Logical decision theories

“Log­i­cal de­ci­sion the­o­ries” are al­gorithms for mak­ing choices which em­body some var­i­ant of “De­cide as though you de­ter­mine the log­i­cal out­put of your de­ci­sion al­gorithm.”

Be­sides the in­tros in the tabs above, a for­mal pa­per on the “func­tional de­ci­sion the­ory” var­i­ant of LDT is now available. This is the best cur­rent in­tro­duc­tion for those who pre­fer to read pa­pers.


  • Guide to Logical Decision Theory

    The en­try point for learn­ing about log­i­cal de­ci­sion the­ory.

  • Introduction to Logical Decision Theory for Economists

    An in­tro­duc­tion to ‘log­i­cal de­ci­sion the­ory’ and its im­pli­ca­tions for the Ul­ti­ma­tum Game, vot­ing in elec­tions, bar­gain­ing prob­lems, and more.

  • Omega (alien philosopher-troll)

    The en­tity that sets up all those trol­ley prob­lems. An alien philoso­pher/​troll im­bued with un­limited pow­ers, ex­cel­lent pre­dic­tive abil­ity, and very odd mo­tives.

  • Introduction to Logical Decision Theory for Computer Scientists

    ‘Log­i­cal de­ci­sion the­ory’ from a math/​pro­gram­ming stand­point, in­clud­ing how two agents with mu­tual knowl­edge of each other’s code can co­op­er­ate on the Pri­soner’s Dilemma.

  • Introduction to Logical Decision Theory for Analytic Philosophers

    Why “choose as if con­trol­ling the log­i­cal out­put of your de­ci­sion al­gorithm” is the most ap­peal­ing can­di­date for the prin­ci­ple of ra­tio­nal choice.

  • An Introduction to Logical Decision Theory for Everyone Else

    So like what the heck is ‘log­i­cal de­ci­sion the­ory’ in terms a nor­mal per­son can un­der­stand?

  • Newcomblike decision problems

    De­ci­sion prob­lems in which your choice cor­re­lates with some­thing other than its phys­i­cal con­se­quences (say, be­cause some­body has pre­dicted you very well) can do weird things to some de­ci­sion the­o­ries.

  • Updateless decision theories

    De­ci­sion the­o­ries that max­i­mize their poli­cies (map­pings from sense in­puts to ac­tions), rather than us­ing their sense in­puts to up­date their be­liefs and then se­lect­ing ac­tions.

  • Fair problem class

    A prob­lem is ‘fair’ (ac­cord­ing to log­i­cal de­ci­sion the­ory) when only the re­sults mat­ter and not how we get there.