Wednesday 27 February 2013

Beating The Bookies At Formula 1 Betting

Being a jockey is the only thing any member of my family has ever been too tall for.  So, while my great-grandfather never got to be jockey*, I was brought up watching horse racing.  So I know sports betting is a mug's game, because the bookies have an in-built advantage because they get to set the odds and aren't ever going to set them in the punter's favour.

So I don't bet, except for the Grand National, which is enough of a lottery to make a mockery of the odds.

When @psychmedia showed that grid position was strongly positively correlated  with final race position (http://blog.ouseful.info/2013/01/30/f1stats-visually-comparing-qualifying-and-grid-positions-with-race-classification/ and following posts) my interested was piqued.  I have to admit my interests were more ... acquisitive than purely scientific.  When I've asked previously to suggest who to bet on to win a race, I've always said "back whoever qualifies on pole**".  The problem, in terms of betting, is that, while you risk less, you also get less of a reward because whoever gets pole position has their odds slashed.

So, I was thinking about how to improve the rewards without risking too much more.

There are some obvious points during a race weekend when you can place a bet based on what has gone before:

1) After Friday practise.  The obvious problem here is that a lot of teams use Friday practise for long runs, to give their reserve driver some track time, and to test new parts.  Who goes fast on Friday might not tell us anything about who goes fast on Sunday.  The obvious plus is that there should be better odds on offer.

2) After Saturday practise.  The cars are more likely to be in race-ish trim and it's not likely that the reserve drivers will be in any of the cars.  The odds are probably going to be shorter at this point.


3) After qualifying.  The cars will definitely be in race trim because of parc fermé rules, and despite possible fuel and tyre issues, the cars are at a point where position is positively correlated with final performance.  Of course, because of this, the odds will be at their shortest here.

At what point can a conscientious, risk-averse punter best place their money?

That's something that's testable.

If I put some hypothetical*** money on the fastest driver at each of these points, by the end of the year (or 20 replicates), I should be able to see if there's any difference.

Now to some more practical details:

I will be using the William Hill website values to give me the odds I would get.  William Hill were chosen because they're also where I put my National bet on.  I will only use simple bets i.e. x to win at a/b with no additional frills****.

I will put the same amount of "money" on for each bet.  I'm trying to decide whether to put on £1 or £10 because with £1 the result will be clearer but with £10 the differences should be more obvious at the end of the season.

Now to list the foreseeable difficulties:

a) Short time windows for putting the bets on, particularly for putting money on at point 2.  I am one of those people who still doesn't have a smart phone so if I am away from a computer (which will happen because of fencing), I might not be able to put on a bet for that time point.  I might require help from friends.

b) The odds are not calculated from 0 each time.  The bookies aren't stupid*****, they will take into account previous form.  So for instance, even if this season had entirely new regulations, Vettel would be at a shorter price going into the Australian GP than his team-mate.  Form through the season is also going to be weighted so if Racer X wins the first 10 grand prix, his odds are going to get shorter so I will win less money betting on him.  You also have local weighting - William Hill is a British bookies so is more likely to have fans that put money on Hamilton to win because he's the local lad.  As bookies don't want to lose money on popular bets, this is likely to cause them to reduce his odds as well.  This may mean the values I get at the end of the season don't actually reflect the money I could have won if I was betting at a bookies based on Mars.

Comments, suggestions, things I've overlooked?

~~~~

*don't worry, he managed to find a different job with horses.

** except Mark Webber who appears to be cursed.

*** hypothetical because I am not made of money.

**** remind me to write something about how a lot of the additional frills are actually bad for the punter's chance of winning.

***** if anyone ever comes to you with a betting scheme that involves the bookies being stupid, it is a bad plan and will not work.  Run for your life.

Friday 22 February 2013

Apparently my hypothesis was wrong

The hypothesis I was talking about in the last post was that the reason the 2012 Formula 1 season's racing was closer was due to the lack of blown diffusers, more particularly, I thought that this was because the diffusers and the way they worked (http://en.wikipedia.org/wiki/Formula_1_2011#Rule_changes) aided smoother drivers, or hindered the less smooth.

Now I admit this was mostly based on a gut feeling that Mark Webber was a much better driver than his 2011 results showed, and that, with the removal of the blown diffusers, his results would be closer to Sebastian Vettel's.

Unfortunately, I can't actually prove my hunch with numbers, because, Webber's average placing in 2012 in races where both drivers finished, was actually slightly lower than his position the year before, in comparison to Vettel (on average he was 1.9 places behind Vettel rather than 1.8 places). I blame being Grosjean'd a few times. However, at the same time, he's finished ahead of Vettel more times this season (five times as opposed to twice). I therefore decided that I needed more evidence to back up my theory. Or, you know, prove it at all. I needed another team where one driver is general regarded as smoother than the other and the drivers haven't change between 2011 and 2012.

Bring on the McLarens.

Thankfully, the McLarens actually do bear out my theory, even with Hamilton's various misfortunes.  There's a swing from Hamilton being -0.71 of a position behind in 2011 to being 3.3 of a position ahead in 2012.  The other thing screaming at me is that really, McLaren's reliability issues have really hurt them.  Both Red Bulls finished 16 times, both McLarens finished 10 times.

Unfortunately, I'd planned to use the difference between the Ferraris as a baseline but it turns out that Massa is less capable of coping with a terrible car than Alonso is.

There would be spreadsheets to back this up, but, unfortunately, I've not been able to figure out how to post them into Blogger.  Any help is gratefully received.