Why Preseason Polls are so fantastic and accurate
Nothing and no one is perfect. However, I just couldn't resist taking a quick look at the AP Poll after one month of play. These were a few of the team results that caught my attention.
ND - preseason rank # 5, currently unranked - 0 (as in zero) "others receiving votes".
MSU - preseason # 15, currently unranked - 0 "others receiving votes"
Miami - preseason # 16, currently unranked - 0 "others receiving votes"
Wisconsin - preseason # 18, currently unranked - 0 "others receiving votes"
As a point of reference - Tulane has "more AP votes" than those four teams.
Any other teams that are catching your attention based on their on field performance? If so, who?
September 27th, 2022 at 1:08 PM ^
WF, Houston, and arguably TAMU & Oregon were all over-ranked preseason as well.
Conversely, Tenn has shot up a lot.
September 27th, 2022 at 1:58 PM ^
Oregon can easily sneak back into the top 10 and TAMU is proving the App state loss was a fluke but ya preseason rankings are always wonky
September 27th, 2022 at 2:28 PM ^
TAMU's OOC, other than the App State loss, was Sam Houston and a 17-9 win over the tire fire that is YTM. Not sure how the body of work thus far suggests App State was a fluke.
Notwithstanding Oregon getting trucked by UGA, Wazoo pulled a same 'ol Lions in their game against the Ducks. I will give the Ducks props however for their win over BYU...
... but if the whole premise of this post is about preseason versus reality, again I say, "body of work".
September 27th, 2022 at 6:31 PM ^
Yes, and A&M got very lucky to beat Arkansas at home. They were outgained by almost 100 yards and survived a FG that hit the upright that would have won it for Arkansas.
A&M's offense is pretty brutal. They are not a top 10 team.
September 27th, 2022 at 2:31 PM ^
Did you actually watch any of that A&M - App St game? Clearly not, because it was not any part of fluke. App St outplayed them for most of that game - they earned the win every bit as much as Marshall earned it's 'W' against ND.
September 27th, 2022 at 4:12 PM ^
A&M had a huge win against arkansas this past weekend so yes i can say with confidence that the app state loss was a fluke on TAMUs part not taking anything away from app state just saying tamu played a bad game. TAMU has some big games coming up so they can make up for the poor start or fail miserably and be a laughing stock
September 27th, 2022 at 4:12 PM ^
A&M had a huge win against arkansas this past weekend so yes i can say with confidence that the app state loss was a fluke on TAMUs part not taking anything away from app state just saying tamu played a bad game. TAMU has some big games coming up so they can make up for the poor start or fail miserably and be a laughing stock
September 27th, 2022 at 9:18 PM ^
SEC always seem to shoot up quicker, and move down slower. That’s at least my observation. I’m still of the opinion that only Georgia and Alabama have earned that status, and sometimes LSU
September 27th, 2022 at 1:09 PM ^
Preseason AP poll is actually a decent prognosticator for the end of the season. But I am still here for any sparty or irish Schadenfreude.
September 27th, 2022 at 1:36 PM ^
Bill Connelly pointed out that his very good system for ranking teams (S&P+) actually got worse when he took out preseason projections. People will always point to 5-10 teams that were overranked but as a whole, preseason polls actually tend to be pretty decent predictors of a team's performance. You just have to realize that no system will do better than like 50-60% accuracy
September 27th, 2022 at 1:53 PM ^
50% accuracy is like playing poker in a casino.
The rankings are so Disney, Fox, & CBS can promote games. That's it.
September 27th, 2022 at 1:56 PM ^
This is a good thing to mention, although as I'm sure you know Bill's preseason projections and the preseason AP poll are quite different.
And since he is using a sort of power-rating system that is meant to be predictive, he is under no obligation to consider head-to-head results, nor the routine of downgrading a team based on won-loss record. Whereas pollsters are compelled to slot teams based on "résumé" or head-to-head results (that may be fluky) and are not intended to be predictive. (Cue TV pundits feigning(?) shock when a lower ranked team is favored vs a higher ranked team...)
September 27th, 2022 at 5:45 PM ^
The 2021 preseason poll was 0% accurate.
I'd also argue that the majority of accuracy granted by pre-season polls is because it has a direct effect on mid season polls, which has a direct effect on playoff rankings, which has a direct effect on post-season rankings. And, it pretty clearly does.
