BuzzerBeater Forums

Yes once I proved to you that this can happen in real life several times while you claimed it was impossible AND asked me to provide data, you want to stop debating about this, I understand perfectly why. Hard to recognize the truth.

I never claimed it was impossible. You have a direct quote where I say it's nearly impossible. You're welcome to keep looking for a game where 3 players got 81.2% from 3 and 3 makes, which is what my original post was about.

So far you found only 1 with the much more lenient requirement of 62.5%. I'll help you: http://www.basketball-reference.com/boxscores/200002200CH...

My opinion is that it is random, IG like IRL.

Backpedaling are we? You gave your explanation:

1/ You can claim high OD on your players but your OD grade is quite bad (strong high = 8.6 when your maximum this season is 10.6) because of your wrong GDP, your bad gameshapes and your C defending PG has no sub (stamina ?). On the contrary his outside scoring grade is pretty good.

You said it was OD, ST, Strong (bad...) GS. You also said it was random. So it was both non random (due to bad OD,ST, GS), random and presumably not scalable (i.e. no evidence same shooting vs worse OD does produce even better results, you ignored this multiple times). Impressive...

And for me that the GE is fine as it is. Your kind of game probably happened also 10 seasons ago before the fine tuning of Marin concerning JR (he likes to take it slow and I like him for that).

Do you know something we don't and has never been announced?

You conveniently forget the part were the shots were guarded by 16 OD and 14 OD or 15 OD. That was said in the very first post, you just dismissed it as the team rating not being good enough. My prolific inside rating was also based on a 12-12-13 IS in the frontcourt, do you think 'prolific' properly reflected it? Yeah I didn't think so either. Nachtmahr's guys shoot better than these 3 players on average and never shot 70% simultaneously.

Your kind of game probably happened also 10 seasons ago before the fine tuning of Marin concerning JR (he likes to take it slow and I like him for that).

I'm sure someone is going to be able to find one such game. BB is a game, probably based on RNG and expected averages, therefore you'd expect it to follow patterns. The long term average is affected by the conditions in the single game, such as among others the number of open/guarded shots and the OD faced. Now, when the patterns are no longer there or are subverted for no logical reason, it's fair to ask if something has been done to the GE in order to produce such results.

Let's agree to something. I will take 176 games from this season and will check the guarded vs unguarded 3pt FG%. Then I will do the same from 10 seasons ago at D2 level in France and compare them. If the guarded shot FG% has gone up then you will campaign with me to have Marin tell us if he's done something.

If you want to make a deep true analysis, take all the games of your league, put them in game analyzer and sort 3s compared to long 2s. Here you may be able to sort a true statistic and prove your point . One game analysis is just useless. You don't study a phenomena by its extrema, that is a basic statistics law.

That data sample is too small and any change in the GE is likely to go unnoticed until something extreme begins to happen when players with particular skills are in the same team. Your statistical analysis is a fairly pointless idea for several reasons:
1) we don't know the skills of the players involved.
2) data would need to be compared to other data which does not exist
3) a sample size of a single league is not big enough for a statistical analysis to yield robust results.
4) it would be impossible to understand whether improvements are due to changes in the relative skillsets between the offense and defense or changes in the GE without knowing whether the GE changed.

Last edited by Lemonshine at 11/16/2016 8:11:19 AM

To me the game is suspicious because the vast majority of the 3 pointers were guarded (by the OD I mentioned). It seems like we may have a similar effect to the 3pt FG% at the C position where the FG% is higher for guarded shots than open looks (because open looks are defended based on team defense which is usually a lot higher than a C OD).

The only problem is that it happens to positions where the OD is actually higher than the team rating (as demonstrated by the number of steals). This is important because it would mean that counterintuitively lower flow may result in higher efficiency.

If no changes have been done then we'll need to investigate further and it may well be a random outlier, but it would be great to know if what you say is confirmed:

As far as I know, the last review was announced and nothing was changed after that.

Last edited by Lemonshine at 11/16/2016 8:25:37 AM

ecause open looks are defended based on team defense which is usually a lot higher than a C OD

True interest : where does this come from ? Analysis from patient game with a guard at center ? You have statistics for that ?

This has been reported and debated over and over. You're a supporter, I don't need to tell you how to use the search function or google. I'm surprised that someone like you never heard of this before.

Let's just say that a C with low JS/JR facing a C with 1 OD and a team rating of 12 has very different results in open and guarded shots. It does not show up that much in patient with a guard offsetting a C as you'd assume that a C with 1 OD will not be able to stay with a guard outside the paint and most of the shots will not be defended.

Last edited by Lemonshine at 11/16/2016 10:40:57 AM

What are your 176 games ? French D2 now and ten seasons ago ? It is a good start for data collecting (way better than one game lol), but statistical deviation and uncertainties must still be computed ;) