eerily on point
|1 week 1 day ago||Its pretty normal for objects||
Its pretty normal for objects towards the edges of photographs to appear larger than things in the middle.
|2 weeks 1 day ago||From the article: 2014 Dallas||
From the article:
2014 Dallas Cowboys at Jacksonville Jaguars
Is 91% really high accuracy? Sounds like a recipe to get gashed by play action to me.
|1 year 13 weeks ago||Davis really did make an||
Davis really did make an argument for more playing time. Hopefully it wasn't a result of the competition level as it was nice to see.
The game also clarified Altidore's role I think. As Brian said, he didn't have many one-on-one scoring opportunities, but he did a good job of boxing defenders out and fighting for possession.
There was a nice sequence in the second half where Altidore controlled a long pass, kept the defender away from the ball with his body, made a quick little pass to Bradley who tried to thread another quick pass to a wide open Davis which was just narrowly deflected.
I have a feeling that more than one goal in this World Cup will follow that model: Altidore fights for possession deep allowing the rest of the team to come up, quick pass to Dempsey, quick pass to Bradley, etc, as opposed to a lone player streaking down the field beating defenders.
|1 year 41 weeks ago||Gardner seems to be reading||
Gardner seems to be reading the MLB ... Does he keep because that LB takes off for the playside? But that doesn't seem right, because the pulling guard ends up on the MLB. Is this supposed to be a handoff but Gardener pulls because the FB botches the block?
|1 year 45 weeks ago||I very much like this theory,||
I very much like this theory, and giving the coaches some credit in general. Regarding important games, I would throw in ND, which would seem to be a datapoint in Borges's favor.
Now, for the tackle-over stuff, I think it makes a lot of sense to deploy it for the two-game Minny/Penn St. stretch. Its a totally sensible gameplan against Penn St. in particular who are thin at the linebacker position. However, it would very much be a datapoint against the coaching if we continued to see it this year.
I think another factor behind running non-optimal plays partway through the season is that we want to hide plays from Michigan St. and Ohio. Make them prep for weird stuff, and then throw something completely new that they haven't seen. There are only so many times we can reinvent the offense during the season, and it makes sense to try to have some stuff off-film going into those games.
For example, it would make a lot of sense to me see this team turn into a short passing team that utilizes the strengths of Gallon, Dileo and Norfleet in those games. But if that (or some other direction) is coming in the future, we don't want it on film yet, so it should stay in the barn against inferior opposition. Furthermore, if that is the plan, then you don't want to be throwing lots of bubbles, because you don't want to see press coverage on every play.
|1 year 47 weeks ago||There's another option, which||
There's another option, which is to change up the playcalling. This line has been better at pass-protection than run protection. It seems to me that an emphasis on short passing routes might get some defenders out of the box and make the running game easier. The problems come when the opponent knows we're going to run and we run anyways.
|1 year 48 weeks ago||30 - 10 Wolverines||
30 - 10 Wolverines
|2 years 20 weeks ago||Thank you! You'll always be||
Thank you! You'll always be remembered for your TREMENDOUS blocks and those sweet throwback screens!
|2 years 30 weeks ago||News like this demands||
News like this demands muppets!
|2 years 37 weeks ago||It seems like you'd see a||
It seems like you'd see a bump (over the team average at least) this whenever your worst rebounder shoots, assuming that getting your own rebounds (a la McGary) is rare.
|2 years 37 weeks ago||Couldn't these teams also||
Couldn't these teams also just be the ones for which your method is the least accurate - for whatever reason? You're essentially saying that teams that won more than you expected got "lucky", but what if your model is just plain wrong for some teams?
|2 years 40 weeks ago||31 - 17 Michigan||
31 - 17 Michigan
|2 years 41 weeks ago||Thanks for all your work BISB||
Thanks for all your work BISB - especially in the (ever chaotic) liveblog!
|2 years 41 weeks ago||25-15 Michigan||
|2 years 42 weeks ago||This is obviously better than||
This is obviously better than the straight numbers, but the problem with this approach is that different teams schedules can have very different levels of difficulty. For example, if team A plays a very difficult schedule, they're not going to change those strong teams offensive output as much as team B, who plays nobody but cupcakes.
|2 years 42 weeks ago||Mich 9 Minn 3||
|2 years 43 weeks ago||45-30 M wins||
45-30 M wins
|2 years 44 weeks ago||20 - 12 M||
180 yards rushing.
State moves the ball but can't get in the endzone. Turnovers keep Michigan's point total down.
|2 years 45 weeks ago||Exactly, forward progress was||
Exactly, forward progress was clearly stopped, end of story.
|3 years 23 weeks ago||I actually used pagerank a||
I actually used pagerank a few years ago to pick my bracket. It did well, especially in the first round or so, but only ended up getting 3rd in my bracket pool.
|3 years 24 weeks ago||You can't draw conclusions||
You can't draw conclusions from data with such low R2. The R2 is telling you that the percentage of the variation you're accounting for with this variable is 2-3%. So of all the wild outcomes that can occur on the basketball court, the 3pt attempts can explain less than 3% of the result.
