Friday 20 August 2010

Scaling Down


The Elo ratings are now caught up to date and, for reasons explained previously, the early results are mixed. One good sign is that the draw predictions in the top leagues of England, France and Scotland are hitting at 50%, but it's early days. Another two weeks or so should see results start to improve, but for now the policy not to bet seriously is the right one.

The Budesliga starts tonight with Bayern Munich hot favourites to win the league. If this was a mid-season game, Bayern would be value at 1.68 to beat Wolfsburg, but the latter have made several close-season signings which, of course, upsets the ratings. Cologne also appear to be value at 2.26 v Kaiserslautern, as do Borussia Dortmund at 2.5 v Bayer Leverkusen.

If I were forced to have one bet on the Premier League this weekend, the value bet would appear to be Manchester City at 2.42 to beat Liverpool. Also noted is that Chelsea have the biggest expected margin of victory this weekend at 2.5, and are 1.27 to win at Wigan. Chelsea -1.5 and -2.0 are available on the exchanges for big (four figures) money at 2.05, and are 1.78 giving 1.5. Wigan are bottom rated in the Elo ratings, and not surprisingly Chelsea are top.

In France, Ligue 1 are already into their third round of matches, and Auxerre and Montpellier at 2.32 and 2.27 respectively look value to win at home, as do Nice at 2.48 v Nancy. Lille at 2.22 should also win at Sochaux.

One of the advantages of my summer changes is that the predicted superiorities are now down to a 0.25 goal. No longer are all prices between 0.5 and 1.49 treated the same at 1. This allows me to more easily compare superiorities with the odds available in the perennial search for value. One early observation is that there is a huge difference in the average price of +1 teams and +1.25 teams. This should lessen as the sample increases, but right now the average prices are 2.29 and 1.72. It may be that I will need to reduce the scale yet further.

For interest, the predicted draws in the top leagues this weekend are the games at Birmingham City, Stoke City, West Bromwich Albion, Newcastle United, Kilmarnock, Hannover, Hoffenheim, Hamburg and Mainz.

Hamilton at 3.0 are way out of line when compared with the prices of other +0.5 teams, which probably means there's something going on that is not reflected in the ratings.

3 comments:

Anonymous said...

I can confidently state a team has never been a 2.5 goal favourite away from home in the Premiership. And nor are Chelsea this weekend.

Easiest explanation for your number would be the 8-0 at Stamford Bridge and, i assume, an over-reaction from your model.

Do you do anything to help smooth the extreme results?

Also, something is seriously flawed if you have a team with an expected one goal superioroity at 2.27 to win.

And a 1/4 goal difference should generally be worth approx 6% at popular superiorities.

Anonymous said...

Lat year chelsea got 34 points away from home. This makes them only as good as Blackburn or Sunderland are at home.


How do you adjust for home/away form in your ratings? I'm guessing an away win is worth more? But often teams have very different home/away personalities; like villa last year who were great on the counter attack when away but struggled to break down teams at home - they ended 3rd in the Away league table, 12th in the Home league table.

If I were running an Elo rating system I'd treat the Home team and Away team as separate entities.

Saying all that though it does make me feel more confident when my weekend punts match your 2 goal superiority picks!!

Cheers, Danny

Anonymous said...

Good to see you back posting! Hope the wedding went ok and married life is treating you well. Good luck with the ELO's this season. Edgar Davids at Selhurst Park eh ? Who'd of thought that....

Talkbet