Friday, July 25, 2008

Baseball Pool Advice

Just for fun and to keep busy between now and the start of the NFL season, I am going to conduct another baseball experiment.

If you are a player in a non-profit baseball pool with points based on odds posted for those fools who bet on baseball, let's try the following approach for the rest of the season.

Always take the following teams:

Atlanta
Cleveland
Seattle
Toronto

Always pick against the following teams:

Florida
Los Angeles Angels
Minnesota
Texas

My system is based on the fact that Florida, LA Angels, Minnesota and Texas have better win-loss records than their talent level simply due to good luck so far this year in the way that their runs scored and runs allowed have been distributed in each game. Conversely, Atlanta, Cleveland, Seattle and Toronto have worse win-loss records than their talent level due to bad luck so far this year. I recommend picking against the "good luck" teams and favouring the "bad luck" teams on the grounds that luck is just luck and cannot be expected to continue. It may well be the case that oddsmakers and baseball bettors are fooled by won-loss records and overestimate Florida, LA Angels, Minnesota and Texas and underestimate Atlanta, Cleveland, Seattle and Toronto.

For the statistical basis of my argument that won-loss records are influenced by luck rather than talent, go to:

http://www.baseballprospectus.com/statistics/standings.php

When two undervalued teams or two overvalued teams are playing against each other, I don't have any advice. For example, today Seattle is at Toronto and my system has nothing to say about which team to back.

But, undervalued Cleveland is hosting overvalued Minnesota. I recommend picking Cleveland even giving odds of -190 (or -1.9 on a 1 point scale). My guess is that oddsmakers are undervaluing Cleveland against Minnesota even with Cy Young contender Cliff Lee starting for Cleveland.

I will report back in the first week of September on the performance of my baseball pool-picking system.

In the meantime, I remain rapt to be still ranked in the top 10 at

http://www.crowdpicks.com

Wednesday, July 16, 2008

Have a Good Summer Godivans

With no interesting sporting activity until the National Football League (NFL) regular season kicks off on Thursday, 4 September, I bid my legion of loyal Godivans a temporary adios.

On Thursday, 4 September I will return with a full set of NFL week 1 picks for players in non-profit pools in which you have to pick winners against the point spread.

To get through the dog days of August before the real season kicks off, I recommend reading Professor Steven Levitt's article:

http://pricetheory.uchicago.edu/levitt/Papers/LevittHowDoMarketsFunction2004.pdf

Don’t worry about the heavy mathematics in the article. The key point is that underdogs won 52% of all regular season contests against the number over 16 NFL regular seasons from 1992 through 2007. The underdog winning percentage exceeds 50% 3 out of every 4 years and betters 52.5% 5 out of every 8 years.

So, my twin mottos are:

1. When in doubt, pick the NFL underdog.
2. Never pick an overdog unless the posted point spread shows value.

The all-underdogs approach to NFL picking is well-known. The value that I add for NFL pool players who consult this website is my ability to find the rare underestimated overdog. Over 7 years of picking NFL games from 2001 through 2007, my 52.4% winning percentage is 0.7 percentage points ahead of the all-underdog winning rate of 51.7%.

But, be warned that last year my 47.5% record was even worse than the all-underdog record of 48%.


Also be aware that overdogs have bettered 50% in 3 of the past 5 years. Has the worm turned to favour NFL overdogs against underdogs after the decade of underdog domination from 1992 to 2002?

I will be sticking to mostly underdogs in 2008. I’m treating overdog superiority from 2003 to 2007 as a temporary deviation from the long-term trend. Let’s see what 2008 delivers.


All-Star Baseball


It took 15 innings, but baseball’s all-star classic delivered more evidence that the American League (AL) is the only truly major league. The National League (NL) has not won an all-star game since 1996.

Factoring in the all-star game result, an all-AL approach to picking inter-league baseball contests in 2008 would have generated 34.1 pool points with
  • 1 point awarded for an AL overdog victory
  • 1.x points awarded for an AL underdog victory (with x denoting the posted odds)
  • -1 deduced for an AL underdog loss
  • -1.x deducted for an AL overdog loss
The AL’s 59.3% winning record in 2008 inter-league contests so far is a touch above the 4-year average of 57.5% (including all-star and World Series games). It’s possible that an all-AL approach paid off this year simply due to random variation above the 4-year average. But, there’s at least a preliminary case worth investigating with further monitoring in future years that oddsmakers are not correctly factoring in the AL’s utter and complete dominance over the NL.

