Statistics we are interested in.

[docpkrdude]

Statistics of interest can be categorized into various "levels" to represent their urgency (some combination of utility and difficulty of coding). The deployment of the various levels might look something like the following.

Level 1: These are the most basic event statistics which might be meaningful to individual players. They will be functionally represented into a mockup with which we can onboard the soccer coach in early December. If we aren't confident we can deliver something with Level 2, we won't include them in the mockup.

Level 2: This is a slightly expanded set of statistics that can be robustly recorded by young users with limited training, limited technology experience, and moderate game-time attentiveness. Our aim is to represent most of the statics that are presently represented on the coach's paper statistical sheets. We'll develop and test this implementation using recorded games available online.

Level 3: Statistics in this level are those which are not on present statistical sheets, which we can visualize a reasonable pathway, and would have clear value to typical coaches.

Level 4: Statics relating to the supporting players and players not immediately playing the ball.

Event to sort (basic statistics....advanced statistics): goal scored (for/against/own-goal) shot on goal (player/place/time/ball-hit-zone...L/R leg, goalie avoidance) goalie stop (position on field, height of the ball) goal kick (result) goal punt (result) gained possession (player/place/time...forced-bool) loss of possession (player/place/time....forced-bool) header (player, force, possession retention) foul committed (player/place/time) foul win (player/place/time) corner (player/ball-hit-place/result) pass (player/time/direction) pass receipt (player/time/?) passing chain (number of passes before turnover) pass grade (player/time/direction/alternatives/pressure/risk-assessment/L-R-leg choice) dribble past (player/time/place/follow-on-event) dribbled past (player/time/place) cleared ball (player/time/place/follow-on-event) impactful pressure (player/time/place) heat map for player heat map for player overlaid with ball location heat map for player color coded for speed

Statsbomb tracks some of this and more for pro games...some of their data is publicly available. https://soccermatics.readthedocs.io/en/latest/gallery/plot_UsingStatsbomb.html

[indohippi]

The stats bomb thing is cool, but its 100% python and you have to have credentials to use their APIs. Maybe for the backend it will be helpful though.