Episode 191: Matt Glassman

11 thoughts on “Episode 191: Matt Glassman”

1. Hi all:

   First, thanks to Andrew and Nate for having me on the show. I had a blast. Although I have regularly corresponded with both of them for years, I had not spoken to or seen Nate in over 10 years, and I had never spoken to Andrew. It was surprisingly enjoyable to do so, even conditional on how much I was looking forward to it. Thanks again, guys. #21stCenturyFriendships

   Second, I’m happy to talk all things political strategy or political analysis here in the comments this week. I’d rather not argue about the merits of the presidential candidates or their parties, because I don’t think those discussions are particularly fruitful in this format, and I’m not sure I have any particular expertise relative to anyone else you might query. But I’m happy to talk of the politics I know. Being a sharp political observer is like being a good cardplayer; the more you can set aside your personal opinions and biases of people and their ideas, the better the analysis usually becomes.

   Third, we barely scratched the surface of what there is to say about the similarities between analyzing politics and analyzing card games, and we even skipped over some of the basics. One of the things I totally forgot to mention on the show was the importance of logically applying statistics, even very basic statistics. For example, in poker if it’s checked to you heads-up on the river, and you know you have the best hand 90% of the time, it does not at all imply that a bet is +EV; what you need to know/estimate is the % of the time you have the best hand if you are called. This is an extremely common error, and an essential theoretical point.

   If you apply the wrong theoretical logic to statistics, they are often worthless, or worse.

   The classic political equivalent to this poker example is the logical fallacy that if 70% of your constituents hold a certain position, it can’t possible hurt you electorally, as their representative, to adopt that position. That belief, although widespread, is totally false. You need to know the distribution of the preference among your electoral supporters and your electoral opponents, as well as the intensity of the position on both sides. If the 70% who support the position don’t really care about the issue, but the 30% who oppose the position are single issue voters on this issue, you could easily lose a lot of votes if you adopt the position, since there’s a large chance current electoral supporters of yours who are opponents of the position will desert your coalition, and little chance those who aren’t in your coalition will join based on the position.

   This very problem absolutely haunts many politicians on the issue of gun control, and enrages many liberal voters. Polls consistently show large majorities for certain gun control measures, such as expanded background checks. But very few Americans are single-issue, or even impassioned, gun control voters; they support background checks but don’t really use gun control in formulating their vote choice. On the other hand, a large number of those who oppose background checks are single issue gun rights voters. Consequently, adopting a pro-background check position can cost a politician the desertion of a lot of former supporters, without the corresponding gain of new voters who were formerly electoral opponents. All within the context of the vast majority of constituents favoring the background checks.

   I discussed this in detail several years ago in a blog post about Obama and Catholic support after the HHS contraception decision (http://www.mattglassman.com/?p=2708). If you need a more nuts and bolts example, or are just further interested in the mechanics or implications, I recommend it.

   Finally, one thing I strongly believe but didn’t relate on the show is that, in a democracy, ideas, interests, and the arguments made for and against them should carry the day, not the credentials of the person stating them. While expert deference is an important and reasonable analytical tool, it is often used in politics as a method of silencing voices and ending discourse. As I often tell my students, there’s no a priori reason to believe they are right and you are wrong simply because they published it in a book. The same is true for things said on podcasts. To the degree my arguments were convincing, great. But if they were not, I’m all ears, and welcome being told.

2. Hey guys,
   I know this was a little different type of episode for the show and thought you might like to know that I for one really enjoyed it. What a fascinating guy with a fascinating job! Thank you so much for having him on.

   • Thanks so much for the note! I’m a big fan of this episode and am happy at the feedback we’ve gotten.

3. What Matt writes above about the 70-30 example, where the 30 care more than the 70 about the specific issue and so get their way, leaving aside the example of gun control, in general that’s a strength of representative democracy rather than direct democracy – the representative tries to weight and deliver the things which are most important to different groups of people rather than let the majority overrule the minority on things it doesn’t care about anyway.

   +1 to it being a great episode. I’m pretty active in the NVG presidential election thread but I learnt plenty but was also pleased to have one or two of my theories about political certainties confirmed. 2 months before the UK had its EU referendum there was still a market operating on “UK to hold an EU referendum by 2020” – given that the law was passed, the date was set and the campaign was underway the odds were now at 1/100 that we would have one but it was still a lot more certain than that (it was pretty much betting against nuclear war in the next 2 months), so there was basically free money available but to actually close out the market requires a real ton of people willing to come along and make the last bets at 1/100 and people don’t want to do that (when I tell people I took those odds they look at me like I’m crazy, till I start to explain that’s the kind of odds insurance companies take on their “bets” all the time).

   The one exception I would say to politics betting favourites being undervalued, is where you are seeing large bets from big betters following conventional wisdom, and on the other side lots of small bets made by the same people who are actually going to be casting their vote in the election you should listen to the voters – that was the betting pattern in the EU referendum and also when my father was unexpectedly elected to parliament (the odds for the previous MP hanging on were 1/5 as late as a month before the election but Ladbrokes suspended betting with a week to go after they saw the number of bets being made in the consituency) – at least that pattern applies in a country like the UK where basically everyone gambles (people say they don’t gamble but you ask more questions and they say things like “well obviously I put a bet on when I go to the horseracing track”), and you don’t have issues with bettors being an unrepresentative sample.

   • Lots of interesting stuff in there, thanks!

4. Yeah, great guest, could have listened to the conversation for hours.

   One of my pet likes about poker interviews is when guests/hosts start saying things about “Poker players do [this]”, and “Poker players are like [this]” – of course you can’t easily capture the totality of a group as diverse as poker players so these statements are very revealing about their maker’s self-image and ideas of what poker is and isn’t . They’re never _wrong_ but they’re always only part of the story.

   • Well said!

5. This episode was phenomenal. I didn’t realize how dumb I am about politics prior to listening to this episode.

   • Agreed on all counts!

6. Outstanding episode: that was a really interesting and sophisticated conversation.

   • Thanks, much appreciated.

Comments are closed.