Numbers, Not Teams

One advantage of work travel is the opportunity to read during flights, and on my latest trip I took the opportunity to re-read Jack Moore's 1996 book The Complete Book of Sports Betting, A New No-Nonsense Approach to Sports Gambling. 

It's a read more for US sports and readers, and it's a little dated, but it does explain how lines and spreads came about, how they work, and emphasises what I have said here many times, which is that when betting, you are not betting against the bookmaker, you are betting against other market entrants. 

Jack Moore writes:

The line, on any given game, is in no way an expression of the line maker's opinion on how the two teams will fare. Rather, it is the line maker's expert opinion on what number will sufficiently induce half of the collective betting public's money to come in on the underdog and and half on the favourite. The line maker is interested in what the public will do, not in what the ballplayers will do.

Later in the book, Jack Moore clarifies:

Keep in mind that it is not the number of bets that come in on one side or the other that matters but the sum total of those bets; thus in the extreme case, a hundred $10 bets are offset by one $1,000 bet.

It's something else I have written about in the past, and it's very annoying that Pinnacle keep tweeting out useless 'bet share' numbers, presumably thinking that their followers are stupid enough to pay such meaningless numbers any heed. No doubt some of them are. We need the total amount being bet too, but that's not free. Here is some analysis based on the money bet on NBA games:

I can't be sure, because it was a few years ago that I first read this book, but this may have been when I realised that betting the 'dime line' (i.e. Pinnacle's 1.952) on 50/50 bets, winning just 51.23% of bets would result in a profit. 

Given that my grandmother would be expected to pick winners at a rate of around 50%, and she's been dead for over 30 years, this didn't seem to be a hugely impossible task.

It's not as easy as it might seem of course, but as explained above, you're not betting against a sportsbook with otherworldly powers of prediction. You are betting against other members of the public, and in some markets more than others, the public aren't too smart. 

Should you require evidence of this, just look at the number of idiots following tipsters on Twitter. 

The goal is to take a contrarian approach, or to use the American expression, to fade the public.

While the following quote has its problems, the gist of it is correct:

Let’s start with a simple premise: The public loses. Not every time, but over the long haul being on the same side as majority of bettors isn’t a sound strategy. Why? Public bettors tend to rely on gut instinct, overvalue recent performance and usually gravitate toward favorites, home teams and overs. By taking a more contrarian approach, you’re capitalizing on public bias and taking advantage of artificially inflated numbers. As an added bonus, you also place yourself on the side of the sportsbooks.

As previously stated, the sportsbooks don't have a side. Their goal is to balance the money so that they win whatever the outcome of the event. Our goal is to identify markets where the money has made the sportsbooks adjust their lines and prices too far.

Bet numbers, not on teams.  

So if one assumes the sharper books know their stuff, and thus a significant move indicates pressure on them to change their 'accurate' prices, does this mean there is value to be had? Certainly in the image above, it would appear so, at least in the NBA, although the sample size was small. 

When it comes to football, I pretty much limit myself to the Draw these days, and I can tell you that if you backed the Draw in the English Premier League in every match where it closed with a 5% or more drift from its 'opening' Pinnacle price, you'd have an ROI of 24.6%.   

Don't try this on Chelsea home matches though. When the Away team shortens by up to 5%, jump on the steamer and back the Away team - 129 bets and an ROI of 34.8%.