September 27th, 2022 at 6:47 PM ^
Kenpom has shown the same thing in basketball and showed that preseason AP polls outperform pre-tourney polls in predicting tourney results.
Conclusion being that ranking systems put too much weight on a small number of game results and end up discounting things like talent and coaching.
September 27th, 2022 at 4:45 PM ^
"Preseason AP poll is actually a decent prognosticator for the end of the season."
It's an even better prognosticator for the end of the previous season.
Would that make it a post-nosticator?
September 27th, 2022 at 1:11 PM ^
Despite being 3-1 with a ranked loss, Pitt has dropped from 17 to 24.
They were preseason 17, and stayed there after beating WVU week 1. Lost to Tennessee and dropped to 23. Beat Western Michigan and dropped to 24, then beat Rhode Island stayed there.
September 27th, 2022 at 1:20 PM ^
They'll definitely be on the rise if they perform well during ACC play.
September 27th, 2022 at 1:59 PM ^
Pitt should win the coastal without breaking a sweat. That division is trash
September 27th, 2022 at 2:36 PM ^
That has been the case more often than not for a long time.
September 27th, 2022 at 1:13 PM ^
I think Kansas and Kansas State are obvious choices. Kansas already exceeding expectations and have a shot at being a top 4 team in the Big 12.
Kansas State has had a weird season but the win at Oklahoma is probably the best win in CFB of the season so far.
September 27th, 2022 at 1:51 PM ^
Teams that beat the Nebraska this year are 0-3 the following week. Getting outhit must take its toll...
September 27th, 2022 at 1:23 PM ^
Gotta feed the hype machine for the season. Who cares if it’s meaningless? It gets people watching and that’s why they don’t wait for week 4/6/8, like they should.
September 27th, 2022 at 4:46 PM ^
On the other hand, savvy football observers can just ignore the polls until week 4/6/8.
September 27th, 2022 at 1:41 PM ^
It's as simple as ensuring TV ratings early in the season. People get way more excited about #2 Ohio St. vs #5 Notre Dame in week one than if there were no preseason rankings
September 27th, 2022 at 1:45 PM ^
While I share your cynicism in general, I don't think the AP voters are all in a big secret cabal with the networks
September 27th, 2022 at 1:56 PM ^
Yeah................The TV networks don't influence college football.............
September 27th, 2022 at 6:10 PM ^
Your response is a perfect example of a Straw Man logical fallacy
I didn't say networks don't influence college football.
What I was trying to imply is that networks do not get on the phone or otherwise collude with AP poll voters, which are typically sports writers and journalists, to get them to rank teams a certain way to create top-5 matchups.
Do ESPN and the like promote teams whose games they intend to cover? Yes
Are AP poll voters influenced by preseason coverage from ESPN and the like to such an extent that they rank teams _significantly_ higher or lower than they would otherwise rank them? That's where I lose the thread
September 27th, 2022 at 2:05 PM ^
It's too simple for anyone to be "in on" anything. AP does it because it generates clicks and even stories/narratives about itself. The networks want a number next to teams to promote match ups. That's why it will continue. It generates discussion and interest and "accuracy" has next to nothing to do with it and there are no consequences for voters or networks. Isn't a conspiracy because there's no secret.
September 27th, 2022 at 6:19 PM ^
What about AP voters that don't cover OSU/ND week 1? What's their incentive to create a top 5 matchup if they actually believe ND is not a top 5 team?
September 27th, 2022 at 5:48 PM ^
You don't think the people who make money by showing and discussing games are working with the people who are showing and discussing games to make more money?
That seems, silly?
September 27th, 2022 at 6:15 PM ^
Define what you mean by "working with"
I agree the incentives of AP voters that cover the sport at a _national level_ are aligned to some extent with network execs.
I highly doubt those people are working together in any sort of organized fashion. It's a little too illuminati to me
September 27th, 2022 at 2:20 PM ^
TBH I'm not sure whom I would have voted #5 preseason. ND is usually overranked but they were as good a choice (at the time) as any. Everyone outside Georgia has issues and everyone outside the top 4 (hopefully not top 3) has some major issues.
