Ultra-UFR Proposal

Submitted by WolverineBlue on July 18th, 2011 at 7:37 PM

This is an idea I started kicking around last year and actually built a crude prototype. I think most of us would agree that the UFR (Upon Further Review) is MGoBlog’s crown jewel of features during the heat of the Michigan football season. I had some thoughts on making it even better by enhancing the presentation and adding filtering options. Before I elaborate, I suppose I ought to jump right out in front and issue the following disclaimer:

This proposal is in no way intended to make Brian’s life more difficult. My crude prototype simply took a current UFR, parsed it and reorganized the information to make the viewing process more powerful and flexible. I would propose that any additional work associated with such a project be undertaken by volunteers such as myself (I probably would need at least some help: I am a software engineer, but not primarily a web guy). The possibility even exists to make Brian’s life easier by including a powerful UFR creator-editor that streamlines the creation process.

Let me also make it clear that I have received no form of pre-approval for this idea, so it could get shot down before lift-off and, of course, would not move forward without the full endorsement of the MGoBlog staff. With all those qualifiers out of the way, here is a quick overview of my Ultra – UFR vision:

1) It could be a standalone application, but would probably work best as a web app hosted directly on MGoBlog. Imagine being able to go to a special UFR section, picking a game, and digging in to the juicy details.

2) Each play description would be linked to video. This depends on the availability of a streamable version of the game and being able to start playback at an indexed location. Karpodiem’s efforts with regard to MGoVideo would seem to be the ideal solution to this part of the project, although I have not yet contacted him regarding feasibility or his potential interest.

3) Standard navigation would let you step through one play at a time, with a scoreboard type game status display similar to the kind of real-time gametracker you would find on MGoBlue.com, ESPN, etc. Each play would have all the information in a standard UFR table (including Brian’s commentary) plus a video link to watch the play within a window of the Ultra UFR (or full screen if preferred).

4) There would also be summary drive charts to quickly zero in on plays of interest at any point in the game.

5) Playlists could be filtered based on UFR data. For example, “show me all the plays where Taylor Lewan received a + rating”, “show me defensive plays with negative yardage”, etc.

I’ll stop there with any detailed descriptions, since the exact specs could be worked out later by the project team with feedback from MGoBlog staff and users.

The proposed timetable would not be overly ambitious and this season could serve as a trial/beta period for developing the concept. Perhaps by the Ohio State game, there might be a serviceable version of a near-complete application. The grand experiment would not interfere with standard UFR publication.

So that’s the broad view of my idea. If there is a general “meh” reaction, then it would clearly not be worth anyone’s time to further pursue it. If, however, there is some genuine enthusiasm, some official endorsement by MGoBlog, and some willingness on the part of qualified individuals to help out, I will follow up and see if we can get the ball rolling.



July 18th, 2011 at 7:54 PM ^

I've done some parsing of the UFRs to try to get them in a MySQL database format. I've been busy lately, so I haven't made much progress since cleaning up the HTML to a point where it could be used to generate MySQL commands. I'd gladly pitch in, and I've got a decent amount of experience with web programming.

Getting the data queryable with some basic filters would be pretty easy (and infinitely easier if the data was added in some standard format, possibly through a web app like you mentioned... player +/- comes to mind as something that's really tough to tease out of the raw HTML data).

Feel free to shoot me an email (same as my twitter handle in my sig @ gmail or umich). We could probably set something up on Github or Google Code or something to keep things orderly.


July 18th, 2011 at 7:59 PM ^

Thanks for your offer to help. I'll send you an E-mail to discuss technical issues. If this thing gets off the ground, I will definitely be looking to establish an online code development environment and am interested in recommendations as to which one people have found to be the most useful.


July 18th, 2011 at 7:50 PM ^

I like the idea. Would also be cool to search by down and distance if you wanted to see tendencies
<br>It would also be neat to search by play name or formation. Then you could see which plays are successful against different looks
<br>Lots of different possibilities, and as a data miner, I'd pretty much love it.


