Candidate Scoring System, First Release

The first re­lease of CSS is out, with ten­ta­tive rat­ings of most of the con­tenders in the 2020 US pres­i­den­tial elec­tion.

Here is the link.

There are five files here—“Scor­ing Cri­te­ria” gives the overview, method­ol­ogy and preferred policy plat­forms; read it first if you want to un­der­stand ev­ery­thing. “Party Pri­ors” sets prior ex­pec­ta­tions to help fill in val­ues for un­known can­di­dates from ei­ther party. “Can­di­date Eval­u­a­tions” gives the ac­tual scores of each per­son, with re­sults and dis­cus­sion. “Model” is the ac­tual spread­sheet run­ning the whole thing. “Sen­si­tivity Anal­y­sis” is an R pro­gram for check­ing re­sults across differ­ent in­put weights.

Any ideas for im­prov­ing the re­port with new/​differ­ing points of view are wel­come. This is the first edi­tion so it has a long way to go be­fore it can be au­thor­i­ta­tive and con­vinc­ing to most peo­ple. You can down­load a copy of the model spread­sheet to try it with your own as­sump­tions. Un­der the sug­gested de­fault in­puts, these are the re­sults:

Booker 123

De­laney 103

Inslee 97

Gillibrand 83

War­ren 82

San­ders 59

Har­ris 58

Gab­bard 57

Bi­den 52

Cas­tro 51

O’Rourke 43

Brown 40

But­tigieg 32

Klobuchar 29

Haley −23

Trump −126

Thanks to Daniel Eth, Ben Henry, Michael St. Jules and other peo­ple who con­tributed ar­gu­ments and sources. Let me know if you want to be in­volved with cre­at­ing the next re­lease.