Einstein

Absolute flag fest

Thank you! I’m happy to send you the project files that you can run locally on your computer. But you will need to have an understanding of R and running servers. Or you can use the website version I will publish this upcoming week. I think i’m going to take the advice someone gave earlier in this thread and have an advanced option for folks like you who want to drill down deep, and a simple version for those who want access to quick info. I just don’t understand why. The data is updated near constantly throughout the game. In fact, I’m thinking about adding a feature that rapidly tracks every play the Bills run throughout the game, producing the EPA for each, and predicting the next. So I dont know why they wouldn’t have that data by now.

Literally anyone but Cinci. They are the only team that I believe has a legitimate chance to beat us in the Wild Card round. Burrow, Chase, Higgins, Brown & Hebert would be a nightmare for our defense to defend.

I also dont think we should have traded with KC. Even if Worthy was a bust, I dont think we should trade with KC. I just dont think you trade with enemies - ever. That being said, look at Worthy's efficiency metrics. They are tied in Rec/Target and Coleman has him beat in Proj. Yards/Target and TD/Target

Sure does. Click the image to make it bigger.

Yep..... but i'm a little scared to put that feature on the online version for everyone. In order to use those calls it scrapes PFF data. I'm uncertain how PFF would respond to that being public.

interesting. Today I learned to never blindly trust Pro Football Reference. Why? Pro Football Reference shows McD with 13 challenges over the past 3 seasons, with 5 of those 13 being wins. My program shows 14 challenges with 6 wins. At first I assumed that my program was wrong, because why would PFR have incorrect data? But then I pulled the actual play by play data, and it turns out that my program is right and PFR is wrong. There are 14 challenges by Buffalo over the past 3 seasons and we won 6 of them.

Added: Daboll with a surprisingly high 70% win rate. PS, with all charts, the team logo can be off by a couple % points. Its minor, but just FYI.

Nope. Shiny is great but a bit less flexible. I prefer to have a seperate front-end that I can customize to my hearts content with javascript and HTML. I can add a “simplify” button that will change the page (remove features). But what are you thinking as far as simplified options?

And all complaints go to Simon!

Haha I totally get that. I mainly wanted to ask for suggestions for any additional features before I get it all transferred. There is an NFL repository for data called NFLverse. Here check it out! https://github.com/nflverse/nflverse-pbp ESPN also used to have an API. They pulled it though.

Site is now active. https://einsteinfootball.com (just registered the domain so not sure if its propagated in all areas yet) or http://34.205.139.226/index.html (this is direct to the server so it definitely works) The site does NOT work on mobile. There is simply too many tools and features that dont work well on small devices. Access it from a desktop or laptop. --------------------- I am making available to everyone my free Bills statistical application that I built and use throughout the season. The backend is coded in R and runs on a plumber server with a CMD prompt server for the front end. For everyone else who wants an easier way, I bought a domain to host it on. I will have it loaded fairly soon. In the meantime, this is how it works: 1) Play Finder This is my favorite feature. You can: Load all plays Load only run plays Load only pass plays Load plays that feature a certain player (example: plays only involving Keion Coleman) Load plays that are more than X yards (example: plays that are more than 10 yards in length) Load plays that are less than X yards (example: players that are less than 10 yards in length) Load plays between X and Y yards (example: plays between 8 and 35 yards in length) Then, once those plays are loaded, you can filter them even further. The "additional filters" section allows you to drill down into exact situations. For example: You can load all plays for a specific game and drive within that game (Game 3, Drive 4). You can pull the stats for run plays depending on which run gap they exploited. You can filter plays by air-yards. You can filter plays to see how the team plays in various temperatures and wind and field types. For example, run plays in 52 degree temperatures and on astroturf. You can examine Allen's stats when he gets hit. Or break down offensive performance when in Shotgun vs when Allen drops back. Filter to plays between X and Y epa. A lot more... Not only that, but it will also show the stats for that particular filter/search. For example, if you search for Josh Allen passes in the 4th quarter, and only on plays when he gets hit - you will not only see all those plays listed in the table, but the stats widget will also show you the fully tabulated stats for when Allen gets hit in the 4th quarter (yards, INT’s, TD’s, etc). 2) QB to WR connection data Pretty simple. You can select any player who has caught a pass from Josh Allen this season and it will show you Allen's stats when throwing to that receiver. This gives insight into what player is most beneficial for Allen. Unsurprisingly, Shakir has the (far and away) best connection with Allen on an output basis. Know whose second? You may never guess. It's Ray Davis. He has caught 14 of his 15 targets and has 2 TD's on those 15 targets. Valdez-Scantling is on the opposite end of the spectrum. And dont even get me started on what this tool has shown for Gabe Davis over the past few years. 3) Bills AI Chat You can chat with a Bills artificial intelligence and ask historical questions. I'm making a few tweaks to this, as it tends to decide its own tiebreaker when the answers are close. For example, if the GIF below, the answer really should be AJ Klein with 5. But it awarded the honor to Hughes with 4.5 in some sort of small-differential tie-breaker scenario. 4) Lots of Charts You can generate a lot of different charts. Some examples: Where the Bills rank against other terms in EPA (offense/defense). Quarterback EPA per play across the NFL Win to INT plots Win to QB rating plots Plotting of teams against each per division A lot more This one requires patience. These are not pre-generated charts. When you click one, it generates in real-time, pulling NFL data, doing the calculations and creating the chart. A chart can take up to 3 seconds to load. 5) Suggestions/Requests Its going to take a few days to get everything onto a hosted server. I need to change lots of little bits of code that run on my local computer to running on a host. In the meantime, if you have any suggestions or requests to make it better, drop them in a reply. I would like to add vector information (speed/direction) but have not found an API to pull this data. Next Gen stats doesn't seem to have an API.

