VaMatrix Posted November 2 Report Posted November 2 Week 11 Game Predictions: https://vafbmatrix.com/schedule (of course filterable by class, region, and district) Also "Playoff Mode" has been a big hit! If you're interested in looking at seeding possibilities, that's where you want to be! Click the dropdown arrow beside each team in "projected seeding" to see their % chance to be any particular seed. (Please report any mistakes) Fairlawncat and wdr31 2 Quote
rdh2443 Posted November 2 Report Posted November 2 Awesome and thanks for this once again! GO PACK!! Quote
Fairlawncat Posted November 2 Report Posted November 2 5 hours ago, VaMatrix said: Week 11 Game Predictions: https://vafbmatrix.com/schedule (of course filterable by class, region, and district) Also "Playoff Mode" has been a big hit! If you're interested in looking at seeding possibilities, that's where you want to be! Click the dropdown arrow beside each team in "projected seeding" to see their % chance to be any particular seed. (Please report any mistakes) Strong work brother. Thanks for what you do. Awesome website and solid website. All the best wishes to you and your hard work. rdh2443 and VaMatrix 2 Quote
Wolf1207 Posted November 3 Report Posted November 3 Question. How much do these predictors take into account home field? I know this isn’t Vegas line, but RV predicted by 5.4 seems pretty high over GC on the road. Quote
jeffro Posted November 3 Report Posted November 3 10 minutes ago, Wolf1207 said: Question. How much do these predictors take into account home field? I know this isn’t Vegas line, but RV predicted by 5.4 seems pretty high over GC on the road. i was thinking same thing! Quote
VaMatrix Posted November 3 Author Report Posted November 3 2 hours ago, Wolf1207 said: Question. How much do these predictors take into account home field? I know this isn’t Vegas line, but RV predicted by 5.4 seems pretty high over GC on the road. great question, using the data collected from last year, there is a 1.5 point home field advantage modifier, so on a neutral field it would give Ridgeview almost a 7 point advantage. obviously in reality that could vary more or less from team to team in a way the computer can't recognize though. There is more to it than this, but from common opponents, Ridgeview is averaging more than 10 points better score differential. Obviously Union is the most telling, which is one where GC has the advantage, but the formula does take that into account and put more weight on the closer matchups than blowouts. I'll add a second post right after this showing transitive point differential by two steps. rdh2443 and Wolf1207 2 Quote
rdh2443 Posted November 3 Report Posted November 3 7 minutes ago, VaMatrix said: great question, using the data collected from last year, there is a 1.5 point home field advantage modifier, so on a neutral field it would give Ridgeview almost a 7 point advantage. obviously in reality that could vary more or less from team to team in a way the computer can't recognize though. There is more to it than this, but from common opponents, Ridgeview is averaging more than 10 points better score differential. Obviously Union is the most telling, which is one where GC has the advantage, but the formula does take that into account and put more weight on the closer matchups than blowouts. I'll add a second post right after this showing transitive point differential by two steps. Just amazing work!!! VaMatrix 1 Quote
VaMatrix Posted November 3 Author Report Posted November 3 8 minutes ago, VaMatrix said: great question, using the data collected from last year, there is a 1.5 point home field advantage modifier, so on a neutral field it would give Ridgeview almost a 7 point advantage. obviously in reality that could vary more or less from team to team in a way the computer can't recognize though. There is more to it than this, but from common opponents, Ridgeview is averaging more than 10 points better score differential. Obviously Union is the most telling, which is one where GC has the advantage, but the formula does take that into account and put more weight on the closer matchups than blowouts. I'll add a second post right after this showing transitive point differential by two steps. By two-step connectors (a two game connecting transitive point differential) Ridgeview maintains a +16.3 net average point differential per game against GC. Again, many of these include blowouts which don't tell us as much, but over 22 connections, it is worth noting, and shows us a glimpse of how Ridgeview is being rated so well in the statewide network. DIRECT CONNECTORSTWO-STEP CONNECTORS 22 TWO-STEP CONNECTORS · RIDGEVIEW +16.3 Gate City→+13(41–28)→Abingdon→−26(40–66)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +49 