Wednesday, 25 March 2009

MLS (Most Logical Statistics)


Whilst it is undeniably true that the standard of MLS falls far short of that of the major leagues in Europe, it does have a few advantages from an investment perspective.

One is that the MLS season starts now, just as the English season starts to wind down and silly-season starts with teams in mid-table just going through the motions. Players have one foot on the pitch, the other on the beach in Spain.

Another is that tracking results and updating team ratings from the English and Scottish leagues (11 divisions and 202 teams) is an incredibly time consuming task. With just 15 teams in MLS, the number of games is much more manageable.

Also MLS attracts nowhere near the amount of interest that the Premier League, Serie A or La Liga do, and thus if one does ones homework, there could be opportunities.

One final advantage is that since I have absolutely no interest in any of the teams, and there is limited TV exposure, I can make my investment and not feel like I have to watch the game unfold. I can just click my balance in the morning and watch it leap up…

The aforementioned ratings have, it has to be said, been something of a disappointment, in that I haven’t yet found a way to make them pay. There are promising signs from the lower reaches of the Football League, but when out-of-form bottom-rated Team A beats in-form top-rated Team B, one can’t help but think that the whole thing is a big waste of time.

But a new league starting up affords the opportunity to apply some lessons learned. One thing I will have the luxury of being able to track is not just the result, but shots, shots-on-goal, and goals. I am thinking that a 1-0 win is overly rewarded in my rankings, and that perhaps shots, or shots-on-goal, (or both), should be given more weight. 

I read somewhere that one goal is scored for every ten shots, so it will be interesting to see if this is true. I would also like to include corners but the MLS web site doesn’t include them, (strange for a nation that loves its statistics), so if anyone knows of a website that does, please let me know.

Looking at the results from the opening weekend, and there are some interesting stats.

Chivas USA 2 Colorado 1 (Chivas USA had 6 shots, 4 on-goal; Colorado had 17 shots, 4 on-goal).

Dallas 1 Chicago 3 (Dallas had 20 shots, 1 on goal; Chicago had 11 shots, 6 on-goal).

San Jose 0 New England 1 (San Jose had 16 shots, 5 on-goal; New England had 7 shots, 2 on-goal).

Should my ratings reward the above winners for their efficiency, or should the ratings reward the losing teams for having more shots, which in the long run ‘should’ result in more goals?

I am going to experiment with a rating adjustment that ignores the final result, but takes into consideration shots, shots-on-goal, and goals. And if anyone can help with corners, I’ll throw them into the mix too.

My thinking is that, not unreasonably, the odds are based on results, but because goals are so rare, these odds can be all too easily skewed, but when odds are wrong, there is opportunity.

Incidentally, Baseball records results in the format Runs - Hits - Errors so for example:

Boston Red Sox 3-11-1
New York Yankees 2-8-2

Only the 3-2 is important from the result perspective, but the Hits count is a useful indicator for future reference too. Perhaps football could adopt a similar ‘box-score’ idea: Goals – Shots-on-Goal – Shots – Corners.

Chivas USA 2-6-4-x Colorado 1-17-4-x would give a much more complete picture.

3 comments:

Anonymous said...

Hi Cassini - A quick google search brought up this: http://soccernet.espn.go.com/match?id=260860&cc=5739 so to complete your stats: Chivas USA 2-6-4-2 Colarado 1-17-4-0

Curly

Cassini said...

Thanks Curly! Now to make money with it all!

strugar said...

Hello! I have just discovered this blog, and I like your approach, hope to follow that through the season!
As you say "And if anyone can help with corners, I’ll throw them into the mix too.", I know the site where you can find corners, it is www.en.7m.cn; it seems complicated, but once you get with it, you can find a plenty of useful data; for example, you can find stats for match Houston Dynamo vs Columbus Crew here: http://data1.7m.cn:13000/GoalData/en/232500.shtml
Cheers! :)