July 19th, 2011 at 10:16 AM ^

No, I don't think coaches have time on a weekly basis to sort through UFRs and then try to interpret the data. Having an  app that was able to parse through the noise would make it worth spending an hour or so looking at third and short tendencies.


July 18th, 2011 at 8:01 PM ^

This would obviously be awesome but sounds like a lot of work.  Then again, I know nothing at all about computers so it could be easier than I think.  I've been spending about a week trying to figure out how to highlight in a pdf and I still can't figure it out.  Maybe it's unpossible.


July 18th, 2011 at 8:29 PM ^

I want to go to there.

Although, I can only offer my +1 to your post as I haven't the foggiest about programming or if that's what this would even be called.


July 18th, 2011 at 8:32 PM ^

I love it.  While the Blue Community is creating state of the art football analysis software, the Ohioans are flipping people off and running into cop cars.

How in God's name did we lose to them so many times in the past 10 years?

I've never seen any threads that featured any analysis of Tressel's dominance of Lloyd.

I'm curious but I think I might just prefer to forget about that season of my life and look forward to what Hoke, Mattison and Borges can do.


July 18th, 2011 at 8:35 PM ^

September: Denard straight ahead for many

October: Denard straight ahead for some

November: Derp



July 18th, 2011 at 8:36 PM ^

On a serious note:

This may be a crucial step for the blog.  If you can get a UFR database up and running as a web app on the site, that would really solidify MGoBlog as a premier Michigan Football resource. 

Not that MGoBlog isn't the premier M site, but if we lose TomVH, MGoBlog loses a valuable portion of its site that cannot be found elsewhere.  A UFR database may help fill that hole.


July 18th, 2011 at 8:36 PM ^

If we get this good at UFR'ing, I think opposing teams will have an ultimate database of exactly what we do, without having to do any work themselves.


But seriously, maybe this is something we could crowdsource, a la the way Google got 3D buildings so quickly on Google Earth. 


July 18th, 2011 at 8:40 PM ^

It's entirely possible that if we can get some sort of webapp for entering UFR data, Brian could get users to do some of the legwork of linking up videos, etc. I know I wouldn't mind watching through a game and keeping track of where plays begin/end.


July 18th, 2011 at 8:59 PM ^


My thought was that if the concept was approved but Brian wanted to steer clear of any personal participation, volunteers could use the editor to tag the start and end segments of the video.

On the other hand, if we had a kick-ass editor with built in video preview, this might be the easiest way to create the UFR in the first place and the video indexing would be mostly automagic.


July 19th, 2011 at 8:52 AM ^

have a crew of grad-assistants that break down film like this already.  After my playing days, myself and a few other GA's would be up all night sunday breaking down tendancies and packaging film for the coaches to use on Monday....and this was at the DII level.  I'm sure they have a refined system at the DI level


July 18th, 2011 at 8:51 PM ^

You have my support, and you'll get the all the video you need for it. Let me know how I can help; a k a r p o @gmail.com


If you could do a quick and dirty prototype with RoR, you might find some people interested in lending a hand. You'd probably have to pour 100 hours (most of which would be parsing data) into this to get something really rough that is production ready, but it would be a start.


July 18th, 2011 at 9:06 PM ^

I greatly appreciate your support. I will e-mail you with a few quick questions relating to the video and some technical issues I encountered in my first prototype. I already did a first cut at parsing the data, so I have something of a head start on that task.


July 18th, 2011 at 8:58 PM ^

Sort of. The thing to be mindful of is that the gist of UFR is Brian's opinion as to what happened on any given play. Who did well and who effed up is in the eye of the beholder and enhancing an Mgoblog UFR is putting refinement to editorial content, which may not be useful, particularly since the video itself doesn't tell the whole story.

What would elevate the art would be better ORIGINAL video, as opposed to better editing of available video.

I mailed Brian about the availability of full-22 video. It is not. Available. The only people who have it are the teams and by extension, the league. And they don't share it. Damn them. There's a camera on the roof, right over the 50, that's recording everything. And we can't see it. So when Brian critiques the play of a safety or cornerback, a lot of times that player isn't even in view for said critique.

