This is the official thread for the 2022 Houston Match Prediction Contest. Here is the contest from 2019 for reference.
We will hopefully get preliminary schedules before the day of the competition, I will keep this thread updated with submission deadlines. If the schedules are not released until very late (which is very possible), then Thursday will be skipped for this contest and only Friday matches will be scored. I just wanted to get this thread started to let anyone interested start preparing.
Scoring
The master spreadsheet is here (There’s not much in it yet, we will update as schedules are released). Add a column with your name and predictions for each division.
I will use Brier scores to determine the best model. Your prediction will be subtracted from the actual result (red win = 0, blue win = 1, tie = 0.5), and then squared. So if you predict a 85% blue win probability, and blue does win, then your Brier score for that match would be 0.15^2 = 0.0225.
The model with the lowest average Brier score, throughout all matches from all eligible divisions (see the 5th bullet point in “Other rules”), will win the contest.
You are welcome to not enter all divisions, but your predictions for all divisions that you do not enter will be considered to be 0.5 (tie) for every match, for the sake of finding the best overall model.
Other Rules:
Limit one submission per division per person
You are free to make submissions for multiple divisions.
I will post in this thread what the final deadline is for predictions as we get closer to the event, but you should assume it to be 1 hour before the first official match is scheduled to start.
Don’t directly copy anyone else’s predictions. I’m sure everyone is fine with you using their predictions as a starting point, but you should really strive to get at least 100 percentage points different than everyone else. I may or may not enforce this.
I reserve the right to remove any submission for any reason (e.g. if I find out you are throwing matches to improve your predictions).
I reserve the right to change or add any rules at any time, also I’m not going to be a great organizer since I’m on a busy Australia trip, so I’m counting on all of you to self-moderate.
Good luck, have fun, and I look forward to seeing what everyone comes up with.
Quick opinion poll, in the past I have made a single submission deadline at some point before the first match starts. Would there be interest in having two submission deadlines, one before the day 1 matches start and another before day 2 matches start? This would allow incorporation of day 1 performances for day 2 predictions. It would be more work for people to enter and a bit harder to manage, but could provide for a more interesting contest.
Would you like a second submission deadline for day 2 matches?
One submission deadline, same as previous contests.
Two submission deadlines, day 2 predictions may be modified anytime up to the first day 2 match.
Alright, seems pretty clear people want to stick with one submission deadline. We will stick with that, just wanted to get a feel for the idea.
We are still waiting on preliminary schedules. Hopefully we get those soon. Assuming we get those in a timely manner, my plan is to make the submission deadline 1 hour before the first match is scheduled to begin. If any schedules have last minute changes from the preliminary schedules (like Turing did in 2019), those divisions will not be included as part of the contest.
If we do not get preliminary schedules, my backup plan is to make the submission deadline around noon tomorrow. Only matches occurring after that time will be scored.
Okay, still no preliminary schedules. If we don’t get them by 7pm Central time, I will be pushing back the submission deadline to the start of the lunch break tomorrow.
They look great! I was busy tonight so haven’t gotten a chance to format the sheet, but thank you for submitting.
Because the preliminary schedules came out so late, and in hopes of getting a few more entries, I am going to push the final submission deadline back to 1pm central time tomorrow, and only qual matches 30 and above will be scored for the contest, although everyone is free to enter predictions for matches 29 and below if you so choose.
Alright, it is now 1PM Central time, so all predictions have locked. Good luck to our contestants:
Bovlb
Caleb_Sykes
Eugene
cobbie
Strategos
I did update my predictions about half an hour ago to include partial Thursday morning data.
This year was really tough because we got the schedules so late, I wanted to give people reasonable time to submit, but ideally I still wanted the submission deadline to be before the first match. Hoping we get preliminary schedules sooner in future years so it’s not so tricky.
Unfortunately, I didn’t get to finish my model in time for this year. But I spent a little time to make the spreadsheet auto-updating (minus ties, which will be skipped and need to enter manually)
Thank you so much for doing that. I’ve been super busy the last few days so haven’t been able to keep up with everything, so I appreciate everyone who’s chipping in to make the shared doc better.
Super late, but here were the final Brier scores over the 502 scored matches (quals 30 and later):
Entry
Total Brier Score
Caleb_Sykes
0.1431
cobbie
0.1481
Eugene
0.1553
Strategos
0.1587
Bovlb
0.1590
Tie Every Match
0.2490
Coiny McCoinface
0.5010
Congratulations to @cobbie for having the best predictions other than mine! Feel free to share how you arrived at your predictions if you so choose.
My entry came from a simple average of Elo and OPR predictions straight from my event simulator. I did use the results from matches 1-27ish in each division when making my qm 30+ predictions, which seemed to give me an edge compared to entries that only used pre-event data.
Biggest upsets according to the average of all entries:
division
match
ROE
106
GAL
108
ROE
80
GAL
31
ROE
35
TUR
45
ROE
44
NEW
81
CARV
34
CARV
71
Most expected victories according to the average of all entries:
division
match
ROE
100
CARV
38
GAL
37
ROE
59
GAL
102
TUR
41
GAL
89
TUR
98
HOP
35
CARV
98
Unfortunately it was difficult to run a contest this year due to how late the schedules were released. I didn’t really want people to incorporate early matches in predictions, but I also didn’t want anyone to feel forced to stay up late or wake up early just to submit, particularly for attendees of the championship event. Open to feedback for how to structure the contest better given the late schedule releases.