BuzzerBeater Forums

A won't get you assists and it's debatable to what extent it will reduce the turnovers (which is why I replied to hrudey). It is very pertinent to this thread that HA may not be important at all while PA certainly is. There are plenty of players with very high HA and no PA, perhaps you could study those and tell us how they perform in both Ast and TOs. I will help you out, check the following who are the reverse of my player:

I think the problem I have with contributing more is that for pretty much the entire time I've played this game until this recent experimentation, I've been obsessive about making sure my guys don't have low PA or HA, and I'd always been among top two or three in fewest turnovers. Most of the low HA/PA evidence I have is anecdotal from Klein, who was 1/1. Hitchcock started 7 HA/1PA, and he's still only 10 HA/3 PA, which could lend more support to the "bad passing is bad" aspect but it a poor testbed for the other side.

The reason I mention the Princeton specifically is that the flow skills of the C in particular do tend to matter more than in most offenses. Passing on big men is extremely useful at all times, of course - the C assist to the PF for a dunk is likely one of the easiest baskets to convert. But Princeton seems to use that even more on passes back to the exterior as well as passing to people for driving attempts, which unfortunately seems to be way too frequent with a solid passing C and a guard/SF who has high driving and low IS.

My other bigs have 11, 12 and 13 PA and when I played Princeton in Utopia against weaker competition I was playing Cornaglia at PF who had 20 HA and 11 PA if I remember right. Needless to say Vadikolias is the only player in my entire roster with such low PA and HA, which I suspected would make it more TO prone. It very unfortunate I didn't take note of another big man who sold a few weeks ago who was similar to Vadikolias (he had 1 HA / DR, and 6 OD / 7 PA were swapped compared to him) but I can guarantee he had no TO issue either.

My point is that after seeing how Vadikolias has done in the last 5 or 6 seasons I'm very skeptical about HA. I'm sure it reduces the turnovers, but the real question is by how much. I know that PA reduces turnovers because the player can find a way to get rid of the ball when he's in a pickle. Turnovers are not only bad passes and it stands to reason that different TOs are affected by HA in a different way. TOs as far as I know can be:
STEAL_ON_PASS
STEAL
BAD_PASS
TRAVEL
OFF_FOUL
THREE_SECOND
SHOT_CLOCK_ZERO
and a Bad Pass includes different situations (such as receiver not ready) although I suppose that's purely cosmetic.

That said, the elephant in the room is DR, obviously. HA always come with DR and it appears that more DR = more touches/shots. That seems to be the case in most tactics at any level. I have no doubt that lower DR reduces the number of TO opportunities and changes the kind of shot a player will take to some extent.

For season 38 and 39 I have full data on every single event which happened in the EBBL. I can't collect more data until I get supporter again as I need the play-by-play. In any case, in S39 my 2 starting bigs fared like this:
- 20% of Vadikolias (1 DR and 17 IS) shots were Dunks and another 21% were Inside shots. He had then 8% Putbacks and virtually no Driving Layups (he literally had 1). He had 170 attempts in 421 minutes, 14 of which resulted in a foul.
- 23.5% of Ma Anhui's (12 DR 13 PA and 12 IS) shots were Driving Layups, 13% were Dunks, 10% were inside shots 16% were Putbacks. He had 205 attempts in 630 minutes, 22 of which resulted in a foul.

You see a clear difference in shot selection there and these guys often played together or at least against similar competition. My guess is that tactics influence the general type of shot, but the rest is down to individual skills. So a close/midrange shot can become a Driving Layup instead for players with DR.

Note that the league-wide:
- Dunks had a FG% of 83% and 13% Foul Rate;
- Driving layups had a FG% of 42% and 10.8% Foul Rate;
- Inside shots had a FG% of 39% and 11.4% Foul Rate
- Hooks had a FG% of 25% and a 9% Foul Rate

Unfortunately there were only 7 games where Princeton was played that season so I would need a larger database with more seasons or more leagues to extrapolate meaningful data for Princeton only. Maybe if I get supporter again I will expand the DB and finish the output stats properly.

Last edited by Lemonshine at 04/05/2018 12:36:20

Offtopic, but I have to expand your numbers, cause they are only half-finished
In 2012 I made an analysis of all types of shots in BB. It's similar to your last paragraph.

- Dunks had a FG% of 83% and 13% Foul Rate

I had: 75% FG ; 70-75% assisted ; 30% of all actions were defended ; 12% foul drawn ; 20-25% foul during defended shot

- Driving layups had a FG% of 42% and 10.8% Foul Rate

35% FG ; 40-45% assisted ; 55% of all actions were defended ; 11-12% foul drawn ; 23% foul during defended shot

Inside shots had a FG% of 39% and 11.4% Foul Rate

38-40% FG ; 45-50% assisted ; 50% of all actions were defended ; 10-13% foul drawn; 20-25% foul during defended shot

Hooks had a FG% of 25% and a 9% Foul Rate

25% FG ; 10-15% assisted ; 85% of all actions were defended ; 7-10% foul drawn ; 5-10% foul during defended shot

And I'm back into hibernation ;)

These are the actual numbers I get. I have put them in a picture because it's impossible to post a coherent table on these forums: http://tinypic.com/r/veauxx/9

I check the Princeton games when I have the chance. These are the subtotals in (hopefully) a readable format:

Type: Inside shot; FGA: 6,904; Pct of total: 21.2%; TS%: 59.6%; Assisted: 3,978; AssistedPct: 57.6%Fouled: 823; FouledPct: 11.9%Open: 4,084; OpenPct: 59.2%Defended: 1,997; DefendedPct: 28.9%