Gate City→+13(41–28)→Abingdon→+56(62–6)→John Battle→−44(6–50)→Ridgeview GATE CITY +25 Gate City→+13(41–28)→Abingdon→−34(42–76)→Union→+3(23–20)→Ridgeview RIDGEVIEW +18 Gate City→+26(33–7)→Central (Wise)→+26(66–40)→Abingdon→−48(14–62)→Ridgeview GATE CITY +4 Gate City→+26(33–7)→Central (Wise)→+28(42–14)→John Battle→−44(6–50)→Ridgeview GATE CITY +10 Gate City→+26(33–7)→Central (Wise)→−1(13–14)→Lee→−43(0–43)→Ridgeview RIDGEVIEW +18 Gate City→+26(33–7)→Central (Wise)→−43(6–49)→Union→+3(23–20)→Ridgeview RIDGEVIEW +14 Gate City→+42(42–0)→John Battle→−56(6–62)→Abingdon→−48(14–62)→Ridgeview RIDGEVIEW +62 Gate City→+42(42–0)→John Battle→−28(14–42)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +22 Gate City→+42(42–0)→John Battle→−14(27–41)→Lee→−43(0–43)→Ridgeview RIDGEVIEW +15 Gate City→+28(42–14)→Lee→+1(14–13)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +7 Gate City→+28(42–14)→Lee→+14(41–27)→John Battle→−44(6–50)→Ridgeview RIDGEVIEW +2 Gate City→+28(42–14)→Lee→−42(0–42)→Union→+3(23–20)→Ridgeview RIDGEVIEW +11 Gate City→+43(49–6)→Marion Senior→−29(25–54)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +22 Gate City→+43(49–6)→Marion Senior→−28(6–34)→Lee→−43(0–43)→Ridgeview RIDGEVIEW +28 Gate City→+49(49–0)→Richlands→−55(0–55)→Abingdon→−48(14–62)→Ridgeview RIDGEVIEW +54 Gate City→+49(49–0)→Richlands→−40(9–49)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +27 Gate City→+49(49–0)→Richlands→−33(0–33)→Lee→−43(0–43)→Ridgeview RIDGEVIEW +27 Gate City→+49(49–0)→Richlands→−40(21–61)→Southern Gap→−40(8–48)→Ridgeview RIDGEVIEW +31 Gate City→+6(21–15)→Union→+34(76–42)→Abingdon→−48(14–62)→Ridgeview RIDGEVIEW +8 Gate City→+6(21–15)→Union→+43(49–6)→Central (Wise)→−36(7–43)→Ridgeview GATE CITY +13 Gate City→+6(21–15)→Union→+42(42–0)→Lee→−43(0–43)→Ridgeview GATE CITY +5 Quote
rdh2443 Posted Tuesday at 07:41 PM Report Posted Tuesday at 07:41 PM 15 hours ago, VaMatrix said: By two-step connectors (a two game connecting transitive point differential) Ridgeview maintains a +16.3 net average point differential per game against GC. Again, many of these include blowouts which don't tell us as much, but over 22 connections, it is worth noting, and shows us a glimpse of how Ridgeview is being rated so well in the statewide network. DIRECT CONNECTORSTWO-STEP CONNECTORS 22 TWO-STEP CONNECTORS · RIDGEVIEW +16.3 Gate City→+13(41–28)→Abingdon→−26(40–66)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +49 Gate City→+13(41–28)→Abingdon→+56(62–6)→John Battle→−44(6–50)→Ridgeview GATE CITY +25 Gate City→+13(41–28)→Abingdon→−34(42–76)→Union→+3(23–20)→Ridgeview RIDGEVIEW +18 Gate City→+26(33–7)→Central (Wise)→+26(66–40)→Abingdon→−48(14–62)→Ridgeview GATE CITY +4 Gate City→+26(33–7)→Central (Wise)→+28(42–14)→John Battle→−44(6–50)→Ridgeview GATE CITY +10 Gate City→+26(33–7)→Central (Wise)→−1(13–14)→Lee→−43(0–43)→Ridgeview RIDGEVIEW +18 Gate City→+26(33–7)→Central (Wise)→−43(6–49)→Union→+3(23–20)→Ridgeview RIDGEVIEW +14 Gate City→+42(42–0)→John Battle→−56(6–62)→Abingdon→−48(14–62)→Ridgeview RIDGEVIEW +62 Gate City→+42(42–0)→John Battle→−28(14–42)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +22 Gate City→+42(42–0)→John Battle→−14(27–41)→Lee→−43(0–43)→Ridgeview RIDGEVIEW +15 Gate City→+28(42–14)→Lee→+1(14–13)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +7 Gate City→+28(42–14)→Lee→+14(41–27)→John Battle→−44(6–50)→Ridgeview RIDGEVIEW +2 Gate City→+28(42–14)→Lee→−42(0–42)→Union→+3(23–20)→Ridgeview RIDGEVIEW +11 Gate City→+43(49–6)→Marion Senior→−29(25–54)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +22 Gate City→+43(49–6)→Marion Senior→−28(6–34)→Lee→−43(0–43)→Ridgeview RIDGEVIEW +28 Gate City→+49(49–0)→Richlands→−55(0–55)→Abingdon→−48(14–62)→Ridgeview RIDGEVIEW +54 Gate City→+49(49–0)→Richlands→−40(9–49)→Central (Wise)→−36(7–43)→Ridgeview RIDGEVIEW +27 Gate City→+49(49–0)→Richlands→−33(0–33)→Lee→−43(0–43)→Ridgeview RIDGEVIEW +27 Gate City→+49(49–0)→Richlands→−40(21–61)→Southern Gap→−40(8–48)→Ridgeview RIDGEVIEW +31 Gate City→+6(21–15)→Union→+34(76–42)→Abingdon→−48(14–62)→Ridgeview RIDGEVIEW +8 Gate City→+6(21–15)→Union→+43(49–6)→Central (Wise)→−36(7–43)→Ridgeview GATE CITY +13 Gate City→+6(21–15)→Union→+42(42–0)→Lee→−43(0–43)→Ridgeview GATE CITY +5 Mind blowing! Great job with explaining this! Quote
