Adv. Stats Schedule Rundown - Week 6 Results
Now that I've achieved the magical 100-pt. threshhold for my account, I can actually create a new post instead of merely replying to prior content. Woohoo!
So without further adieu, here's a table I first ginned up after week 3 (before the BYU game) that gives a comparison of the advanced stats discussed above (the root sources coming from Football Outsiders). In addition to S&P+, FEI, GE and F+, I've included FPI, which is ESPN's Monte Carlo analysis. The basic idea for this comparison I got from SaxonRBR on CFBSH, who does something similar for the upcoming SEC games each week.
This table summarizes the actual values and national rankings for M and all its opponents this season (both past & future games), and also applies a pseudo-color scale to the values relative to M's values, to make differentials more visually apparent. The color-coding goes along the lines of a threat-level:
Red => DangerTeam
Grey => sort of rhymes with "meh" ... statistical equivalence
Blue => as cool and inviting as the other side of the pillow
Some interesting things to note from the results:
- The statistical basis for unbridled hype and hysterical smack-talk is now complete. UM is favored in all of its remaining games in both the S&P+ and FPI indices, although clearly the difference between the associated spreads is significant.
- As of this writing, the only game UM is not favored in (including it's prior games), is Utah, and only based on the FPI spread.
- As has already been mentioned in other posts, but bears repeating here: the most difficult remaining game per S&P+ is Penn State, followed closely by Minny.
- Lastly, shameless jumping to conclusions compels me to also mention that UM winning out would also include an additional victory over most-expected West division champion Iowa to earn its first outright Big Ten Championship since 2003. Yup. It's been 12 long years...
So, can anyone recommend any good hotels in Indianapolis? I think I might go ahead and at least reserve a room! BTW, my SO & I prefer non-chain, more boutique-type accommodations.
October 14th, 2015 at 2:16 PM ^
Holy cow.
October 14th, 2015 at 2:25 PM ^
Love this breakout of the stats. Not sure who could tackle this, but it would be interesting to see how statistical comparisons like this actually translate to W/L results, and specifically, is there a trend that can be found related to the likelihood of upsets (against predicted spreads and in general).
As a starting point, just simply averaging the S&P/FEI spreads would seem to tell us that we would beat Rutgers by somewhere around 32 points, which seems to pass the sniff test as compared to our other wins against comparable teams. HOWEVA, I wonder what this same table would have told us MSU would have been expected to beat Rutgers by before last weekend, and how that expected outcome compared to the actual game outcome.
Put simply: Could the numbers themselves help me to avoid putting all my faith in the numbers?
October 14th, 2015 at 2:30 PM ^
Additionally, said statistical wizard could analyze which of these many different statistical models has been the most accurate year in and year out. Until that happens, we can just assume it's whichever model has us highest.
October 15th, 2015 at 11:23 AM ^
If someone could look back on the accuracy of the models used and chrun out a percentage of games that each predicted correctly over the past few years that would be wonderful. Anyone?
October 14th, 2015 at 9:34 PM ^
already do track how their spreads match up to actual games to a great extent on the FootballOutsiders and FootballStudyHall sites.
The F+ is a composite of the FEI and S&P+ models, so if you're looking for some sort of average, there you go. But of course there's no associated spread or probability/odds.
As for the MSU-Rutgers spreads for last weeks game, since I still have the spreadsheet with the old numbers I can tell you:
MSU -21.0 (S&P+)
MSU -14.0 (FPI)
October 14th, 2015 at 2:27 PM ^
I think I'm still too hardened from the advanced stats' adoration of the Denard-era RR offenses that never seemed to go anywhere against decent defenses. This is all still too rosy to possibly believe.
October 14th, 2015 at 2:39 PM ^
and it passes the eye test, and we don't play any good offenses anyway -- unless Ohio State fixes things, which they probably will.
October 14th, 2015 at 6:25 PM ^
Sent from MGoBlog HD for iPhone & iPad
October 14th, 2015 at 9:02 PM ^
I'll look into it!
October 14th, 2015 at 6:27 PM ^
October 15th, 2015 at 10:18 AM ^
Sent from MGoBlog HD for iPhone & iPad
October 15th, 2015 at 10:26 AM ^
Is it weird that I am scared of these stats because they make us look so good, even though, on the eye test, we look really good too.
October 15th, 2015 at 12:47 PM ^
Congratulations on your ascendance to the 100+-point-club, or the "Centurions" as we like to call ourselves. The secret key code ring is in the mail.
Perhaps you have looked at essay writer? I must look at the web site yet I need to read a couple of critiques initial. I really believe you understand a lot concerning creating companies and also would certainly very benefit the view with this make a difference.
Comments