And Big Red.....Quote:
Originally Posted by LITTLEGUYSINGREEN
Printable View
And Big Red.....Quote:
Originally Posted by LITTLEGUYSINGREEN
Am I reading correctly that Sagarin only predicted the actual winner in 47 out of 106 games? My homemade system was at about 74% last week, and I expect Sagarin to do better than that.Quote:
Originally Posted by thebootfitter
That doesn't seem right. I'd believe that in one week AGS. But not straight up. I picked 10 college games a week over 10 weeks for three years in a contest against maybe 100 others using Sagarin for every pick and finished on the podium at seasons end twice. Straight up should be competitive with your number in my experienceQuote:
Originally Posted by RadioBison
Quote:
Originally Posted by RadioBison
I read it as predicted the winner vs. the spread. Not sure if that's factored into your discussion.Quote:
Originally Posted by bisonaudit
Not living in Vegas, I don't do much sports betting, however, Audit, you have any idea how one would do just using Sagarin to pick 10 college games each week?Quote:
Originally Posted by bisonaudit
Well, just for shits and giggles in the Week 8 thread, after NDSUStudent posted the lines for the Top 25 FCS games, I used Sagarin to decide which ones to bet and how much to wager. Assuming -120, I returned 20% that week.Quote:
Originally Posted by THEsocalledfan
http://www.bisonville.com/forum/show...highlight=week
Extremely small sample and I have no idea if it would hold up over time, but it does illustrate some important concepts:
1) If I'd have sized the bets equally the return would have decreased from 20% to 13%. Bet sizing is important.
2) There were 20 games on the board that week. Sagarin only liked 13 of them. Of the 13 plays, Sagarin won 8 and lost 5. Sagarin was 4 and 3 on the games that it wasn't confident enough in to wager. Playing all the games would have further reduced the return from 13% to 10%. Game selection is important.
So, basically, were you looking at games where the spread was out of proportion to the Sagarin predictor? Then, perhaps bet more on games that have a bigger discrepancy?Quote:
Originally Posted by bisonaudit
I was surprised too. I will go back to double check to be sure. But yes, per my analysis, Sagarin only predicted the winner straight up 44% of the time. By "actual spread," I meant the actual difference of points on the board at the end of the game. Perhaps I misused the term.Quote:
Originally Posted by RadioBison
Correct, the more distance between Sagarin's prediction of the result and the line the bigger the wager.Quote:
Originally Posted by THEsocalledfan
Examples from Week 8:
NDSU -9 @ S Ill: NDSU - S Ill - Home Field = 75.41 - 64.07 - 3.66 = 7.68 NDSU favored per Sagarin. The spread is -9 so Sagarin preferred a wager on S Ill, but only by 1.32 points. That's not enough at -120. No wager.
McNeese St -1.5 v. SHSU: This line surprised a lot of people but Sagarin liked McNeese even more than Vegas. On a neutral field (65.31 - 59.95) by 5.36. But this was at home and McNeese didn't play the week before so Sagarin actually made McNeese an 11.63 point favorite, more than a 10 point edge. The juiciest line of the week. Play 5 units on McNeese. Winner.
Tenn-Martin +3.5 v. Tenn St: Sagarin agreed that Tenn-Martin was the underdog but not by 3.5 points, just 0.82. That 2.68 point edge was just enough to overcome the vig. Play 1 unit on Tenn-Martin. Loser.
I cherry picked the examples. You can see on the other thread that I was a loser on the 3 and 4 unit plays.
Not trying to be a smart ass, but Vegas never had a line on this game and they don't put lines on fcs vs fcs regular season games.Quote:
Originally Posted by bisonaudit
Maybe NDSUStudent can tell us where he got them from.Quote:
Originally Posted by SamsRams
Just in case he was blowing smoke I'll add this:
I applied a similar approach to the NFL for parts of 2010 (wk 9 - super bowl) and 2011 (wk 7 - wk 14) using Bodog's lines and saw comparable results. 15.7% average return. 7 losing weeks in 21. 110-86-4 against the spread.
Think LVH will run lines on the FCS playoffs?Quote:
Originally Posted by SamsRams
Quote:
Originally Posted by RadioBison
Good catch, guys. I should have caught it myself when I questioned the low number the first time, but I didn't double check that formula until now. The 47/106 figure is when the Sagarin predicted spread is at least as large as the actual point differential AND the predicted winner is correct.Quote:
Originally Posted by bisonaudit
Straight up in terms of predicting the actual winner based on the points on the board, Sagarin was right 81 times. That seems much more in line with expectations. I also noticed that I failed to exclude a DIAA-DII game because Incarnate Word played at McMurray (DII). So the total number of games between DI opponents was actually only 105. I guess I need to get someone to peer review my work, eh?