gwloyal Posted Wednesday at 05:52 PM Report Posted Wednesday at 05:52 PM As a long time bettor, an ol sharp told me a long time ago “Betting on the transitive property will make you a very broke man, son.” And I’ve seen numerous examples of that advice save me a lot of bankroll over the years. So whenever the words “transitive” and “common opponents” pop up, I cringe just a little bit. However I am fully aware that there’s only so much you can play out with only 9-10 data points when it comes to quantitative forecasting. By no means being critical at all. Not sure if two-step transitives were your brainchild or a borrowed one, but it definitely intrigues me more than just common opponents. It certainly provides a more solid data pools and provides a little bit clearer pattern to them. As a numbers geek, I appreciate the explanation and demonstration. I’ve enjoyed the site @VaMatrix. Awesome stuff! Hope to see you continuing this for years to come! rdh2443 and VaMatrix 2 Quote
VaMatrix Posted Wednesday at 06:34 PM Author Report Posted Wednesday at 06:34 PM 41 minutes ago, gwloyal said: As a long time bettor, an ol sharp told me a long time ago “Betting on the transitive property will make you a very broke man, son.” And I’ve seen numerous examples of that advice save me a lot of bankroll over the years. So whenever the words “transitive” and “common opponents” pop up, I cringe just a little bit. However I am fully aware that there’s only so much you can play out with only 9-10 data points when it comes to quantitative forecasting. By no means being critical at all. Not sure if two-step transitives were your brainchild or a borrowed one, but it definitely intrigues me more than just common opponents. It certainly provides a more solid data pools and provides a little bit clearer pattern to them. As a numbers geek, I appreciate the explanation and demonstration. I’ve enjoyed the site @VaMatrix. Awesome stuff! Hope to see you continuing this for years to come! I couldn’t agree more. The rating/predictive engine is a tweaked version of a Massey Matrix (which I guess technically is the transitive results taken to the extreme!) But to clarify, just these direct transitive connections are not what is driving the predictions, just a fun feature of the site. You are right about limited data points. Ideally a prediction model like I’m purporting would have per possession efficiency ratings for team offenses and defenses etc. I’m trying to squeeze as much juice as I can out of just raw scores! I would expect humans familiar with the teams to outperform these predictions, but on a statewide scale hopefully we are getting in the ballpark for most games! rdh2443 1 Quote
gwloyal Posted Wednesday at 06:53 PM Report Posted Wednesday at 06:53 PM 10 minutes ago, VaMatrix said: I couldn’t agree more. The rating/predictive engine is a tweaked version of a Massey Matrix (which I guess technically is the transitive results taken to the extreme!) But to clarify, just these direct transitive connections are not what is driving the predictions, just a fun feature of the site. You are right about limited data points. Ideally a prediction model like I’m purporting would have per possession efficiency ratings for team offenses and defenses etc. I’m trying to squeeze as much juice as I can out of just raw scores! I would expect humans familiar with the teams to outperform these predictions, but on a statewide scale hopefully we are getting in the ballpark for most games! Didn’t mean to insinuate that it was all based off that. My apologies if I did. And I think you absolutely get as much as you can with what you are working with. One thing is for sure….even if the model doesn’t predict the correct winner, it’s very much a good indicator of who has performed better over the course of the season up to that point. The only thing I don’t care for is now I’ll have to talk myself out of looking at two-step connectors on college football Saturday mornings now 😂 rdh2443 1 Quote
VaMatrix Posted Wednesday at 07:19 PM Author Report Posted Wednesday at 07:19 PM 26 minutes ago, gwloyal said: Didn’t mean to insinuate that it was all based off that. My apologies if I did. And I think you absolutely get as much as you can with what you are working with. One thing is for sure….even if the model doesn’t predict the correct winner, it’s very much a good indicator of who has performed better over the course of the season up to that point. The only thing I don’t care for is now I’ll have to talk myself out of looking at two-step connectors on college football Saturday mornings now 😂 Nah you’re good! I could talk about this stuff all day! Quote
Unchained Posted Wednesday at 09:29 PM Report Posted Wednesday at 09:29 PM 2 hours ago, VaMatrix said: Nah you’re good! I could talk about this stuff all day! You're knocking it out of the park. Great work! gwloyal and VaMatrix 2 Quote
gwloyal Posted Wednesday at 10:29 PM Report Posted Wednesday at 10:29 PM 3 hours ago, VaMatrix said: Nah you’re good! I could talk about this stuff all day! I was referring to applying that to my own "degenerate work" haha! But YES I could chit chat about it a good while myself. VaMatrix 1 Quote
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.