|11/01/2017 - 5:40pm||I see that the OL had a good week...||
Unprecedented in fact. Wow
|10/25/2017 - 5:43pm||Only have to....||
... go back to 2014 for that coverage number. The saddest thing about the below chart is that the coverage against PSU was basically the coverage numbers every game in 2009.... Ouch....
|10/25/2017 - 12:45pm||That's rough but.....||
at least there is plenty of time left to offer to others or potentially bring him back into the fold. Good luck to him if he does go elsewhere, but we'll get someone else to be excited about in his spot.
|10/16/2017 - 3:17pm||With our D||
we want the opposite of a high variance offense. If we want high variance, stick with O'Korn and hope his close plays turn out more like the Purdue game than the MSU game.
If Peters can operate a safe, basic passing offense and we can couple that to the run game of the last two weeks, that is a much better fit for us right now. No indication yet that he can. If we're debating two high variance options, the coaches seem to think O'Korn is the better one and I'll trust them on that one.
I'll keep my fingers crossed that Peters has started putting things together in practice and that he breaks out to take the job from O'Korn ASAP.
|10/16/2017 - 1:35pm||Purdue O'Korn was the same as every other game||
O'Korn is a high variance player. He doesn't make great reads consistently, he throws into very tight windows, he bugs out at pressure and runs around trying to find somewhere to make a tough throw downfield. He did this against Purdue as much as against Indiana or MSU.
Against Purdue it just worked because he hit on 17 and got a 4 instead of a J. He was in a tough place a number of times and managed to have things work out so that he could make the play. This throw in a tight window was completed whereas other week's it's broken up or picked off: https://www.youtube.com/watch?time_continue=10&v=moYO8YubAYw Similarly, the play where he was almost sacked by Purdue, but spun away scrambled around the backfield for a second and then found Perry downfield was a good play, but also a lucky one.
O'Korn lives on a very thin edge as a QB. Some games he'll come down on the good side but he's as likely to gamble away a big plays for the other team.
|10/13/2017 - 3:14pm||Holy Hell Hurst||
I never expected someone to break Mike Martin's +25 against UMass, even less to do so against a non-Rutgers BIG opponent.
|09/22/2017 - 11:10pm||It's definitely not cats||
that's for sure.
|09/21/2017 - 5:15pm||So of course I post this||
and Brian switched back to one big table for the offensive plus/minus instead of separate tables for OL and Backs/Receivers, which screwed up the web scraping script. It was a small tweak to fix it and now it should work for automatically updating in both cases.
|09/21/2017 - 5:04pm||Sorry||
I dont know how to do that on images in the diary post. Feel free to go to the link and the images are much clearer.
|09/20/2017 - 4:39pm||Hurst current max is||
+13.5 against Cinci this year. His best last year was +13 against Maryland.Wormley had a +17 against Iowa and +16 against UCF last year.
Go to Defense > Top/Bottom 10 > DL > Game at the link below to see the top 10 DL games in UFR:
|09/20/2017 - 4:04pm||Winovich with the #2 all time UFR single game score for a DL||
behind Mike Martin's +25 game against UMass in 2010. Craziness!
|09/18/2017 - 10:06pm||I think 2timeloozer||
is commenting that, if DPJ is the first read, why doesn't he throw it to him? He looks bracketed in the still, but in motion he looks like he has space.
By the time Speight's read turns to DPJ, he has separation from the LB behind him and a step on the safety as well. He has an LB in front of him, but the LB was trying to bail to the wide side of the field and was in no position to make a play on a throw to DPJ unless Speight throws it right at the LB's numbers.
In fact, the last thing Speight sees before turning to DPJ is the LB opening his hips toward Perry to head to the wide side of the field. If the design is to go to DPJ no matter what's going on with Perry/Crawford, then I'm surprised he doesn't try to hit the window there. It's not a huge window, but he's got space and it's not a long throw.
Given where the other safety must be in terms of bailing toward Perry/Crawford, I think DPJ has a really good chance of taking that to the house if Speight hits him in stride. Maybe I'm overly optimistic there about his chance to get away from the trailing safety, but still....
|09/15/2017 - 6:50pm||As an example of the last idea, for this UFR:||
Punt 1: from M32, 56 yards in air. Hart +11
Punt 2: from M28, 68 yards in air. Hart +23
Punt 3: from M42, 45 yards in air. Hart +0
Punt 4: from M15, 58 yards in air. Hart +13
Punt 5: from M14, 32 yards in air. Hart -13
Punt 6: from M24, 31 yards in air. Hart -14
Punt 7: from M29, 58 yards in air. Hart +13
Total: Hart +33 on 7 punts
45 yards per kick might be a little too easy, maybe 50 is better. But still, I think that idea works out pretty well for summarizing the punter's field position impact.