I applaud Brian for what he's done but he's always had to do it with less than full vision.


July 18th, 2011 at 10:52 PM ^

This.  This whole idea sounds somewhat like my college team had... but better.  Being able to parse through plays with that level of detail would be fantastic.  However, without the proper angles (endzone views for proper OL/DL evaluation, wide angle view for WR routes and DB coverage) this tool would be a bit wasted.  I still believe that this would be a fantastic addition and would love to be able to help with such an idea.


July 18th, 2011 at 10:51 PM ^

But the weekly schedule during the season us pretty full. I think offense ufr is wednesday, defense is Thursday, preview is Friday. And I think Brian kills himself doing that.


July 19th, 2011 at 12:08 AM ^

But the premise is that this would not add to his workload. Best case scenario: we could provide him with a nifty tool that would significantly cut down the time required to create the UFR and both regular and enhanced versions are released simultaneously. Worst case scenario: after the UFR is out, someone else runs it through a utility to parse out the content, manually index the video links, and post the enhanced version shortly after the regular version comes out.

M - Flightsci

July 19th, 2011 at 1:51 AM ^

This sounds to me like it would take a team of developers a year to create, at least for the implementation I have floating around in my head.  That being said, I was surprised how quickly the Android/iPhone app came to market, and how relatively well it works.  It has been interesting to see how UFR has evolved since the old blogspot days.  Look here for an example: http://mgoblog.blogspot.com/2006/10/upon-further-review-defense-vs-penn.html.  I sorta miss the old website.  Also, haloscan was pretty awesome, if only for the more interesting commentary generated (chunks the hobo?), and for the fact that Brian would frequently chime in with some hilarious gems.  The addition of video was a helpful tool instead of forcing the reader to TiVo everything, or just pull it from memory, a new facet to a great product.  Another commenter mentioned pulling this off would solidify mgoblog as the premier Michigan football blog.  I'd contend this would make it the premier college football blog, and possibly best overall sports blog too.


July 19th, 2011 at 2:49 AM ^

Perhaps I'm delusional, but I really don't think it would require anywhere near that level of effort. It's not necessary to reinvent the wheel at every step. There are html parsers and video players out there and much of the task consists of assembling available components.. My brief experiment last year was able to extract usable information from published UFRs, reformat it, and display it on a web page with game video (not synched, though). I spent about 10 hours on the project.


July 19th, 2011 at 8:22 AM ^

It's a great idea with one substantial legal issue.

Including every play ripped from a broadcast feed would almost certainly violate fair use provisions of copywrite protections. Brian already had to take his clips down a few years ago b/c either vimeo or youtube (can't recall the service) found them in violation of ther TOS. He's since found a more amenable means to post a limited number of clips, but mgoblog's prominent enough that posting essentially the entire game would almost certainly result n a cease and desist letter from ESPN and the BTN.


July 19th, 2011 at 10:31 AM ^

This is an awesome idea.  I love to review data and it would be great to have a tab with all of the UFRs in some searchable/downloadable format.



July 19th, 2011 at 1:17 PM ^

I have one suggestion.  A section for reader input, whether wiki-style or stream of mgoconsciousness for each play would allow the community to chime in with thier own insights, etc in a way that has not been possible outside of real time live blogs.

I'm imagining two scenarios.  One, Brian UFR's a play and says, "I think ...(something)... but the coaches may say something different in the comments"  Let the coachy types comment right there, specific to that moment, and make us all smarter.

Two, a certain moment in the game induces a seismic shift in someone's personal philosophy, or maintains their status quo, which was previously in considerable peril.  Like when Greg Mathews caught that TD on the wheel route against State in '07, or Braylonfest; or conversely when Appy State and then Dennis Dixon tore away the curtain earlier in '07; some enlightened mgouser probably remembers, "It was at this point that I realized ..."  Let us all witness the moment and perhaps be enlightened ourselves.

Just a thought.  And speaking of considerable peril: