DWibby's Week 11 NFL predictions based on intuitive confidence scale (rated from 0-5, and mapped to 50-100% at 10% intervals) will be at least 5% less than true accuracy at each 10% interval.

Created by dwibby on 2012-11-16; known on 2012-11-20; judged wrong by dwibby on 2012-11-20.

• dwibby estimated 80% on 2012-11-16
• dwibby changed their prediction from “My Week 11 NFL predictions based on my intuitive confidence scale (rated from 0-5, and mapped to 50-100% at 10% intervals) will be at least 5% less than my true accuracy at each 10% interval.” on 2012-11-16
• dwibby changed their prediction from “DWibby's Week 11 NFL predictions based on my intuitive confidence scale (rated from 0-5, and mapped to 50-100% at 10% intervals) will be at least 5% less than my true accuracy at each 10% interval.” on 2012-11-16
• dwibby said “Just getting into calibrating my intuitive confidence, and I expect that my initial scale is a bit greedy.” on 2012-11-16
• RandomThinker said “What does this mean? Your calibration will be 5% overconfident? Or underconfident?” on 2012-11-16
• RandomThinker said “I’m guessing you’re overconfident – a 4 maps to 70% at most. A 5 to 85%, given random upsets.” on 2012-11-16
• RandomThinker said “@dwibby if you want more NFL predictions, I made a bunch of them earlier. Search for NFL:” on 2012-11-16
• dwibby said “Yes, at least 5% overconfident (thanks for the better term), and likely more overconfident at the higher scales.” on 2012-11-16
• RandomThinker said “Be interesting to see if you’re more overconfident at higher scales. I think 3’s and 4’s are probably most overconfident. ” on 2012-11-16
• RandomThinker said “I’m probably overconfident on a few of my 75%‘s (my 4’s).” on 2012-11-16
• dwibby said “@RandomThinker: Thanks, I’ll take a look. I’m just getting into Bayes’ Theorem, and tracking my predictions with a spreadsheet was just not cutting it.” on 2012-11-16
• RandomThinker said “I should make a separate NFL name so I can track my sports predictions :)” on 2012-11-16
• RandomThinker said “Hmm not so easy to set up multiple account names. Need to make up a new email address” on 2012-11-16
• RandomThinker said “So far your calibration is better than mine” on 2012-11-19
• RandomThinker said “One week is too short to measure calibration. Average pundit is only 63% correct; best at 71%. http://predictionbook.com/predictions/9993. ” on 2012-11-20
• RandomThinker said “Vegas favorites are about 66% to win outright. http://www.advancednflstats.com/2011/10/prediction-accuracy.html” on 2012-11-20
• RandomThinker said “Wrong link. http://blog.pundittracker.com/nfl-pundits-breakeven-so-far-in-2012-wickersham-and-mortensen-shine/” on 2012-11-20
• dwibby   judged this prediction wrong on 2012-11-20.
• dwibby said “This was a poor question, as I didn’t have the sample size for a good result. The results are 50% confidence at 0% accuracy; 60% at 0%; 70% at 0%; 80% at 88%;90% at 100%;100% at 100%. I was overly confident on too little data.” on 2012-11-20