←
Chris Monoki4,829 posts
Chris Monoki avatar

Chris Monoki

@Monoki

OC | Sprint Coach | @BradenRiverFTBL | #AirRaidCert 3 | @USATF Level I | Altis Short Sprints Cert

◷ Joined February 2009▧ 176 Photos and videos
Chris Monoki avatar
Chris Monoki@Monoki·Nov 9

End/ Confused? You bet. It's a very complex model that is difficult to explain, leaving many bewildered. Greater transparency and explanation is needed in order to build confidence, in my view.

Chris Monoki avatar
Chris Monoki@Monoki·Nov 9

8/ Quality-win bonus: beat an opponent rated above ~74 and earn extra credit. It tapers in date order: the 1st qualifying win gets full credit, the 2nd ~62%, the 3rd ~45%. Close-loss credit: lose by 7 or less to a team rated above ~61 and get back 7.4 per 10 points they exceed that threshold. It cushions a competitive loss to a strong team. @RickFarmer53 @CoachShack_ @darlington_rick

Chris Monoki avatar
Chris Monoki@Monoki·Nov 9

7/ s = +1 win, −1 loss. K = 8.4, so a win vs a loss is about a 17-point swing in that game's score before margin. m = point margin, capped at ±55. Each point is worth 0.66 in a win and 0.87 in a loss, so getting blown out costs more than running it up gains. @CoachBraselton

Chris Monoki avatar
Chris Monoki@Monoki·Nov 9

6/ R_j = opponent's rating. A major input: each game's score starts at your opponent's number. Against a 90-rated team, it starts at 90; win/loss and margin then adjust it. @BrodyWiseman @Bat_t3

Chris Monoki avatar
Chris Monoki@Monoki·Nov 9

5/ What contributes to ratings in MY model? Point Margin 39% Strength of Schedule 35% Preseason prior 12% Win/loss record (%win) 12% Close-loss credit 1% Quality-win bonus 0.5% Model explains 98.4% of all MaxPreps ratings, with a variance of +/-3.1 pts in rating. So, pretty accurate.

Chris Monoki avatar
Chris Monoki@Monoki·Nov 9

4/ Fit vs the 11/6 MaxPreps release: R² 0.984 (0.983 on held-out teams). Typical miss is about 3 points (CV RMSE 3.13 vs 3.20 for v3). It's a close approximation, not MaxPreps' actual code.

Chris Monoki avatar
Chris Monoki@Monoki·Nov 9

3/ All 557 FL teams are solved at once. Beat a team and their rating feeds yours; if they keep winning, you can rise too. That's why schedule matters every week, not just the night you play. @Cbradley2218 @WhyNot8Teen

Chris Monoki avatar
Chris Monoki@Monoki·Nov 9

1/ I built a model to approximate the @MaxPreps Florida football power ratings from public data, assuming Ned Freeman's CalPreps system is the foundation. Here's how it works. @PayFLCoaches 🧵

Post attachment
Chris Monoki avatar
Chris Monoki@Monoki·Nov 4

Check out Maddox. Impact player. What doesn't show on film is his vocal leadership and field presence. @JimCordle63 @kwamsutton_9 @Coach_DeWall @CoachBaney @qbdietz @KenyonFootball @UnionCollegeFB @DigInMules @SEUFireFootball @GburgFB @BWfootball @FB_KnoxColege

🚨Maddox Hale🚨 📍OLB/EDGE 📏 6’2/ 205 Ibs 🎓 3.5849 I had a very productive season playing a new position. I was able to learn a lot of things on how to be a leader and a better player on and off the field. With a 8-1 record we have secured a spot in the playoffs and there is a lot more to come. I couldn’t have asked for a better group of coaches and teammates. @Monoki @CoachShack_ @RickFarmer53 @JSmith263626 @coachcollier1 @coachporter1010 @Brade_Ice @Coachpatwins @CoachArcadi @941Blitz

3:39
Chris Monoki avatar
Chris Monoki@Monoki·Nov 1

Here's a @MaxPreps proxy formula that explains 97% of ratings. R ≈ O + 0.86(M), where R is predicted MaxPreps rating O is the avg opponent rating M is the avg margin per game, each capped at 60. There's no subjectivity to MaxPreps ratings. @CoachArcadi @CoachPost70

Chris Monoki avatar
Chris Monoki@Monoki·Nov 1

Proud of our boys and all they've accomplished. Still a lot of work to do. Speed. Intensity. Hit like a truck.

Post attachment