I had a decent debate on twitter yesterday, with a knowledgeable guy who was adamant that my definition of value was flawed. It all stemmed from my BORE DRAW BEATER post, where I show how you can earn a decent amount by backing one goal or more in a game. The basis of his argument was that value needed to be objectively quantified by the use of stats and figures, and that value can never be subjective. This became a debate as I disagreed. I see a bet as value when you think the risk (financial outlay) more than justifies the potential return (potential winnings).
For example, it you walk into a bookies, or log on to one of many betting sites to place a bet, you place bets on outcomes you think are going to happen, despite the fact the offs may suggest there is only a 60% (or any other %) chance of it happening – you have decided that this is good value.
He said that the only way you can get value is through the use of statistical models and databases. While I agree that stats are important when deciding what to bet (and in my case, what to tip), they will always have some form of subjectivity. You might say this is impossible if a statistical model calculates the likelihood of a certain outcome. He stated:
“Value is never subjective. For example, if the odds are 2.0, but you have calculated a probability of 55%, then 10% value.”
I agreed with the above statement as it falls in line with my definition of value – where there is more chance of the bet winning than the odds suggest. My point is where he says “you have calculated”. Using “you” implies subjectivity, whether a statistical model is being used or not. At the end of the day, all models require manual creation. While creating such a model, you apply ‘weightings’ or importance on certain stats and figures that help shape the predicted outcome. If you think the number of crosses is important, you apply a heavier weighting. If you think home advantage is non-existent, you remove it from your calculations. This is where the subjectivity comes in.
Using a statistical model isn’t being objective; it is simply being consistently subjective.
What do you think? What’s your definition of value?