BuzzerBeater Forums

OK so this is a continuation thread from something that someone suggested in the Suggestion forums, which ended up becoming a discussion about How much of the hidden aggression stat contributes to a players fouling frequency, and whether or not you could 'train' this out of him. ie; Bulk up his OD, Experience, etc etc.

Ive done some quick analysis of this and am just presenting the stats
AIM - To try and find if there is an obvious correlation between player stats and fouling frequency and thereby prove how much of an influence the hidden aggression stat makes on the player.

Assumptions - Im excluding game tactics, game shape, enthusiam etc etc from these stats. Im simply looking at the player, his stats, and his fouls per 48 mins.
For my initial sample group, I took players aged 28 - 30 (thus also making the assumption that none of these guys popped in OD, Handling, or Experience although, they may have) and the stats for the previous season. My initial sample group was 30 Pg's and SG's.
My first sample group was also players with a maximum of 5-Mediocre Outside Defence. as well as a max of 8-strong Handling.
So its just a simple ratio between those 3 stats, and the fouls per 48 mins.

Ironically, I found absolutely no relation, not even when i applied weighted values to OD, or Exp or anything. If anything, it was a negative relationship. (which makes me think i need a bigger sample, and not to limit the stats ranges)

More ironically, was I only had one player whom, with OD of 5, was averaging more than 5 fouls for 48 mins, which was surprising.
More work to be done yet, and I think once I incorporate players with broader ranges of stats, the relationship might improve, so watch this space.

Nice work and nice team name.

I think the biggest problem there is that you exclude opponents from the analysis. If your opponent is good at driving, for example, I guess your player will foul more ceteris paribus.

Also, there's a question of how many of these fouls are offensive versus defensive.

However, it does seem to contradict the oft-repeated advice to train OD to reduce an outside player's tendency to foul (assuming the sample size is sufficient, which is may not be - especially if you're not careful to select players from equivalent strength leagues).

Yes these are two of the many other factors. Good to points to raise though and thanks for the input.

Other factors:
- Opposition - quality of the opposition, the defensive/offensive strategy, etc etc.
- Offensive fouls - does this mean I take into account Driving into my formulae?
- Team mates - If that player is the worst defensive player on the team, he will be targeted by the game engine as the weakest link and hence, be more prone to be defending the opposition and hence, more fouls.
- End of game - some players may have more of a tendency to be playing at the end of the game, when, if its close, lots of fouls occur. Very difficult to take these into account.

At the moment, ive sampled only 30 players off the transfer market. Im hoping to double and perhaps triple that over the next few days, and to include better defenders into the equation.

But the general idea is that, a large sample size, and for each player, their stats for all of last season (22 games or thereabouts). If the sample size is large enough, it should be able to provide generalisations about the data, and smooth any of the 'exception' scenario's as mentioned above, from the equation. ie: Gives a rough idea!
Once I get all the data and remove any anomolies, I can see what the data says.
The end result may very well be that all I do is prove my method to be way too flawed, to be of any use. But then ill be able to sleep at night :)

Any more comments let me know peoples! I might be missing something really obvious here.

Any more comments let me know peoples! I might be missing something really obvious here.

I may know what you have been missing: Inside Defense on Guards. Low ID makes guards make a lot of fouls on driving opponents. I had a SG with OD 12 and ID 2. OD was quite good for the league I was in, but the player kept fouling out a lot. Then I started putting him on players who didn't drive that much and my SG stopped fouling out.

I noticed the same on two more SGs I bought - both has good OD and bad ID. And they were fouling out more than normally.

In addition, I have a SG with OD 12, ID 7 and SB 7. I also have a PG with OD 13 and ID 7. They are in my team for four seasons now. SG fouled out two times: once when he guarded a PF and once when he guarded a go-to player who had 42 shots (his owner played Patient).
The PG never fouled out in his career. SG started as 19y old and PG as 18y old.

Yes good point, id actually thought of this for the Forwards and Outside D, but totally forgot about the guards.

Its logical. Because if you take offence, a guard, who has excellent driving, but poor Inside Shot, tends to miss his drives more often than a guard who has good or even average inside shot.

Im already half way through some of the data, so i might have to take another sample pool to take into account inside D.

Alright so ive decided to present some findings now that ive taken a larger sample pool.

Still focusing on guards only, and ive chosen to ignore Inside D for this sample. The stats used OD, Handling, and Experience, and fouls per 48 mins, over a variety of levels, and stat combo's. (high OD, low handling, high handling, low OD etcetc).

Sample size was just ove 50 players. and stats were their averages for season 16. All SG/PG aged 28-30 to ensure stat stability over the season.