Sweet! NDSU cracked the top 30 again. So K-State ' loss to us gives them another loss to a top 30 program. And with K-State at 22, we still have a win over a top 30 team. We jumped all the way up to 79.27 in Sagarin's predictor score. That's about where we ended the season last year when we were 32nd at 79.53.
Wow. Just wow!
Sent from somewhere using my Windows phone.
Today's numbers are baffling to me. Huge jumps both in ratings and positions all up and down the list. It feels like the predictor formula changed or there are errors in the data. I'll try and remember to come back later today and see if anything changed(wouldn't be the first time).
Big Sky takes a huge tumble; Montana State drops out of the top-25 altogether. Jacksonville State comes out of nowhere to take #15. Two Ivy teams crack the top-25. Illinois State makes it back into the top-25 with a loss to us. That makes 70% of the Valley in the top-25. The OVC passes both the Southland and Big Sky to take third place. CAA teams take 4, 5, 6, 13 & 24. Not a single Southern team makes the top-25. Mass chaos. Bison are still nearly a two TD favorite at Youngstown.
Sagarin's Week 12 Predictions
Sagarin FCS Predictor Top-25Code:Sagarin Predict Result Diff
31 North Dakota State 79.27
23 Kansas State 83.48 -8.32 3.00 11.32
NA Ferris State NA NA 46.00 NA
232 Delaware State 34.14 49.24 51.00 1.76
111 South Dakota State 60.68 14.48 20.00 5.52
91 Northern Iowa 63.33 20.05 1.00 -19.05
95 Missouri State 62.63 20.75 15.00 -5.75
107 Southern Illinois 61.18 13.98 21.00 7.02
199 Indiana State 40.50 34.66 46.00 11.34
121 Illinois State 57.92 25.46 18.00 -7.46
100 Youngstown State 62.07 13.09
163 South Dakota 50.61 32.77
162 Western Illinois 50.81
Total Diff: 4.70
Home 4.11
Sagarin Predictor Top-6 FCS ConferencesCode:1 31 North Dakota State
2 48 Eastern Illinois
3 69 SE Louisiana
4 74 Villanova
5 75 William & Mary
6 78 Towson
7 83 Eastern Washington
8 91 Northern Iowa
9 93 Montana
10 94 Sam Houston State
11 95 Missouri State
12 100 Youngstown State
13 101 Maine
14 103 McNeese State
15 106 Jacksonville State
16 107 Southern Illinois
17 108 Coastal Carolina
18 111 South Dakota State
19 112 Cal Poly
20 113 Harvard
21 114 Old Dominion
22 115 Eastern Kentucky
23 118 Princeton
24 120 New Hampshire
25 121 Illinois State
Code:1. MVFC 59.24
2. Colonial 56.11
3. Ohio Valley 52.84
4. Southland 52.68
5. Big Sky 48.81
6. Southern 48.65
The difference for NDSU decreased by like 14 points. I don't think it's moved that much all year. Maybe the whole model snapped to a new equilibrium?
A couple thoughts:Quote:
Originally Posted by Hammersmith
1) Wow
2) LOL@BigSky
3) The entire MVFC, with the exception of Indiana State, would be above average in the Big Sky.
4) UNI is still getting love from beating Iowa State and playing NDSU tough?
4a) losing a bunch of games close in the toughest conference in the division. It's all about the score with the predictor it doesn't see W/L.Quote:
Originally Posted by Tatanka
4.a.1)Sorry should have been a period not a question mark.Quote:
Originally Posted by bisonaudit
If I'm remembering right, the home field figure was less than 4 points last week. I think something did change.Quote:
Originally Posted by bisonaudit
Look at the list of teams below us! Some big names. At least 11 teams that have won a national title.
Sent from my DROID RAZR using Tapatalk
4b) Kicking the crap out of McNeese State versus the prediction (nearly 30 points)Quote:
Originally Posted by Tatanka
The only terrible game for UNI was South Dakota(-24). Two bad games in SIU(-13) and ISUr(-11). Good games were NDSU(+19) and Iowa State(+11). Fantastic game was McNeese(+29). All the rest were within 5 points of the prediction.
Quote:
Originally Posted by gumby013
3.59 last week; 4.11 this week.
That systems like a crazy swing in home field, too. USD/Mont. is probably a partial explanation for what happened to the Big Sky but it seems like this isn't the first late season reshuffle we've seen in the ratings in recent years either.