Type: Layup; FGA: 6,026; Pct of total: 18.5%; TS%: 42.0%; Assisted: 2,582; AssistedPct: 42.8%Fouled: 649; FouledPct: 10.8%Open: 2,727; OpenPct: 45.3%Defended: 2,650; DefendedPct: 44.0%

Type: Putback; FGA: 1,781; Pct of total: 5.5%; TS%: 48.4%; Assisted: ; AssistedPct: 0.0%Fouled: 183; FouledPct: 10.3%Open: ; OpenPct: 0.0%Defended: 1,598; DefendedPct: 89.7%

Type: Jump Shot; FGA: 11,850; Pct of total: 36.4%; TS%: 32.9%; Assisted: 3,824; AssistedPct: 32.3%Fouled: 951; FouledPct: 8.0%Open: 4,082; OpenPct: 34.4%Defended: 6,817; DefendedPct: 57.5%

Type: 3 pointer; FGA: 6,021; Pct of total: 18.5%; TS%: 37.5%; Assisted: 1,343; AssistedPct: 22.3%Fouled: 90; FouledPct: 1.5%Open: 1,447; OpenPct: 24.0%Defended: 4,484; DefendedPct: 74.5%

Type: Total; FGA: 32,582; Pct of total: 0.0%; TS%: 41.7%; Assisted: 11,727; AssistedPct: 36.0%Fouled: 2,696; FouledPct: 8.3%Open: 12,340; OpenPct: 37.9%Defended: 17,546; DefendedPct: 53.9%


Edit: I included also the TS% for Assisted, Open and Guarded shots in the Picture. Still planning to do the one for Princeton but this has already taken enough time for now.


Last edited by Lemonshine at 04/05/2018 20:40:47

Very nice stats

I'm not sure why but I got supporter now (I asked Marin how that happened but he hasn't replied yet).

I may launch the code to download all the s40 and s41 games (and complete the s38 for the games that got broken because of slow connection) in the EBBL.



If you know teams other than hrudey who play Princeton please post them here. If we get 2 or 3 teams more, I will get the stats from the leagues where those teams play (let's say 6-8 leagues in total, that's like 1056-1408 games) and then do a table for the whole dataset and a table for only the teams and games where Princeton was played.

We need teams who play Princeton at least 8-10 times a season and in leagues without bots (better if at least D2).

Last edited by Lemonshine at 04/07/2018 21:02:28

Hi mate, you can look at my box scores for your stats. I played Princeton 9 times this season and I played it in the game that offers me the path to the title as well. Hope it will helps you :)

Will do.

Season 19-22 was my Princeton time. I played it for most of the time during this 2.Div turn.

All right I have made some progress with my VBA code and I have done the analysis for the BBBL season 41 (Tom 10 above)

I used all RS, playoff and relegation matches. In 193 total matches 33 teams used Princeton (so 33/296).

The results I get are as follows:

League as a whole
http://tinypic.com/r/2rcsl1f/9

Princeton teams
http://tinypic.com/r/2hq83g0/9

Note that in these 2 tables I revised the TS% formula to account for FT and Foul rate. Now instead of calculating PTS/(2*FGA) which is more of a eFG% calculation, I changed it to (PTS+SH_FOULS*1.9*FT%)/(2*FGA+SH_FOULS*1.9). The 1.9 (and 2.85 for shooting souls on 3 pointers) parameter is there to account for the fact that a small portion of fouls result in 1 FT instead of 2 (or 3). I factored this in also in the TS% for defended shot as we can safely assume that all shooting fouls are the result of a defensive play.

We can already see the difference here:
- Jumpers+3 pointers instead of the league average 56.5% Princeton teams had 71.2%; shot efficiency in these shot categories is bumped by 6-7% (point per shot)
- dunks+inside shots+put backs instead of the league average 26% Princeton teams had 15%; shot efficiency dropped by 7% for dunks (only on guarded ones), 15% in inside shots and 17% on putbacks
- layups dropped a little in number but not in efficiency and they are still more efficient shots than jump shots and 3 pointers.

Conclusions:
- Inside shots need to be limited as much as possible and getting people with DR may help or not depending of whether the result is an inside shot or a layup. We'd have to go further into this to understand what causes the drop in efficiency (i.e. it may be simply a drop across the board, or also the fact that more shots are taken by the worst inside shooters)
- If you don't consider FTs, then overall the TS% (41.7% for both) is very similar between the whole league and Princeton teams. However the foul rate in Princeton is lower than the average tactic used in the league and FT are the most efficient shot in the game after dunks, so Princeton end up being slightly worse at 46.4% v 45.9% (no doubt some tactic, LI?, will have more than the 46.4% average)
- There should be no reason for jump shots to have such a lower foul rate than inside shots (it's more logical for 3 pointers)

I will do an analysis of B3 KO games before returning to this. I suspect hrudey's numbers would show something similar. Nachmahr's games are too old, before changes to 3 point shooting and SB, so they will probably show a worse situation than the current one.



Last edited by Lemonshine at 04/14/2018 21:35:08

Oh yes: compared to the previous tables here I also included the fouls into the defended shots so TOTAL=OPEN+GUARDED, while previously TOTAL=FOULS+OPEN+GUARDED. In the future I may put also the points scored as well so that people can do their calculations if they so wish.


Another thing the numbers seem to indicate is that Assists help more for inside shots than outside shots. For 3 pointers the difference in efficiency between an assisted shot and a defended shot is relatively low (assisted shots are always open shots).

Again we need to check these numbers by position as well, because we may find that the 3 point percentages are beefed up by C and PFs shooting surprisingly well when guarded...which is a theory that has gone around for a long time and may explain how the open v guarded drop is not so large (the theory goes that for big man is better to take a contested 3 pointer than an open one as the FG% is higher when guarded)


Last edited by Lemonshine at 04/14/2018 21:29:05