As a simple test, I tried to establish correlation between each of the individual stats, and Fouls per 48 mins (from now on referred to as FP48).
These are the results
OD -0.3
H -0.13
Exp - 0.41

Yes those are negative correlations. So nothing obvious unfortunately, which suggests other factors, and weighting and etc etc.
So looking at it more closely, I gave a weighting to each of the above stats.
Using a weight of 1 for each stat, and then correlating that as a total, between FP48
Correlation = -0.35 Still no luck.

So I played around with some weightings, I wont go through each scenario, but what i did find was that when i increased the weighting for Experience, it just made the correlation worse, and when i remove experience from the equation all together, it made the correlation significantly better.
Thus ive deduced that Experience has no input wih regards to a players fouling frequency. In fact, those numbers above suggest that handling has more of an influence than OD, but with negative correlation values, this is just a statistical fallacy.

I decided to introduce a randomized 'Aggression' number, for each player, value 1-5. totally random, but equal distribution.

Random Aggression value correlated with FP48 +0.16
well at least i got a positive correlation!

I gave my aggression statistic more of a bell-curve distribution (that is, more players have an 'average' aggression level, and less players have the extreme's) - I also increased this value distibution to match BB stat ranges 0 -20.
Doing this, i increased the correlation to about 0.25,

Using these correlation values, I can deduce that, statistically, players stat levels are less likely to contribute to a player fouling out, or making fouls. The hidden aggression stat does seem to provide more of an impact towards fouling, compared to other stats. But at a correlation value of 0.16, its impact is hard to determine.

Therefore, I would say that foul ratio is very much related to style of defense, the opposing player, whether or not the player is the weak 'defensive link' in your team, what position in D he is playing, is it a home game? opposition tactics, game shape, comparative stamina etc etc etc. All these other factors must obviously contribute in a generous factor.

So in conculsion, can you train-out a players agressive nature and frequency to foul?
Probably not.
BUT - Could you train him such that he is no longer the weakest link in your team, thus is less susceptable to fouling?
Yes. OD/ID. Other factors? Very hard to not be the home court isnt it!

hope some people found this useful.

Interesting study. To be honest, foul troubles in BB always gets on my nerves.

It´s pretty hard to establish any kind of correlation with foul tendencies.

The only conclusion I could arrive on 16 season playing BB, is that the hidden foul attribute is as powerful as it gets.

It really ignores the OD and ID of your players, your team defense, experience, you name it.

One could think that -yes- your rival´s offense has something to do with it, but I got proved wrong again and again, being my last frustration last wednesday on the B3 Cup: (38851282).

23 fouls to 7 (you can take off some late fouls on the last 20 secs), still I´m probably tripling my rival on fouls.
Definetively not a record, but it reminds you again that even if it looks that your teams can play Pistons-calibre D, nothing tops the hidden "Charles Oakley" foul attribute. When you like to chop some rival´s hands...just sit an prepair a comfy place at the bench for your starters.

And no, there´s nothing you can do about it (besides shipping your guys to neverland), a whopping 9 sub-levels of difference between my OD and his scoring, and 4 sub-leveles of difference between my ID and his inside scoring. With 4 rival players below 75.0 on MR and the fifth guy on 79.0, you could wonder how my rival could buy 85 points.

Well, all that difference meant nothing for the HA (Hid. At.)

My hitting gang always delivers (never injuring a rival though :-) and rarely is benefited with the same treatment of receiving the same amount of fouls they commit on offense. (not even close).

They turn an emotionless game into a lucky-fans-contest drama.
I´m not good at sampling and estimation (barely passed that exam in college) but it´s pretty easy to tell that the foul system is not logical at all.

HU-HA! Beware of the Hitting Gang!

Yer great example.

In fact, you would almost call that a 'perfect storm' of a game, whereby the fouls pretty much won the game for the opposition.

There is obviously a combination of factors which a) Led you to having significantly better team ratings in pretty much everything except offensive flow and inside scoring. b) Comparable individual ratings of players, yet c) a significant difference in fouls with a distinct lack of steals by your team.

I wonder. - 3-2 zone + Run n Gun. Stamina issue?

Maybe its the steals/turnovers that led to fast-break mismatches? I dont know if they even exist in the game engine... either way, that game is just. well. weird.

Thanks for doing some research on this topic. The results are really no surprise to me though.

However I am still convinced, that this random aggression stat needs to see a change. Those foul prone players were advertised as players that would draw a significant number of fouls in exchange (tit for tat) - if that was true (which it obv. isnt), those aggro players would have some benefit.

My hitting gang always delivers (never injuring a rival though :-) and rarely is benefited with the same treatment of receiving the same amount of fouls they commit on offense. (not even close).

If you're playing Run and Gun, your team hurries down the field and tries to jack up a quick outside shot. When you have shorter possessions and don't try to go inside, it's not at all unusual to expect that you may not draw a lot of fouls.