{"metadata":{"kernelspec":{"name":"ir","display_name":"R","language":"R"},"language_info":{"name":"R","codemirror_mode":"r","pygments_lexer":"r","mimetype":"text/x-r-source","file_extension":".r","version":"4.0.5"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"**Gunner and Jammer Win Rates**\n![](https://operations.nfl.com/media/4012/gunner-slater.jpg?anchor=center&mode=crop&width=1417&height=740&rnd=132168383060000000)","metadata":{}},{"cell_type":"markdown","source":"Code: https://github.com/tuckerboynton22/bigdatabowl22.git","metadata":{}},{"cell_type":"markdown","source":"The unsung heros of any punt play are the gunners and the jammers, the former in charge of chasing down the punt returner and the latter tasked with impeding the gunner's pursuit. As soon as the ball is snapped, the gunner fires off the line of scrimmage, sprinting as fast as he can towards the punt returner at the other end of the field. Meanwhile, the jammer, also known as a vise, does runs along with the gunner, doing everything in his power to get in the way and slow down the gunner. Typically, we don't see the jockeying between the gunners and jammers on television, but we do often observe the result of this hand-to-hand combat: if the gunner prevails, he often arrives at the returner at or before the punt lands, forcing the returner to call for a fair catch, let the punt bounce, or be tackled for little to no gain; if the jammer effectively prevents the gunner's forward progress, the returner has time and space to field the punt and return it for a sizeable gain.","metadata":{}},{"cell_type":"markdown","source":"Despite observing the final outcome of the punt play, we have no real metrics to quantify who wins the battle between the gunner and the jammer. In the last several years, ESPN has popularized the metrics Run-Block Win Rate and Pass-Block Win Rate (as well as their defensive counterparts Run-Stop Win Rate and Pass-Rush Win Rate), which measure how often offensive (defensive) linemen maintain (beat) their blocks in the 2.5 seconds following the snap. I propose two new statistics in the same vein: Gunner and Jammer Win Rate. Because the average punt play is longer than a typical pass play (~4.2 seconds when there is a return, fair catch, or downed punt), I extend the period of observation from 2.5 to 3.5 seconds and award the gunner a \"win\" if he has passed his opposing jammer(s) on the horizontal axis at the 3.5-second mark and the jammer a \"win\" if the gunner fails to do so. Here's an example of how one gunner wins and the other loses:","metadata":{}},{"cell_type":"markdown","source":"![](https://user-images.githubusercontent.com/70855507/148473962-a981c7dd-a7e9-4bbe-96a2-5934f25cd28e.gif)","metadata":{}},{"cell_type":"markdown","source":"In order to calculate win rates, I isolate my analysis to plays during which there was a return, fair catch, touchback, muff, or punt downed/out of bounds that happened after at least 3.5 seconds had elapsed. I do this because gunner/jammer responsibilities may change after a punt is fielded/muffed/etc., and my goal is to examine exclusively the period during which the gunner wants to run forward and the jammer wants to stop him. I classify gunners and jammers based on PFF scouting data, and for each frame of the play after the snap, I isolate the X-coordinate of the gunners and jammers, introduce indicators for the gunner being farther downfield than the jammer, and record whether the gunner has \"won\" 3.5 seconds after the snap. Finally, I isolate plays that had no penalties and at most one gunner lining up on the each side of the field, as plays with multiple gunners on the same side create ambiguous blocking responsibilities, and modeling those assignments is beyond the scope of this metric.","metadata":{}},{"cell_type":"markdown","source":"I now have leaderboards for Gunner and Jammer Win Rates in both single- and double-coverage situations:","metadata":{}},{"cell_type":"markdown","source":"![](https://user-images.githubusercontent.com/70855507/148472397-6789b5f3-b511-415f-b4f8-a397d6390fb8.png)\n![](https://user-images.githubusercontent.com/70855507/148472400-2cca467b-78a7-43f1-80bc-6c40a4732e75.png)\n![](https://user-images.githubusercontent.com/70855507/148472401-a3308af7-6601-425a-a364-10b93a47d507.png)\n![](https://user-images.githubusercontent.com/70855507/148472402-0a00534f-d845-4334-91f4-47a1795dee3b.png)","metadata":{}},{"cell_type":"markdown","source":"Additionally, I plot each player's win rate versus their double-team rate as ESPN's Seth Walder often does with Rushing/Blocking/Stopping Win Rates.","metadata":{}},{"cell_type":"markdown","source":"![](https://user-images.githubusercontent.com/70855507/148472405-2a683d07-c8ca-437a-b418-c12313a4edae.png)\n![](https://user-images.githubusercontent.com/70855507/148472406-43ad8c82-13dc-43f1-a4b9-057d95f80f4b.png)","metadata":{}},{"cell_type":"markdown","source":"Finally, I create the table below, which shows that at a high level, \"winning\" leads to better outcomes at the play level––when more gunners win, both mean and standard deviation return yards fall.","metadata":{}},{"cell_type":"markdown","source":"![](https://user-images.githubusercontent.com/70855507/148472693-2e14aa3d-ae59-499b-9124-aca3ca0b1141.png)","metadata":{}},{"cell_type":"markdown","source":"","metadata":{}}]}