Anyone on here know a good data source I could use for analysing the effectiveness of different ways of defending set pieces? Don’t mind paying for it if not open source.
It was interesting hearing Oggy on the way home after Blues, talking about 2nd phase from long throw ins.
For those who missed it, the point he was making was that the pace and trajectory is much different to a corner pinged in, that enables the defender to get a good strong clearance on, where with a thrown in ball, you need to generate the pace, which is of course much more difficult. It's basically how BTA's goal came about.
Not something I'd heard before, but makes sense and why so many teams are now using that weapon, as it seemed to me to be overused, with not enough direct goals resulting from them.
You’d first need footage of set pieces, but without full matches you’re probably not getting the well defended ones. So you’d need a large amount of full games and outside of pro services like wyscout or your own fire stick and massive dvr I’m not sure you could get it. Once you’ve got that there’s annotation software available for free or you could chuck it into an LLM to extract just the set pieces and use standard extraction algos to count numbers of players on each team maybe. A lot of work though if no one is paying you.
My point is clearly set piece coaches have access to this sort of thing which explains why so many teams keep going for something that's leading to really poor goals being conceded on the regular.
Don't mind paying for the service either, within reason
You’d assume set piece coaches either do this or have access to someone who has. I agree it’s mental to me not to leave anyone up. Especially when your keeper has no out ball when he collects it.
Anyone on here know a good data source I could use for analysing the effectiveness of different ways of defending set pieces? Don’t mind paying for it if not open source.
My point is clearly set piece coaches have access to this sort of thing which explains why so many teams keep going for something that's leading to really poor goals being conceded on the regular.
Don't mind paying for the service either, within reason
I recently felt the urge to get a subscription to Statsbomb (one of the large data recruitment models used) until I discovered it was something like £12k per month.