Too many choices can lead to unexpected outcomes
Bill James is a sort of living legend among baseball writers. Mr. James is the creator of “sabermetrics” which fuses statistics and baseball as seen in the movie (and read in the book) Moneyball. This model-driven approach to building a baseball team is not so different from our model-driven approach to investing. We both seek solutions that remove subjectivity and bias.
In 2001, Mr. James reflected on a rather odd case where the way the rules were written and applied drove an extremely unlikely result. The case involves Rafael Palmeiro’s ludicrous Gold Glove victory in 1999. The Gold Glove award is presumably given to the player who exhibits the best defensive/fielding skill at that position. In that year, Mr. Palmeiro won the Gold Glove at first base even though he only played 28 games at first base, with the rest being played as a designated hitter. As Mr. James and many others pointed out, it’s hard to be the best defensive first baseman in the league when you are not the best defensive first baseman ON YOUR OWN TEAM.
It was a bad joke, perhaps the worst award choice in baseball history. Many people had theories about why it happened. Some said it was because the voters (managers and coaches) weren’t really paying attention. Some said it was because Mr. Palmeiro had a fantastic offensive season (he hit .324 with 47 homers) and that distracted people from the fact that he no longer played first base well (or often). Some said it was because there was a dearth of good candidates.
Mr. James saw the problem as one of mathematics.
To quote:
“The larger point, it seems to me, is that a badly designed voting system will fail sometimes, no matter who votes. The Gold Glove is decided by what could be called an unconstrained plurality, meaning:
- A voter can vote for anybody.
- If the top vote-getter gets 15% of the vote, he wins, the same as if he had received 80%.“
You might already see where this is going. Still, it’s worth the big finish. Mr. James continued (and please remember once again that this was 2001):
“A voting structure like this is an open invitation to an eccentric outcome. If the United States were to use a system like this to elect the President, the absolutely certain result would be that, within a few elections, someone like David Duke, Donald Trump, or Warren Beatty would be elected President. If you can win an election with 15% of the vote, sooner or later somebody will. An unconstrained plurality vote gives an opening to someone or something who has a strong appeal to a limited number of people.“
OK, do you see that? Bill James predicted that DONALD TRUMP could be elected President if you had a voting structure where there’s an unconstrained plurality. He predicted this 16 years ago. And then, what happened in this election cycle? The Republican Party had a WHOLE BUNCH of candidates (roughly the same number as first basemen in the American League in 1999. And Trump moved to the forefront by getting an unconstrained plurality, first in the 15-20 percent range, then higher, then higher. No matter where you stand on the Trump candidacy, it was this math that paved the way, and Bill James saw it many years ago.
Bill offered suggestions for how they could improve the Gold Glove voting:
- “Establish some loose statistical limits on who is eligible for the award. No one should be eligible for the Gold Glove at first base unless he plays at least 500 innings at the position.
- Have a panel of experts-not a different group, but a sub-group of the current voters-review the candidates, and trim it to a list of perhaps five candidates.
- Get weighted ballots-three points for first, two for second, one for third.“
The Gold Glove award architects did not take Bill James up on his suggestions. The Republican and Democratic parties, though, might consider taking Bill’s suggestions to heart.
(The above article is abridged from Joe Posnanski’s article found here: Back from the future)
Polygraph
While we are running a tangent from the investment world (because this week markets are doing little new), we offer this interesting ‘experimental publication’ that shows how aggregating data can be used to drive interesting output:
Market Update
The market faced some weak economic data Monday but chose to view it as a positive in that it delays future interest rate hikes. Also a factor in trade early in the week was month-beginning rebalancing of portfolios. The first trading day of the month of May delivered a +0.8% gain. Weak global economic data did not receive such a positive review Tuesday. Poor reports from China and the UK along with a reduction in Eurozone GDP outlook pressured stocks to a -0.9% day. The cautionary tone continued Wednesday with additional weak global economic readings. Stocks struggled to a -0.6% dip. Thursday saw a flat day as investors awaited Friday’s monthly employment report.
The report came in Friday morning with new jobs lower than expected. The weakness could have been read as favorable to those worried about rising interest rates. However, investors seemed to take a modestly negative view of the report coming on the heels of other subpar reports earlier in the week.
Warm wishes and until next week.