Consider the following senate races:
Alabama
Alaska
Arizona (special)
Colorado
Georgia
Georgia (special)
Iowa
Kansas
Kentucky
Maine
Michigan
Montana
North Carolina
The goal is for each race to assign a value, P(D), between 0 and 1 that represents the probability that a Democrat wins the seat, the value P(R ) = 1 - P(D), that represents that a Republican wins the seat is implied by P(D).
Note: there are two senate races in Georgia this year. Be sure to mark which one is the special election.
Entries will be submitted by posting in this thread. The most recent post by a user with a timestamp on or before August 31, 11:59:59 PM CT will be that user’s entry (i.e. you can change your entry until the end of August).
After the election results are known, I will compute the Brier score for each entry. Participants will be ranked by Brier score where lowest is best.
Here’s my August prediction (PredictIt derived) which I will likely change (slightly) by the end of the month.
Race |
P(D) |
Alabama |
3.0% |
Alaska |
5.6% |
Arizona (special) |
93.7% |
Colorado |
96.6% |
Georgia |
19.1% |
Georgia (special) |
19.1% |
Iowa |
35.8% |
Kansas |
5.6% |
Kentucky |
2.2% |
Maine |
80.1% |
Michigan |
91.7% |
Montana |
34.0% |
North Carolina |
76.9% |
May thread
June thread
July thread
Here’s mine:
AK 19%
AL 6%
AZ* 94%
CO 90%
GA 51%
GA* 20%
IA 56%
KS 18%
KY 10%
ME 85%
MI 91%
MT 66%
NC 73%
Lance, Note AK and AL are reversed on our columns; and, I think you mistakenly repeated your GA-special number (19.1) for the Ossoff race.
Thanks for looking out.
Firstly and annoyingly, when you order these states alphabetically by name, and when you order them by two letter postal abbreviation, you do not get the same order. This came up in one of the previous threads, and at once led me to submit my Alaska number for Alabama, which I corrected before the deadline.
Secondly and weirdly, my algorithm was producing the same number to three decimal places for both Georgia races. We’ll see if that still occurs when I update at the end of the month.
(Hey, discourse, here’s your freakin’ compete sentence)
On second glance, Excel vlookup was interpreting the * in GA* as a wildcard and was returning the value for GA.
I think I’ll just have my Python script export a markdown table and skip the Excel step entirely.
Ha! I’m a teacher, and something similar happened to me once — some students ended up getting the wrong grade because Excel would autofill a value as one previous in the list. I immediately created digital folders for the following semester’s classes called “Final grades — AUTOFILL OFF.”
There’s still 12 hours to get an entry in.
Here’s my final update to mine…
Race |
P(D) |
Alabama |
2.7% |
Alaska |
7.5% |
Arizona (special) |
92.4% |
Colorado |
95.9% |
Georgia |
18.6% |
Georgia (special) |
11.9% |
Iowa |
35.4% |
Kansas |
5.4% |
Kentucky |
2.8% |
Maine |
75.6% |
Michigan |
81.7% |
Montana |
27.9% |
North Carolina |
73.8% |