Biased Pandemic

Re­cently, Port­land Less­wrong played a game that was a perfect trifecta of: difficult men­tal ex­er­cise; fun; and an op­por­tu­nity to learn about bi­ases and rec­og­nize them in your­self and oth­ers. We’re still perfect­ing it, and we’d wel­come feed­back, es­pe­cially from peo­ple who try it.

The Short Version

The game is a com­bi­na­tion of Pan­demic, a co­op­er­a­tive board game that is cog­ni­tively de­mand­ing, and the idea of role­play­ing cog­ni­tive bi­ases. Our fa­vorite way of play­ing it (so far), ev­ery­one se­lects a bias at ran­dom, and then at­tempts to ex­ag­ger­ate that bias in their ar­gu­ments and de­ci­sions dur­ing the game. Every­one at­tempts to iden­tify the bi­ases in the other play­ers, and, when a bias is guessed, the guessed player se­lects a new bias and be­gins again.

The Pieces

First, Pan­demic. Pan­demic is a co­op­er­a­tive game with a win con­di­tion and three lose con­di­tions that are sep­a­rate. It pro­vides each player a list of available ac­tions and then al­lows them 4 moves in which to mix and match those ac­tions. Be­cause of the com­bined win and lose con­di­tions, play­ers are con­stantly forced to de­cide be­tween tac­ti­cal and strate­gic ob­jec­tives, with a strong em­pha­sis on mak­ing it easy to choose tac­ti­cally good moves at the ex­pense of failing to win the game and thus los­ing by tak­ing too long. Pan­demic is a fun game, and it’s a great game for peo­ple look­ing to stretch their brains. Ob­vi­ously you want to be fa­mil­iar with, prefer­ably ex­pe­rienced at Pan­demic be­fore at­tempt­ing Bi­ased Pan­demic. We did have one in­ex­pe­rienced player at Pan­demic (out of 4) and that seemed to work okay, though it may have been harder for him than the rest of us.

En­ter the bi­ases. Each player se­lects a bias. We printed out the bi­ases listed here and used them to se­lect our bi­ases. One of our play­ers just made a TOC for se­lect­ing bi­ases. There hap­pen to be 104 bi­ases listed in that doc­u­ment, so a deck of cards com­bined with a coin flip al­lowed for bias se­lec­tion. A com­puter’s ran­dom num­ber gen­er­a­tor, dice, or any other ran­dom method should suffice. Some bi­ases may seem un­playable to some play­ers—cer­tainly, the mon­e­tary bi­ases seem un­playable to most of us—but other play­ers may find a way to play it, so we’ve re­fused to cross off bi­ases and just al­lowed play­ers to re-roll if they get a bias they’re sure they can’t play.


This can be a lit­tle difficult to wrap your brain around, so let me give a cou­ple of ex­am­ples. One player, play­ing the Nega­tivity Bias, went around the board treat­ing cities which had out­bro­ken ear­lier in the game and ig­nor­ing other is­sues. Another player with Hyper­bolic Dis­count­ing went fur­ther: he treated cities, any city near him, while car­ry­ing 5 red city cards in his hand and point­ing out, in re­sponse to en­treaties to cure red, that red wasn’t much of an is­sue right now. A player with Re­ac­tance had the win­ning yel­low card and sim­ply re­fused to be told to go some­where to give it to the player with the other four. He even went so far as to re­fuse a half a dozen offers of an air­lift so he could give up that card. A player with Hind­sight Bias claimed that he had pre­dicted that the player with 1 red card would get two more on his next draw, and was up­set that he’d let the other play­ers ar­gue him oth­er­wise. A player with The Ul­ti­mate At­tri­bu­tion Er­ror sug­gested that if we weren’t do­ing well be­cause no ra­tio­nal­ist could ever win this game be­cause we were ter­rible at it. A player with the Author­ity Bias at­tempted to sug­gest that we should do things be­cause it’s what Eliezer would want us to do. A player with Illu­sion of Con­trol de­clared that his next draw, he sim­ply would not draw an epi­demic. There were many oth­ers.

Recom­mended Rules Of Play

We played it some­what hap­haz­ardly the first game, but at the end we agreed on a struc­ture for the next game that we think is bet­ter. In our next game we plan to have the or­der of play go like this: dur­ing each player’s turn, all play­ers can dis­cuss what the player should do for a timed in­ter­val, per­haps 1-2 min­utes. The player then de­clares their in­tended move. Now each other player gets an op­por­tu­nity to make a sin­gle bias guess. If a guess is cor­rect, the player stops play­ing the bias, and be­gins the round again. At the end of their turn, if their bias was guessed, they se­lect a new bias. We con­sid­ered a bias to be guessed cor­rectly if the player guess­ing fully de­scribed the bias, not just the bi­ased be­hav­ior. Bias names, how­ever, were not re­quired.


One way that you can not do well is by fal­ling into the trap of mak­ing the same bi­ased state­ments re­peat­edly. After a few rounds of this, the bi­ased state­ments were pretty ob­vi­ous. The guesses are an in­di­ca­tor of what the other play­ers are see­ing, and we went out of our way to look for ways to re­spond to the guesses by play­ing up the as­pects of the bias that the other play­ers weren’t see­ing. For ex­am­ple, a lot of the differ­ent bi­ases look like sim­ple over­con­fi­dence. One player was play­ing The Illu­sion of Con­trol in such a way that the rest of us thought he was over­con­fi­dent. His re­sponse was to start declar­ing that he sim­ply wasn’t go­ing to draw an epi­demic card, and when he drew one, he de­clared that it was my fault for mak­ing him draw the card. This was ob­vi­ously not sim­ply over­con­fi­dence.

Be­fore play­ing, you should figure out how fa­mil­iar you are with the bi­ases. Play­ers who are in­cred­ibly fa­mil­iar with all of the bi­ases may want to play a game where ev­ery­one plays as sub­tly as pos­si­ble and your goal is to pre­vent other peo­ple from notic­ing your bias. For us, our goal was to learn the bi­ases bet­ter and iden­tify them in other peo­ple, so we tried to ham it up and play them as ob­vi­ously as pos­si­ble at first. It was in­cred­ibly difficult to speci­fi­cally iden­tify the bi­ased think­ing be­hind ob­vi­ously bi­ased state­ments, even with that, so I’d sug­gest at least try­ing it with ob­vi­ous­ness first.

One of the most difficult things to re­mem­ber is that your goal is not to win the Pan­demic game. Sure, that’s nice, but your real goal is to fa­mil­iarize your­self with bi­ases, and to have fun roie­play­ing and iden­ti­fy­ing bi­ases. Los­ing Pan­demic, es­pe­cially be­cause the play­ers are fol­low­ing their bi­ased think­ing, is a to­tally ac­cept­able out­come. We won, and do not credit our think­ing for it.

We’re look­ing for­ward to try­ing the game again, and maybe you’ll have sug­ges­tions for im­prov­ing it.