value? – leave it or lay it
Once you have established a system for identifying value that you trust then we suggest you apply it consistently and if you don't find value, don't make the bet. Over time you will be able to monitor the level of success you are achieving and start to refine the system to give even more reliable results.
Due to the bookmakers over–round you will find that there is very often no value in any of the outcomes. The betting exchanges often (but by no means always) offer more competitive odds and tight spreads between the back and lay prices. This means that if you find that value doesn't exist for backing a particular team or runner it naturally follows that there could be value in laying it instead. But you'll need to factor the spread (difference between back and lay prices). Whether backing or laying, don't forget to factor in the commission that is usually taken by the exchanges from the winnings you make. You may remember this from an earlier page:
True odds = (quoted odds – 1) x (1 – commission rate) + 1
At time of writing, Mansion were only charging commission on the lay side, not for backing.
| Previous - Betting as investing | Next - Pricing of odds |