Warrior General Games

Originally posted by a_esbech
You mean that there's sense to be made from the frames array? It seems to me it would be quite a heavy script to keep track of every player's movement. I made a script last night that extracted the json data from all plays in a game. However it would be a lot more useful if the data contained player stats and yards gained by the play. That way you'd be able to cross reference player stats with plays. Is your WR more useful on long or short throws. How many yds does your HB gain when running through the middle as opposed to running on the outside.

There's a lot of potential for a well structured API.

No doubt statistical information 'can' be provided in the json data at a pbp level in a structure at game level, but what level of effort would it be for the devs to include it that way. I would be willing to bet how it works is as such. Each play is ran independently and when it comes to end of game totals, it parses each play for statistical info. Hence probably why TFL stat was not recording correctly a while back. Personally I look at as this, does the value dictate the effort? For me it doesn't as I know the sim changes, teams change and builds are constantly evolving and reacting to the sim. That only leads me to provide statistical information to take with a grain of salt. Again personally is how I see it, not all feel this way.

Originally posted by Stobie
No doubt statistical information 'can' be provided in the json data at a pbp level in a structure at game level, but what level of effort would it be for the devs to include it that way. I would be willing to bet how it works is as such. Each play is ran independently and when it comes to end of game totals, it parses each play for statistical info. Hence probably why TFL stat was not recording correctly a while back. Personally I look at as this, does the value dictate the effort? For me it doesn't as I know the sim changes, teams change and builds are constantly evolving and reacting to the sim. That only leads me to provide statistical information to take with a grain of salt. Again personally is how I see it, not all feel this way.

I agree with the sentiment "No challenge, no reward". However I believe that once the data has been gathered, that is when the challenge really begins. Making good statistics is all about finding out what is usable data and what is clutter.

I've just analysed two plays and tracked two players in each. The first one was a handoff, if I were analysing the data I got from the frames array all I could tell you was that the QB held the ball right up to the frame where the HB held the ball. The only indication the HB eventually was tackled was the fact that towards the end of the play, the HB had a move called fall..tackle. Yet the player who tackled him wasn't in the same XY coordinates and there's no sign at all that the defender made the tackle. (HB was at y 65.25 x 173.25 and FS was at y 68.5 x 176.75) - this play: http://glb2.warriorgeneral.com/game/replay/40367/160153/1

The other play was a pancake play. Again there are no evidence that the two players are actually trench fighting. Both the ROT and LDE are entangled in a block until eventually the LDE is prone and the ROT moves. Once again the positions are hard to compare. ROT is at y 79 x 153.5 and LDE is at 75 x 154.25. There is no mention of the fact that the reason the LDE is prone is due to a pancake from the ROT. - this play: http://glb2.warriorgeneral.com/game/replay/40367/160153/1

So I maintain the claim that the data in the frames array are useless.

Originally posted by a_esbech
I agree with the sentiment "No challenge, no reward". However I believe that once the data has been gathered, that is when the challenge really begins. Making good statistics is all about finding out what is usable data and what is clutter.

I've just analysed two plays and tracked two players in each. The first one was a handoff, if I were analysing the data I got from the frames array all I could tell you was that the QB held the ball right up to the frame where the HB held the ball. The only indication the HB eventually was tackled was the fact that towards the end of the play, the HB had a move called fall..tackle. Yet the player who tackled him wasn't in the same XY coordinates and there's no sign at all that the defender made the tackle. (HB was at y 65.25 x 173.25 and FS was at y 68.5 x 176.75) - this play: http://glb2.warriorgeneral.com/game/replay/40367/160153/1

The other play was a pancake play. Again there are no evidence that the two players are actually trench fighting. Both the ROT and LDE are entangled in a block until eventually the LDE is prone and the ROT moves. Once again the positions are hard to compare. ROT is at y 79 x 153.5 and LDE is at 75 x 154.25. There is no mention of the fact that the reason the LDE is prone is due to a pancake from the ROT. - this play: http://glb2.warriorgeneral.com/game/replay/40367/160153/1

So I maintain the claim that the data in the frames array are useless.

My apologies on my assumptions. There has to be something else or they are making some grand assumptions as well.

Originally posted by Stobie
My apologies on my assumptions. There has to be something else or they are making some grand assumptions as well.

No worries, if I had concluded that the frames were useful, then I would've made more progress than I had imagined I would. So it was a good thing testing both of our claims.

Hopefully the admins will consider adding a bit more information on each play.