Also, none of the kicks this game were really ones where he was likely trying to take something off to pin them. Maybe punt 3, instead of being a +0 for kicking 45 yards, should be either a -3 or +2, for pinning at the 13 instead of the 10 or 15 respectively.
Just a thought.
|09/15/2017 - 6:41pm||Looking forward to more of these!||
A couple thoughts:
Why the blank line in the table for the botched return? I was surprised not to see any mention in the table.
The distance for FGs would be nice in the FG tables, even if it's somewhat inferrable from the location of the kick.
I love the summary of players return impact being measured in gained vs lost yards. A way to adapt this for the punter could be to estimate a "reasonable" outcome for an "average" college punter, then give the punter credit or blame for the difference from that. Obviously average and reasonable are pretty subjective, but it would give a way to get an idea for their impact on field position.
For instance, at a punt from out own territory, maybe a 40 or 45 yard punt would be "reasonable" so that a 60 yard bomb gives the punter big credit and a 20 yard shank gives correspondingly poor blame. On the other hand if we are punting from their 45, maybe a "reasonable" punt would pin them at the 10 or 15 yard line so that pinning them inside the 5 gives a big + but a touchback or worse would give a -.
Just some thoughts! Awesome to see!
|09/14/2017 - 4:50pm||Where's the Jake Ryan love?||
He seems to be being forgotten as an impact LB in recent history. He was outstanding before the knee injury and still solid, though seemingly playing out of place, after he came back.
|09/13/2017 - 9:14pm||Disagree with your assessment||
about momentum being very different there. At least for the DL. #96 is inside of Ulizio with about the same momentum on the Issac run as #44 and #48 have in the Higdon run. Both are at the LOS at handoff and in postion to hit the RB 1 yard deep.
Higdon makes it seem that way because of his momentum.The difference is that Issac immediately throttles down and hard cuts backside. He gets the ball 5 yards deep and has slowed to cut in ~ 1.5 to 2 yards. Higdon gets the ball 5 yards deep and starts trying to slow/cut after 3 yards, so his momentum takes him right into the penetrating DL.
It might be that Higdon's aggressiveness to the hole is just right for him, and while he misses an opportunity here, it'll allow him to break other plays later. It also could be that his play was meant to more aggressively follow the lead block and Kugler/Ulizio just failed. Either way, his momentum was what kept him from being able to have a chance to read the cutback.
|09/03/2017 - 5:27pm||If they benched Kadarius Toney||
because he missed that block on Hill, they messed up. That was clearly a designed CB blitz and Toney had no prayer of making that block other than a pretty blatent block in the back.
Toney did the right thing to go block Kinnel and hope someone else, probably Perine making a play as the RB, deals with Hill. If he goes and blocks Hill in the back, he both likely draws a flag, and has Kinnell unblocked cleaning up behind Hill, likely holding the run to minimal gain excluding the obvious flag.
|08/30/2017 - 2:57pm||Dwumfour||
I don't get the poor recruiting ranking for Dwumfour. I mean, I get that a highlight video is just that, but he looks incredibly quick and mobile for a guy his size. The type of plays he's making look a lot like Hurst and generally scream potential star and disruptive force.
I get that the competition may not have been the best, or he may have been really inconsistent or something, but watching his senior year Hudl and that clip from the Hawaii game, I just see an excellent high-impact DT in Dwumfour.
|04/15/2017 - 12:43pm||Good point||
I've gotten used to scraping the data automatically so my brain automatically went to that method of acquiring the data. You're absolutely right though that copy/paste from tables is way quicker and more sensible in this case.
Also because you asked, I use the rvest package in R or beautifulsoup in python. I dont have a particular good tutorial offhand, but if I look and remember one I'll post it back here.
|04/15/2017 - 11:10am||If you know how to scrape data of web pages||
then going team by team from here seems doable:
or if you want their full contact amount instead of just their 2017 salary:
I agree with wesq's advice though if you don't know how web scraping tools work and don't have time to manually copy down all the team salaries.
|04/08/2017 - 2:32pm||Sorry that the plots are squished here||
They look better if you open in a new tab or even better, look at them at the link provided
|04/08/2017 - 2:30pm||More RPS plots||
You can look at the RPS data over time at this link: https://mikegros.shinyapps.io/ufr_analysis/
by going to Single Player -> Team -> RPS for either the offense or defense
The offensive OPS has definitely taken a step forward under Harbaugh, while the D mainly improved after Rich Rod's departure, as you'd expect. The lack of D RPS this last year is likely because the players got most of the on-field credit. Informative plots below:
First with the Offense:
Then with the D:
|11/24/2016 - 3:34pm||O'Korn criticisms||
A couple plays that O'Korn is criticized there seem to only be a little on him.