USC has had good results but they are like +25 in turnovers which isn't sustainable -- on a down-to-down basis they are characteristically (for Riley) mediocre on defense. Clemson has issues on both sides of the ball, it seems. They'll be very lucky to go 4-1 in their next five games unless they improve. Kentucky... come on. Tennessee has promise but could lose some shootouts, and they have three games left where they will likely be underdogs.
September 27th, 2022 at 1:49 PM ^
Poll-logic is not just about brands but simple inertia. 1995 me can't comprehend 4-0 Florida St at #23 and 4-0 Kentucky at #7. Nor can 2013 me, actually. Frankly, 2022 me is a bit puzzled.
That LSU win for Florida St may end up looking really good. I partake in Brian Kelly schadenfreude as much as anyone, but he may have provided the coherence that LSU's embarrassment-of-riches talent pool needed. Meanwhile Kentucky looks like they beat another mediocre Florida team.
September 27th, 2022 at 2:01 PM ^
There was a lot of Gator smack talk down here before the season started. AR15 not playing at a high level as expected.
September 27th, 2022 at 1:57 PM ^
Kansas being 4-0 is kinda mindblowing
September 27th, 2022 at 2:06 PM ^
Agreed. And it feels kinda legit. They're a pretty good team.
September 27th, 2022 at 2:42 PM ^
Seems like every 15 years or so, they out of nowhere have a really good team
September 27th, 2022 at 2:11 PM ^
Surprised he only got the Kansas job after doing amazing work at Buffalo. (Kinda wanted him in 2020 *if* the searchbits thing were to happen.). He'll get named for more traditional powers now, if interested.
September 27th, 2022 at 2:03 PM ^
The problem with preseason polls is they take the previous season into account and assume the teams with talent will just reload. This is only the case for ohio state, alabama, georgia (recently), hopefully us from now on. It isnt easy to replace talent especially when you have one player doing all the work making the players around them look so much better (see sparty)
September 27th, 2022 at 2:41 PM ^
Texas? Texas.
September 27th, 2022 at 3:04 PM ^
Syracuse is 4-0 and broke into the Coaches Poll this week! Not bad for a team picked to be in the basement of the ACC Atlantic division.
September 27th, 2022 at 6:43 PM ^
I'm pretty surprised Babers wasn't fired before this season, but his contract runs through 2023 so maybe Syracuse doesn't want to pay a buyout.
September 27th, 2022 at 3:24 PM ^
Honestly posts like this just encourage people to want to put out their pre-season rankings more because they are validated that people are looking at them. It seems that just by you creating a post like this complaining about them is going to do the opposite of what you want which I think is accurate polls. If people ignored the preseason polls then fewer "experts" would publish them.
September 27th, 2022 at 4:48 PM ^
The preseason rankings don'l look that bad relative to 247's team talent composite (which theoretically is a good measure of a team's talent)
Wisconsin 247 talent ranking #25
Miami- #13 Must overcome Gattis' play calling
MSU- #34 - reverse Disrespekt? Shouldn't this depower MSU? : )
ND- #10 Maybe Kelly is a good coach?
U of M #14
September 27th, 2022 at 5:09 PM ^
To defend Gattis the bigger issue for Miami is they have a dogshit defense that still doesn't know how to tackle players to save their lives. I watched their game against MTSU and it was atrocious how they flailed at tackling, similar to last year when they played MSU. Gattis's offense is as predictable as you'd expect but he still put up 30+ points; that should be enough to beat a middling team like MTSU.
Kelly is a good coach and an awful person, and I do think we're seeing the issue with bumping up a young coordinator to a major role so quickly. ND's talent is also somewhat focused on spots like offensive and defensive lines + TEs, which absolutely is important for a team to succeed but also may mean you don't have the playmakers at the skill positions to really move the ball downfield consistently.
September 29th, 2022 at 11:19 AM ^
REM - great chart - however, I think there is definitely an inaccurate data point. That data point is - the Bucks - in 2019. They definitely were not the 14th ranked team in talent.
My view of 247 sports Team Talent Composite has the Bucks ranking as follows - beginning with the current season (2022) and working "backward / earlier" - 3, 3, 3, 2, 1, 2, 5, 3. Which is an "Average" Team Talent of 2.75 over the past 8 years.
September 27th, 2022 at 6:02 PM ^
Polls should come out until AFTER week 3 or 4. I say it every year.
September 27th, 2022 at 6:43 PM ^
Kansas could win the big 12 this year