I think that home field advantage is tricky, because the results are skewed so much over the first few weeks. I wonder if the model might have just switched to using data from only this year.Quote:
Originally Posted by gumby013
Speaking of home field, Delaware State vs un_ at Delaware State would be pretty much a dead heat.
Had a great time this morning razzing my friend who is a huge Sooner fan.
Sent from my DROID RAZR using Tapatalk
I probably won't continue to post these stats here anymore beyond today, since they don't really have any direct relevance to the Bison, except to help illustrate whether Sagarin is a reliable predictor of outcomes for future games. But if anyone is interested in more, send me a PM. I plan to continue this analysis through the end of the season and see if any reasonable conclusions can be drawn then.
Overall, the average predicted vs actual point differential is quite small this week (only 0.3 points), but the variance is up a bit with a standard deviation of 16. Sagarin correctly predicted the winner 76% of the time.
Out of 108 games between DI opponents:
--> Average difference between Sagarin's predicted point differential and the actual point differential is only 0.3 points (1.3 points last week), with a standard deviation of 16 points (15 last week).
--> (75.9%) 82 games - Sagarin correctly predicted the winner (77% last week)
--> (15.7%) 17 games were within 2 points of Sagarin's predictions in either direction (7.5% last week)
--> (12.0%) 13 games were between 2 & 5 points off Sagarin's predictions (15.1% last week)
--> (21.3%) 23 games were between 5 & 10 points off Sagarin's predictions (24.5% last week)
--> (15.7%) 17 games were between 10 & 15 points off Sagarin's predictions (23.6% last week)
--> (35.2%) 38 games were greater than 15 points off Sagarin's predictions (29.2% last week)
Massey lists us at #32 in all of DI currently. Interestingly, Fresno St. (who is also 9-0) is at #40, with very slightly tougher schedule strength (Fresno's SoS is 106th vs ours of 110th).
Fair enough as I'm sure it is a chore, but if you are stopping because you think nobody cares that is a mistake. I love this stuff and rarely post on it because I have nothing useful to add.Quote:
Originally Posted by thebootfitter
With that said I love this stuff not because I understand Sagarin (or anyone else's) methodology, but just that an impartial and unbiased observer of objective data (compared to polls which are waaaaay too biased by many things like program history, preseason predictions, and much other BS) like Danny Sagarin's formula has us ranked right in there with so many big name programs. That is cool stuff. Because of scholly depth I don't see us lasting long in a tourney like the hoops one but in round 1 it sure looks like we'd go toe to toe with all but the major contenders. How cool is that?Quote:
Originally Posted by mnriverbison
Like mnriver said, if you are doing the analysis anyway, by all means, post it.Quote:
Originally Posted by thebootfitter
I hope you reconsider and continue to post here. Interesting stuff!
Agree please keep posting. Data doesn't lie.Quote:
Originally Posted by Green1
Sounds pretty kewl, does that include a moral victory for holding a team below their average point margin?Quote:
Originally Posted by mnriverbison
Quote:
Originally Posted by thebigund
It's no victory over Northern Colorado...
Based on Sagarin's numbers (4.11 home field) and a 2.61 pt kicker for coming off a bye week, here's the outlook for the remainder of the regular season:
YSU 83.4% (+2.6 percentage points from last week)
USD 99.2% (+0.4)
11-0 82.7% (+5.1)
10-1 17.2% (-4.8)
9-2 0.1% (-0.3)
Assumes a normal distribution of outcomes w/ Sagarin's projected spread as the mean and a standard deviation of 13.86. In round numbers this means that if the spread is 14 the model predicts that roughly 2/3 of the outcomes will be between a tie and a 28 point victory. Half of the remaining outcomes or about 1/6 of all the outcomes would be a victory by more than 28 and there would be roughly a 1 in 6 chance that the team favored by 14 would lose.
Playoffs:
Sagarin gives NDSU a 55.7% chance of winning the SportsNetwork bracket.
Most likely opponents by round (assuming NDSU advances):
Montana 82.9% (San Diego 17.1%)
#8 SE Louisiana 91.7% (Tennessee State 5.8%)
#4 Maine 43.3% (#5 Sam Houston St. 35.0%)
#2 Eastern Illinois 66.2% (#3 E. Washington 14.7%)
CUSA 59.82 (-0.29 from last week)
MVFC 59.24 (+1.13)
MAC 58.72 (-1.10)
Sn Blt 58.08 (-0.13)
And a Gatorade bath to boot. I'm not familiar with protocol, so I may have the order wrong.Quote:
Originally Posted by thebigund
Actually, well played. Immediately when I read that post, I knew your reply would be swift and pretty much spot on.