Some Points on Betting

See Adam Gurri for a review of where we are so far.

  • As I told Kevin Grier a few days ago, I usually lose arguments with Tyler, and even with my model of “Tyler” in my head. I would not be very meta-rational, therefore, for me to “attack Tyler” as Bryan wants GMU bloggers to do. Even writing this post, I have moved somewhat in Tyler’s direction.
  • Everyone takes Noah’s point that an individual bet does not necessarily reveal beliefs, because through clever hedging strategies, it is possible to engineer a portfolio that has a different exposure to the original proposition than the individual bet. Everyone also takes Tyler’s point that a big part of your “portfolio” is always going to be non-traded assets.
  • One reading of Tyler is that to a considerable extent, we are all a (mostly) non-tradeable part of our own portfolios. I am betting on my own success, because what choice do I have? What portfolio theory would suggest is that in terms of optimal risk allocation we are all too exposed to ourselves. We should try to find ways to hedge against ourselves that don’t undermine alpha. Which is hard to do. Ideally, we would simply “bet against ourselves” to reduce aggregate risk, but psychologically it is difficult to bet on your own failure without reducing your probability of success. Instead, we need to self-deceive to some extent, to find ways to hedge against ourselves without realizing that that’s what we’re doing. Betting on one’s beliefs is therefore counterproductive in two ways. First, it actually increases portfolio variance because we are already exposed to ourselves. Second, precisely because of the qualities that Alex, Bryan, and Robin laud, betting one’s beliefs makes self-deception, and therefore optimal portfolio allocation, harder.
  • Nevertheless, these drawbacks of betting should often be outweighed by the expectation of actually winning the bet. Therefore, refusing to bet still has informational content for those who observe the refusal.
  • Furthermore, Tyler says that he disagrees with “virtually every sentence in the second paragraph” of Alex’s post, but that still does not follow from my reading of Tyler. As Alex says, because the number of possible states of the world exceeds linear combinations of assets, portfolios cannot reveal a complete set of beliefs. In the presence of non-tradeable assets and psychological limitations on hedging, portfolios a fortiori don’t reveal beliefs. Bets are new assets that are difficult or costly to otherwise synthesize, and therefore add to a portfolio’s capacity to express beliefs. This is true even if we accept that bets have implications for bettor psychology; as long as bets covary with the rest of the portfolio in a unique way, they increase the information content of portfolios.
  • Internal mental accounting” is simply not consistent with truth-seeking plus meta-rationality. Tyler knows this, which is why we need to do some creative reading of Tyler. He is nevertheless right that betting does pay off in terms of pride in betting.
  • Tyler says to Bryan: “You seek arrangements which maximize your glee.” In my view, this is both true and an argument in favor of betting. Bryan’s glee brings me joy and amusement. If it is relatively cheap to make Bryan gleeful, why not do it?
  • Tyler’s fame relative to the other participants in this discussion provides some escape. Tyler has more than 20 times as many Twitter followers as I do, and of course that metric probably overstates my relative influence. Nevertheless, I am frequently bombarded with belligerent or offbeat challenges to my claims. I imagine that for Tyler, receiving many times more than I do, this is somewhat exhausting.
  • Alex says “a bet is a tax on bullshit.” Despite some concessions to Tyler, I still think that’s basically right. However, for the sake of completeness, someone should consider whether bullshit is all bad. Maybe society advances through the adjudication of competing bullshit claims. I call this dynamic the bullshit dialectic. Bullshit claims slow down our convergence to a reflective equilibrium because they present more (poorly founded) possibilities that we must consider, but they might speed up our convergence to the truth because, simply in virtue of there being so many of them, some bullshit claims end up being true. A betting norm probably cows people into parroting conventional views so that they will not be asked to bet, and that chilling effect is bad to the extent that truth-seeking is not merely an activity for atomistic elites.
  • I invite Straussian readings of Tyler in the comments.

9 replies to “Some Points on Betting

  1. Adam Gurri (@adamgurri)

    1. You know him better than I do, but it seems to me that you may be trying to hard to impose coherence upon Tyler’s various views (“that still does not follow from my reading of Tyler” “which is why we need to do some creative reading of Tyler”) while he seems to me to be more the type who collects ideas and theories and doesn’t mind if they don’t fit together like a jigsaw puzzle.

    2. My observation is that Tyler enjoys his vagueness (he recently called out Ryan Avent for saying that he conceded he was wrong in a post which included “I was wrong” in the title) and that whether or not bets reveal beliefs (and i’m not sure they do) they definitely impose a tax upon vagueness. In order to make a bet you have to formulate a very specific claim with a very precise criteria for evaluation. My sense is that Tyler would find this highly annoying.

    The whole exchange seems like a case in point for Alex’s aversion to vagueness and Tyler’s equal aversion to precision.

  2. Pingback: On Bets and Bullshit | Adam Gurri

  3. Eli Dourado Post author

    For Straussians, apparent incoherences in stated views are signals to look for the esoteric meaning. Tyler is probably more tolerant of the non-jigsaw nature of some of the theories he collects than others would be, but I think we’re talking about a relatively limited domain in which having a coherent view is not too hard. That’s why I say do the creative reading. It is a fun mental exercise in any case.

  4. Michael

    I am not sure all bullshit is useless. Many times bullshit is a rudimentary falsifiable claim that is simply not refined into a hypothesis.

    This has always bothered me about the betting logic. People make mistakes in setting up the contracts, which on net help to clarify the debate, but that mis-characterize the coherence of the author’s vision. An important part of the scientific process of falsification is to go back and create nuisances in the testable hypothesis.

    The discussion of weaseling was particularly interesting to me because it matters if our actors are revising their hypotheses or simply rejecting any evidence which does not confirm their priors. Losing a bet is not synonymous with learning.

  5. Pingback: Assorted links

  6. Pingback: On Bets | This is Ashok.

  7. Enrique

    I want to explain why both sides of this discussion (i.e. Tyler and Robin) might be right, depending on whether we look at bets at a “micro” or “macro” level. On the one hand, if we look at a large number of bets on a given topic or question at a macro level, then Robin is definitely right: all these bets in the aggregate tell us something, and this is why prediction markets are so powerful. But at the same time, each individual bet on a micro scale may not necessarily reveal all that much information, for the reasons Noah and others have given (and, I would add, because individuals might make inconsistent bets over time), so Tyler is right when we look at bets on a micro or individual scale …

  8. Pingback: Do beliefs contain useful information? | Fifth Estate

  9. Wonks Anonymous

    Robin Hanson wants to promote betting because it would advantage people with well-founded unconventional views. Or at least views which are not normally made explicit. The more minority your viewpoint, the better the odds you can get in a bet.

Leave a Reply