I can tell you right now that no matter what teams end up in the World Series and no matter what odds are posted, my advice to baseball pool players will be to take the AL team in all World Series contests.

I remain rapt to be in the top 10 for CrowdScore and W/L picking at

http://www.crowdpicks.com/


Cheerleaders Put the Blue in CFL Winnipeg Blue Bombers

Toy boy Archibald insists that I tease his bro-in-law Doug the Slug, Winnipeg resident and season ticket holder for the Canadian Football League’s Blue Bombers, by including this apologetic link from the so-far-winless Bombers about their cheer squad’s escapades:


http://www.winnipegfreepress.com/local/story/4199449p-4791118c.html

Sunday, July 13, 2008

Baseball All-Star Game Pick

Lady G fancies the American League (AL) -1.4 over National League (or -140 depending on how the points work in your baseball pool)

I usually avoid the cliché “no-brainer” -- unless I’m describing an ex-boyfriend or explaining my correct decision to part ways with an ex. But, picking the AL in the all-star game certainly qualifies as a no-brainer. Leaving aside the 11-inning tie in 2002, the AL has won the last 10 all-star contests to be decided and 16 of the last 19.

The AL has won 7 of the past 10 World Series and 11 of the past 16.

The AL sports a 57% record in a total of over 1,000 inter-league regular season games over the past 4 years.

There haven’t been enough all-star and World Series games to rule out AL dominance as part of random variability. But, the record from the past 1,000 inter-league regular season games provides clear evidence that the AL has a clear majority of the best players in baseball. And, preliminary evidence from this season indicates that oddsmakers may be underestimating AL dominance. Take the AL if you are playing in a non-profit baseball pool that includes picking the all-star game winner.

Thursday, July 10, 2008

Final Inter-League Baseball Pick

If you play in a non-profit office baseball pool with points based on posted odds, you must take

NY Yankees -1.5 (or -150) at Pittsburgh.

See my previous posts on the outstanding record of American League (AL) teams in inter-league play, particularly AL teams visiting National League (NL) parks this year. The posted odds this year did not correctly adjust for AL dominance, particularly when AL teams were on the road.

Check out

http://mikesrants.baseballtoaster.com/archives/1036914.html

for a good summary of inter-league history with the AL winning 57% of all inter-league contests from 2005 on. The AL’s 59% record this year is second only to 61% AL dominance in 2006.

All sporting streaks eventually regress to the average. But, as long as the AL is dominating and preliminary evidence indicates that the oddsmakers are not correctly allowing for this dominance, stick with the AL team in any inter-league game.


I certainly hope my own sporting streak does not revert to the mean anytime soon. I'm as chuffed as chuffed can be to be ranked 5th for season W/L and 9th for CrowdScore at

http://www.crowdpicks.com/

Thursday, July 3, 2008

Inter-League Baseball Results

In my 13 June post, I reminded my loyal Godivan readership that American League (AL) baseball teams have dominated inter-league play over the past few years. And, this year was no different. With one rain date to be made up later this season, AL teams have won 149 out of 251 games so far or 59%.

AL superiority is no secret. What I wanted to investigate was whether the odds offered generally foolish baseball bettors correctly reflect AL dominance. Suppose you were a player in a non-profit office pool with points based on posted odds. Our expectation is that gaming houses set odds to line up an even proportion of betting on each team so that the house profit comes from the 5% commission implicitly built into the odds. If the odds are calibrated correctly, we would expect an all-AL strategy picking inter-league winners (or any other baseball pool picking strategy) to generate net point losses of 5% over a long period of time.

In fact, with an all-AL strategy you would be +34.6 so far this year with one point awarded to an AL overdog winner, the odds awarded as points to an AL underdog winner (e.g., 1.25 for a +125 underdog), -1 deducted from an AL underdog loser and the odds deducted from an AL overdog loser (e.g., -1.25 for a -125 overdog). It’s interesting to note that this breaks down to +25.8 from AL teams’ 57% record when visiting National League (NL) opponents and +8.8 from the AL’s 62% record playing host. In other words, odds favouring AL teams hosting inter-league games turned out more accurate than odds set for AL teams on the road in NL parks.

So, we have preliminary evidence that the fools who bet on baseball may not be correctly valuing AL dominance. Of course, one season of 251 inter-league games is not a large enough sample to draw a firm conclusion. But, I will be interested to follow this trend in future years to see whether the pattern recurs.

Thanks largely to my simple approach of picking all AL teams in inter-league match-ups, I am ranked 6th at


http://www.crowdpicks.com/