Don't get mislead by the fact that the conclusions line up in a way that you can rationalize - this data isn't conclusive at all.
|3 years 25 weeks ago||Its been fun to watch Ace's||
Its been fun to watch Ace's writing style improve so much over the last few months. The only problem now is that he sounds just like Brian. Given another year and he's going to be among the best sportswriters in the sports blogosphere.
|3 years 25 weeks ago||I just read about it on the||
I just read about it on the internet.
|3 years 27 weeks ago||Exactly. Reductio ad||
Exactly. Reductio ad absurdum: QBs have higher PAN than offensive guards, so we should stop recruiting so many linemen. Clearly this is wrong.
There is a tradeoff between defending the run and defending the pass. Teams that sell out to stop one open themselves up to the other. I think its reasonable to say that one reason Russel Wilson's numbers look so good is precisely because Wisconsin's running game was so dominant.
|3 years 28 weeks ago||For a "meta stat" post, this||
For a "meta stat" post, this is very light on data. Instead of debating the merits of EV vs SR, why not just show some data on which one is more predictive of wins?
I'm looking forward to the case study.
|3 years 28 weeks ago||I must be a super football||
I must be a super football nerd, but when I read the bit about the inside zone minutia my first thought was: "I hope Brian turns that into a post".
|3 years 32 weeks ago||The play-by-play breakdown||
The play-by-play breakdown adds depth that isn't available from summary statistics but I think everyone understands that this is a busy recriuting season, and you're looking for a break after football. Maybe you could focus on ufr-ing "interesting" games (like this one), probably games where something either went really well or horribly wrong, on either side of the ball.
|3 years 37 weeks ago||This is really great - stick||
This is really great - stick with it!
|3 years 41 weeks ago||No discussion of 3rd and long?||
I was expecting some praise for the defense on the 3rd and long plays. At one point during the game I found myself saying "I love watching this defense in 3rd and long".
|3 years 41 weeks ago||Smart football has a pretty||
Smart football has a pretty good list (though not all are strictly football related):
"The Inner Game of Tennis" is a classic with lessons that extend beyond tennis.
|3 years 41 weeks ago||He's a baller. The coaching||
He's a baller. The coaching helps.
|3 years 42 weeks ago||I've can't shake the bad||
I've can't shake the bad feeling I have about this game. Homefield is huge (see Ohio vs Wiscy) and they're going to be out for blood after an embarassing loss. . .
|3 years 48 weeks ago||Happy Birthday Denard!!! On||
Happy Birthday Denard!!!
On my 21st I purchased and drank a six-pack of Sierra Nevada Pale Ale. It was literally at 12:01 when I bought it, and the cashier did a double-take when she saw my id.
Good god, I'm such a dork.
|4 years 4 weeks ago||talent / experience||
With the previous coaches we saw a tendency to play an upperclassman over a less experienced play with more innate talent. I remember a quote about the coaches preferring to know where a player's deficiencies were, and scheming around those as opposed to being surprised (Example: Moundros starting over Demens). How will the new coaches weigh experience vs raw talent - i.e. what do the young guys need to do to get on the field?
|4 years 14 weeks ago||Yeah, that makes a lot of||
Yeah, that makes a lot of sense. Football Outsiders actually breaks down defensive performance into passing and rushing, but I think the responsibilities are too intertwined to split the team's talent / experience into pass and rush.
One thing I might do is to look at the standard deviation of talent on the field - that might get at your point in a round-about way.
|4 years 14 weeks ago||Yeah, you're absolutly right,||
Yeah, you're absolutly right, I shouldn't have worded that statement to imply causation.
|4 years 22 weeks ago||That is a shame. I'll always||
That is a shame. I'll always remember that (game saving?) tackle on Leshoure in the Illinois game.
|4 years 36 weeks ago||Really good suggestion. I||
Really good suggestion. I added a new chart above. It does clean up the data somewhat, but the correlation is still not very good.
|4 years 36 weeks ago||All I'm doing is scraping rivals||
No, time on the team isn't perfect, so if you can find me a good data source for years as a starter, i'll happily use it.
Regarding my intetions, all I'm doing is scraping from rivals:
There's no decision on my part to leave out or include certain players
|4 years 36 weeks ago||Here's what the data for||
Here's what the data for Michigan looks like with my score in parentheses:
ryan van bergen jr. rs (3 + 1)
4+4+5+4+3+2+2+3+2+4+5+5+5+2+4+3+3+2+2 = 64 / 19 ~ 3.37
This is a pretty accurate two-deep at the beginning of the year. In a perfect world, this data would include transfers and injuries already, but since it doesn't we can try to correct for it. Watson should probably be replaced with Patterson (sr rs +1), Emillen with Vinopal (fr -1), and Jones with Demens (so rs +1). We'd get 65 / 19 ~ 3.42.