Yeah, I’d still go Daboll. Brady is giving him a run for his money with this recent streak, but the offense was just effortless with Daboll most of the time.

Sounds like they’re pretty big. But I don’t really care if they only had a circulation of 3 people. They were simply quoting a legendary coach. The publication that quotes that coach is irrelevant. To try to bring relevance into it as a variable that is meant to mean anything is just spin.

The article writer wasn’t giving any expertise. They were simply quoting the greatest coaches to ever do it. Lacks context? How much more clear could Belichick be? What more context would you need? “I don’t know how you can be comfortable giving up 40 points every week and feeling good about where your future is… regardless of what their record is going forward. If Buffalo can’t play better defense than they’ve played recently against good teams. I don’t think their season will last much longer.” Belichicks quote has all the context you need wrapped right up into it. He is directly stating that our defense is not playing well and it will lead to our seasons demise.

That’s a terrible spin job Beck. The article publisher is irrelevant. The point is that one of the greatest coaches in football history… the man with more SB rings than we have playoff wins in the past two decades… made a comment that was fairly unflattering. Trying to spin it as a joke because a rando website quoted it is poor form.

Like all things NFL, when you play a team is very determinant of your chance of beating them. They played the Ravens for the first game of the season during their 0-2 opening garbage stretch when the Ravens looked awful and lost to the Raiders! They played the Bengals during their bad opening stretch when they started 1-4. And the Chiefs still needed some handy ref calls. They played the Texans down Diggs and Dell. Their best two receivers. They were a blocked field goal from losing to Denver. .

No it was a good penalty. Lifting your shirt like that is a signature that you’re showing someone the gun in your waistband. The NFL has consistently flagged that move.

I genuinely don’t think we had a place for him in our offense. He just doesn’t fit here. That being said, I am still upset that Beane traded that pick to KC. He fits their offense perfectly and he is going to be a pain in the butt to defend for years to come. I have seen two primary rebuttals: 1) “KC would have picked him anyway”. My response; At their original pick? Are you suggesting that KC gave us a free draft pick out of the kindness of their heart when they could have stayed put and taken him away? 2) “They would have just traded with another team to move up and get him”. My response: Common sense dictates that they took our offer because it was the best available offer. In other words, if they had another offer to move up, it likely would have cost them more. I didn’t like the trade then and I don’t like it now. You just don’t do business with your enemies.

Stroud is … yikes

Worthy is tough to defend when he gets a head start

I would accept that challenge. They were scoring regardless - literally every drive except for one. I would much rather take my chances blitzing. Yes - they may score more quickly. But they may also force a turnover. Regardless, either of those are better than dying a slow death. The absolute worst thing we could do is allow them 10 play drives that take 5+ minutes off the clock. Our entire defensive philosophy can be summarized as “hope the opposing offense makes a mistake”. Well, they weren’t making mistakes that week. So you have to do *something* to try and force them into a mistake. I would have blitzed them relentlessly.

Agree. Allen is superior to Kelly in literally every way.

Ill check the API for errors but as it stands we were only counted for 2 blitzes on 3rd/4th down in the second half. EDIT; I just realized that of course it will show 2 because the penalty makes the 3rd irrelevant. But either way, 3 out of 10 is certainly not blitzing almost every 3rd down. I also noticed that both the catches are just incredible plays. The blitz forced the Rams to make circus catches rather than routine catches. I didn’t pull 2nd down stats because we were all talking about 3rd down defense.

Yeah go back and watch it, I think you’ll be surprised. We barely blitzed at all on 3rd/4th down. They were eating us alive because Stafford had time. The rare time we blitzed, it definitely sped the Rams up but they made some amazing catches. Im going to release a free Bills stat website next week. I’m finishing up the coding but you’ll be able to drill down these stats.