What game theory taught me about last night’s presidential debate

30 September 2020, 0430 EDT

Last night’s debate might go down as one of the greatest in recent memory, and I am prepared to die on that hill. It was ugly. But it was also raw, unfiltered, and honest. It was thin on policy substance which is why I think the majority of my Twitter feed thought it was a shitshow.[1]

A few things stood out which I think are worth mentioning, even on an IR blog. The performances last night got me thinking a lot about game theory. I am not a game theorist, and if I were, I would not subject you to a model of this year’s electoral strategy by Teams Trump and Biden. (I doubt the game theorists could even pull it off.)

Still, some of it can instructive here because although a lot has already been written about Trump’s and Biden’s divergent strategies, much of it is mediated by the virus. So it has been hard to say until now how much of that divergence is the result of their competing theories of electoral victory, and how much is due to covid–19. But in the midst of all the chaos, there are few clues.

Trump’s theory of victory seems to rest on base turnout. His approach to the debate was to bring a more serious, slightly more presidential version of the debate performances he delivered in 2016. The main difference is he now has nearly four years of governance which must be defended. Absent the virus, I suspect he would be on much stronger ground because rightly or wrongly, voters would tend to reward economic growth.

Biden, for his part, believes this election will be won or lost in the marginal districts. So much digital ink has been spilled over what lessons can be drawn from the past four years, but I’ve long thought much of it has been wrong: viz that Trump’s 2016 is not a sign of a declining America, nor a populist rejection of elite malfeasance.[2] Suffice it to say, Biden was focused less on policy substance and more on appearing presidential. I noticed he looked directly into the camera many times and said “you” as though he was speaking to you and me. Clearly 2008 made an impression Unless I missed it, I didn’t notice Trump break the fourth wall once.

How that plays when voters who are not political scientists watch is an open question. I doubt any of this will show up in the polls. The one battleground tracking poll looks fairly consistent with all the national and state level polling. We know this already but it bears mentioning that most opinions on this year’s race are set, and have been since this time four years ago.

Having written all that, there is cause for deeper concern. As @jennifervictor tweeted, we are sliding below weak democracy into competative authoritarianism. The stability of the polls seem to confirm that. The CBS battleground instant poll gave Biden the edge by what my back of the napkin looks largely consistent with national polling. Put another way, most of us are not changing our minds. And we do not even seem to be listening to anyone with whom we might agree, even they are on their on our side. I’ve not yet read Bruno Maçães History Has Begun, but after tonight’s debate, I am all the more intent on doing so.

Then again, maybe Tocqueville was right to observe wonder over the crisis of presidential elections.


  1. Actually, I watched without social media and went back after. I deactivated Facebook until I finish a book manuscript early next year and removed Twitter from my phone. It’s helping.  ↩
  2. More precisely, it is not entirely either or both of those. I want to avoid editoralizing here, but as many of contributors and readers of the Duck already know, institutional design matters. And the path dependent effects of institutional change obscure how we might best evaluate any given poltical phenomenon. Had the GOP instituted a different set of rules in the 1970s after the Democratic party’s reforms, we might be looking at a President Rubio or Bush, drawing difficult conclusions about what the election taught us. Or, as my colleagues who study elections tell me, a small shift in vote choices in a few swing states four years ago would have us all evaluating President Hillary Clinton’s covid–19 response.  ↩