Ok, so lets go crazy and remove Floyd (which I'm iffy about since he did play much of the year) (-3) and add in Avery, Christen, and Talbott (all fr +1). This gets us below a junior's experience: 65 / 22 ~ 2.95.
I've updated the graph with a second vertical line showing Michigan with injuries / transfers / freshman.
|4 years 36 weeks ago||A very good point||
And something I'd really like to do. However, I can't find a good data source for years as starter. If you like at Michigan's rivals for example:
they only list year and redshirt status. Some teams do have information on player's years as a starter, but it isn't consistent.
|4 years 36 weeks ago||Need a comparison||
If you want to draw conclusions from this you need to provide a basis for comparison. For example, what did this chart look like for three representative years under Carr?
|4 years 38 weeks ago||Thanks!||
Thanks for the comments everyone. I'm going to do offense next, and I'll try to take your feedback into consideration as much as possible.
|4 years 39 weeks ago||Bowl bound||
Michigan is going to a bowl and Ohio State is good. The team is young. 7-5 is an improvement.
|4 years 40 weeks ago||Tate?||
Since we throw with Tate significantly more, I think you have to separate out his snaps from Denard's. Could it be that Tate has been in more when we're down?
|4 years 42 weeks ago||Vinopal||
All three played hard. That stop by Vinopal at the line of scrimmage against Leshoure was amazing. Leshoure has him by 40 pounds, and he doesn't get an inch after the hit.
|4 years 44 weeks ago||approach?||
Nice analysis. Can you give a few more details on how you assessed the team statistics. Sounds like its some sort of linear classifier?
|4 years 45 weeks ago||Odoms => Miller??||
With Odoms out do we know who gets the start? I'd love to see Ricardo Miller get some big plays.
|4 years 45 weeks ago||Worst Denard of the year?||
"I'm willing to wager small amounts of money or pocket lint that MSU is Robinson's worst performance of the year."
I want to believe this, but I'm afraid the worst is yet to come against one of Wisconsin / Ohio State.
|4 years 47 weeks ago||Schilling shoves Shaw||
On the Shaw 12 yard run in the first series, Shaw spins into Schilling who shoves him upfield for at least 4 yards. Awsome.
|4 years 47 weeks ago||Opponent Quality?||
The analysis needs to take opponent quality into account. I.e. if a team has a terrible red zone defense it is less impressive to have a good efficiency against them than it is to be efficient against a team with a great defense.
I have some ideas on how to do this, let me know if you're interested.
|4 years 49 weeks ago||cam gordon 53 yarder||
I've been re-watching this play a bunch because it is really bugging me. Lets say Gordon makes the tackle* - this coverage is still totally broken because Gordon is responsible for two wide-open receivers. At the snap it was clear that there are only two defenders for 3 receivers. No matter what this is going to be a big play. Further, its really hard to see how to take this away. Maybe some kind of Tampa-2 with Kovacs playing the strong safety role? Or maybe the plan is just to load up underneath and hope that most qbs can't make this throw?
Man and do you see how wide open Micheal Floyd is coming across the middle? He could have the first down too.
* I think the camera angle is making it look like he's closer to the outside receiver than he really is. He jumps when the ball is thrown, and still isn't particularly close when it arrives. Just to make the tackle I think he needs to drop a few more steps.
|4 years 49 weeks ago||Cam Gordan and the 53 yard touchdown||
Thanks for this - these are great.
But, I have to disagree with the -5 for Gordon on the 53 yard td. I think it should be RPS -5, as well as on that linebacker (I can't see who it is on the replay). The defensive call probably gives the corner responsibility for the flat. That leaves Gordon with two receivers to deal with. I don't think he can take a better angle, because he still has to worry about that second wide open receiver in front of him, so he can't break until he sees where the ball is going.
I think its just a great offensive call - running two deep receivers into that zone and taking away the corner with the TE. To defend this I think you want the linebacker in coverage to take the TE, and let the corner take the outside receiver with help over top from the safety. Unfortunately, I think we'll be seeing this call a lot all year.
|4 years 49 weeks ago||Awsome, one of the best||
Awsome, one of the best diaries I've ever read - keep it up!
|4 years 50 weeks ago||Size / weight||
I wonder if there's a correlation between a quarterback's size and the amount of injury time they're forced to take. Guys like Vince Young and Tebow are much bigger than Denard.
|4 years 51 weeks ago||Safety play||
Watching last year, it seemed to me that the linebackers were hesitant at least in part because there was essentially no-one behind them. If we have just a little bit of consistency in play at the free safety spot I think the linebackers will be free to think less and commit to a gap.
|5 years 7 weeks ago||Nice work!||
Two comments -
1) It'd be great to see error bars on the PBRM Results.
|5 years 8 weeks ago||Defensive Size||
Bud's analysis is begging for a scatterplot and some linear modeling. I'd be happy to do it if someone can point me to the raw data he's using.