On his horrendous case of happy feet, Cole gets beaten badly straight up the middle, so O'Korn can't step up in the pocket. I don't know how Cole doesn't get a -1 or -2 for allowing immediate pressure right up the middle. O'Korn fades back and that causes Kalis to lose his block, so I wouldn't put that added pressure on Kalis. O'Korn does well to keep his eyes downfield and find Darboh there. He probably shouldn't have faded so far back, but he did okay with it.
On his running play, I don't see how he could go anywhere there - there was a trailing defender right behind Bredeson. Slowing up to either use Bredeson's block either to cut back or to fake inside to set Bredeson up and then head outside, has a good chance of getting caught by 17 there. Even if he avoids 17, Smith misses his block and there is another unblocked linebacker within 3 yards of the LOS, so he maybe turns -3 into 0. I don't see what he's expected to do there.
|11/24/2016 - 12:32pm||My code is...||
... up on github at the link below. It's not really well commented at all and probably not the most efficiently written, as I was using this as an excuse to try out Shiny as well. Feel free to check it out though!
|11/14/2016 - 2:00pm||You are right.||
I retract my comment. Bredeson had the other DT and still tried to block that LB that Hill took. I missed Cole releasing to the second level when I first watched, so thought he had Bredeson's guy and Bredeson was peeling off to take the LB.
|11/14/2016 - 1:53pm||On the safety...||
if Hill executes his wham block on Johnson instead of blocking the same guy Bredeson's got, Smith is one-on-one with a safety 10 yards downfield. I don't think that one's on Kalis.
Too many mistakes to win Saturday.
|09/29/2016 - 4:33pm||That Perry route++||
.... some credit needs to go to RPS. Perry route starts cutting underneath Butt in what appears to be a nicely designed rub route. The safety jumps hard inside to avoid falling behind Perry just as Perry cuts out.
Perry executes it perfectly, but that has to be at least an RPS +1 as a counter to the D's counter to rub routes.
|09/22/2016 - 1:36pm||Seriously?||
Excellent teams get upset all the time in games they should win. As it stands, we should win this game solidly, BUT Penn State has talented football players and, as we saw against Colorado, our team is not immune to making mistakes.
Should we win? Yes. Should we win handily? Yes again. But we aren't guaranteed a win. A couple big mistakes or some bad luck can turn a likely win into a toss up. Hell, even last year we were a play away from losing to Minnesota (and Indiana but CHAOSTEAM explains them).
|09/15/2016 - 4:50pm||About that DL||
They crushed the previous record. Wormley had the 6th best UFR score for a DL on record. Glasgow was tied for 11th (though did not set a personal best, being +15.5 last year against Minnesota), and Winovich was tied for 16th best.
Glasgow and Wormley combine to be 7 of the top 15 UFR scores for DL all-time and approach Mike Martin/Brandon Graham levels of awesome. (the 4 of them make up all but one of the top 15 UFR scores for DL all-time.
Link to larger image: http://imgur.com/a/SuukX
|09/14/2016 - 5:13pm||The player coming around Poggi||
is irrelevant to the cut Brian is referring to. Brian is commenting that he should cut hard to the right, into the hole that Asiasi went into. This is away from Poggi entirely. The backside DE and OLB have been kicked out leaving a big hole. Kalis is looking for someone to hit and Asiasi would be able to pick up the ILB that's heading to the hole that Smith actually heads to. That ILB is committed there and would be likely unable to recover.
You could say this is too hard of a cut, but Brian correctly points out that Smith made nearly that same cut against Florida in the next clip below. Florida fills the frontside A gap, so Smith cuts hard into open space on the backside.
|09/12/2016 - 4:18pm||I think Brian meant the gap behind Asiasi||
I could be reading it wrong, but it looks like Smith has 2 potential lead blockers there for the double iso - Poggi and Asiasi. He heads toward the hole off Poggi's butt, but if he reads the LB and cuts to the gap to his right, he has a big hole and Kalis has realeased is looking for someone to hit upfield.
Again, maybe I'm reading it wrong but that lane seems to be designed to be there and is definitely open when the LB crashes the lane that Smith ends up taking.
EDIT: Didn't see someone else posted this already because I hadn't refreshed my page since opening it earlier.
|09/07/2016 - 6:14pm||UFR visualization tool fixed||
Looking forward to it breaking again tomorrow for the offensive UFR.
URL here: mikegros.shinyapps.io/ufr_analysis/
|09/07/2016 - 3:46pm||I love the additional of the snap count for each player||
though the new table format broke my scraping tool for the UFR visualizations. I'll fix it soon and look forward to adding some new features given the snap data.
