BuzzerBeater Forums

Also the problem are not JR or JS, but the way open shots are simulated. Open shots should have a higher chance of going in, but it appears that team defense is what determines the success of an open shot.

Yes, yes, a million times yes. Did I mention yes?

I still remember a SG I had (Jack Poirier) who consistently over a two plus season span had a higher FG% on contested jump shots vs. "open" ones. I don't believe I ever played him out of position (so always PG, SG, or maybe SF) so it's not a case of lighting up the three skill donkey big men, either. And he was my best pure jump shooter pretty much all of that time, so it wasn't like it was just a guy who wouldn't shoot unless given a tremendous matchup.

Did you see the other post?

S39 unguarded turnaround jumpers 0/77, fade aways 0/100 and hooks 0/85: all 0% success rate.
Guarded turnaround 78/238, fade aways 306/1092 and hooks 80/271.

S38 (on incomplete data) unguarded turnaround 1/62, fade aways 1/55 and hooks 0/56
Guarded turnaround 60/194, fade aways 295/881 and hooks 67/196

So it appears that if you want to score on those shots it MUST be a guarded shot. Interesting and completely bollocks.


The FG% for Jump Shots is 37% unguarded and 30.5% guarded. The number for 3 pointers is 28.6% and 23.6% respectively. But of course to add insult to injury Jumpers and 3 pointers are also the most guarded shots (with only 34% and 24% open shots respectively).

Last edited by Lemonshine at 4/7/2018 5:41:44 AM

Did you see the other post?

S39 unguarded turnaround jumpers 0/77, fade aways 0/100 and hooks 0/85: all 0% success rate.
Guarded turnaround 78/238, fade aways 306/1092 and hooks 80/271.

S38 (on incomplete data) unguarded turnaround 1/62, fade aways 1/55 and hooks 0/56
Guarded turnaround 60/194, fade aways 295/881 and hooks 67/196

So it appears that if you want to score on those shots it MUST be a guarded shot. Interesting and completely bollocks.


The FG% for Jump Shots is 37% unguarded and 30.5% guarded. The number for 3 pointers is 28.6% and 23.6% respectively. But of course to add insult to injury Jumpers and 3 pointers are also the most guarded shots (with only 34% and 24% open shots respectively).

Yeah, I saw it. It's good stuff (well, good in the sense that I appreciate the effort and considered it very useful to know, not that I think the stats are indicative of a positive situation in terms of the game's balance). I kind of wish I had the patience to catalog the wild disparities in shooting percentages that occur with Princeton - I suspect if I went back through, I could probably identify at least one third of the games where I had a six minute stretch with under 5 points and a six minute stretch with at least 18 points. Maybe even more than that.

A French user you know very well bought supporter for me without saying anything, so I might try to complete my code before it expires and we can have a look at some stuff. He's also interested in the same table I posted but for specific tactics.

My file currently has 2 problems:
1) I can easily extract the games list and all the other info from the BBApi and I extract all the boxscores like that to double check that my number from the pbp are correct. However, downloading the pbp for, say, 176 games (1 season for 1 league, but could be easily more) is another thing altogether, because the website logs me out at regular intervals and I'm not skilled enough in dealing with internet explorer through VBA to understand how to fix it, so I worked around it but I need to clean it up.
2) All the output part is not done. I was going to make tables like advanced stats, regular stats, stats /36, defensive stats and shooting stats by league/team/player. I couldn't be bothered to spend the time to build that up, but maybe I will do the advanced stats, shooting and defense now. I'm open to suggestions about what to implement.

So that we're clear what I have after I import the data is something like this for every event in the pbp and every game (s39 in the EBBL had 87,152 lines):
http://tinypic.com/r/2pq7rth/9
plus the lineups
http://tinypic.com/r/n689ra/9
plus all the information on the games from the BBApi (which include tactics and GDP)

Note that I also have information on the lineups on the floor for any given event in the game and the time on the shot clock. Both are useful to check different things. The lineups will be useful if we decide for example to look for some evidence for or against the theory that C shoot better from 3 when they are guarded.

I will do the Princeton analysis on 6 seasons: 2 of yours, 2 of Nachtmahr and 2 of the other guy who volunteered in the thread and we can compare league-wide (or league ex Princeton playing teams) stats and distributions and Princeton only stats and distributions.

Last edited by Lemonshine at 4/9/2018 8:09:45 PM

Did you see the other post?

S39 unguarded turnaround jumpers 0/77, fade aways 0/100 and hooks 0/85: all 0% success rate.
Guarded turnaround 78/238, fade aways 306/1092 and hooks 80/271.

S38 (on incomplete data) unguarded turnaround 1/62, fade aways 1/55 and hooks 0/56
Guarded turnaround 60/194, fade aways 295/881 and hooks 67/196

So it appears that if you want to score on those shots it MUST be a guarded shot. Interesting and completely bollocks.


The FG% for Jump Shots is 37% unguarded and 30.5% guarded. The number for 3 pointers is 28.6% and 23.6% respectively. But of course to add insult to injury Jumpers and 3 pointers are also the most guarded shots (with only 34% and 24% open shots respectively).

Was there any response from BB about the issue? It might just be representational bug or something more serious...

No reply and I'm 99% sure this is not a problem of the viewer.

Well all shots are tested against defense as they should or you get in trouble with way too low percentages in low divisions or way too high in upper divisions.

There are two possibilities. First all shots are calculated equally and only best quality shots are reported as open. Second GE determines if a shot is open or not and then does the calculations. If the first is true then it's most likelly just a representational bug within GE not the wiever while the calculations are still correct. In the second option it does sound like quite a huge bug sadlly.

I've gotten the impresion that GE is more or less a boogie man to the BBs and there's very little chance of fixing such bugs???

Last edited by Lonietas at 10/19/2019 4:52:36 AM

I'm not an IT guy: I can code in VB and Python, but that's about it. They should have a test environment to try (and FIX) stuff and my understanding is they never had one. The GE is apparently tens of thousands on lines of code, it takes time to change anything to the point that it may be better to rewrite the GE altogether and of course not having the test environment means you can't troubleshoot. Bugs like the ones mentioned above are not necessarily easy to track down, as they may be due to a variety of reasons.