Awesome writeup as always!
|09/02/2016 - 2:40pm||I dont think so||
Wormley graded out better than Henry last year on average for the UFRs. He's very very good.
|09/01/2016 - 4:52pm||Haven't found one in Vancouver||
in the couple years I've been here. I tried asking in the other "Where to watch the game in ______?" thread the other day but got no response. I don't spend a lot of time downtown though (live in East Van) so that may be part of it.
If I hear of something I'll post here for you.
|08/29/2016 - 11:48pm||Vancouver, BC||
Anyone else across the border in Vancouver? I know few bars are open at 9, but curious if there are any other fans around to gather.
|08/23/2016 - 2:49pm||Yeah...||
... I'll fix that soon. I scrape the year from the URL for the UFR post and so the bowl game was placed in 2016. I'll have to fix that by hand, as well as the fact that the URL doesn't have a year for the OSU posts but I'm waiting for the defensive UFR for the OSU game to go up first to fix it all at once.
|08/23/2016 - 2:46pm||The colors are by season||
The plots are in chronological order by game, with the color changing each season. Currently I haven't fixed the fact that the Florida bowl game was read as 2016 and so not placed in the same color as the 2015 season, as it probably should. I'll fix that as soon as the OSU defensive UFR goes up, as I'll want to make sure they are ordered properly also.
(I'm not Ace but made the graphing tool for the UFR data found here: http://mikegros.shinyapps.io/ufr_analysis/)
|07/29/2016 - 12:32pm||As much as it would be ugly.....||
..... and like slowly swimming through a pool of glass shards for Brian, I'd really like to see an OSU UFR.
Mainly because it would be informative for player and scheme strengths and weaknesses for the upcoming season, as well as good for adjusting expectations with respect to the OSU game this year. For instance, I'd like to get a feel for how much defensive blame can go to: (a) Durkin, (b) Bolden, (c) Lack of Glasgow.
Any chance it happens?
|11/20/2015 - 1:14pm||Code on github||
Have at it!
|11/20/2015 - 1:13pm||My code is up on github||
I recently "followed" you on github. I don't have a ton of extra time right now, but I would love to connect and really put together some good work as much as I'm able.
|11/20/2015 - 12:45pm||I'll warn that...||
.... I'm not always great about writing clean code, but I try to use pretty clear variable naming and "enough" commenting. But I'm sure it's not actially "enough". I was using this as an excuse to learn/practice Shiny so I hope it can be good help to you too!
|11/19/2015 - 4:16pm||Indiana seems like the BIG version of Air Force||
in the way that they are just an outlying team to prepare for and can really screw with your views of your D. I wish there'd been more charts for the OSU losses in the last couple years, because it would be interesting to see whether bad showings against IU translate to bad showings against OSU or if IU is still the big outlier - especially for our DL and our RPS.
Especially two years ago and so far this year, Indiana has been completely out of line with the rest of the season numbers:
|11/14/2015 - 2:34am||Really nice!||
I'd love to see more.
|11/13/2015 - 10:14pm||Consider your case lobbied for....||
I responded in the other thread too, but sounds great. I will get on it soon (when I can scrape up the time).
Thanks for the idea!
|11/13/2015 - 10:10pm||Great idea!||
I should probably clean the code a little first, but I will try to share it as a github repo soon (I'm actually really annoyed with myself now that I didn't put this under version tracking to begin with.... sloppy).
I have a lot of other things on my plate at the moment, but I will try to do that soon because I'd love to see the tool keep improving.
Thanks for the idea!
|11/13/2015 - 6:21pm||Thanks for the shoutout||
I've added the ability for some sort of normalization for number of plays to my list of things to add/fix. I'd really love to normalize each players numbers by playing time, but I need to get individual playing time data per player. I'll look at it in the future.
|11/13/2015 - 10:47am||Brian didnt....||
.... chart +/- for offense for a long stretch. He only made the Hennechart and then at some point started the receiver chart before adding the +/-.
|11/12/2015 - 7:06pm||Not really....||
... but I could justify it to myself as "professional development" as it's good to practice web scraping and data cleaning; plus it was a good toy to use for learning RShiny.
It's also just fun data to look at.
|11/11/2015 - 3:41pm||A few things....||
- I'm a little concered about Hill, he's had a number of rough weeks in a row. I guess that's life as a safety who isn't boring, but it seems like opposing offenses have been attacking him pretty effectively as the weak link in our secondary.
- Morgan, other than his epically bad week last week, is having quite a year. He's a good counterpoint to the argument that Bolden's numbers look rough only because LBs are having all the pluses stolen by the awesome d-line:
- Wormley had his first day with any real noticable mistakes in 2 years. It's great having such awesome D-linemen:
There are more plots at the link below. Plots are up to date but the text probably isn't (also forgive my poor writing skills):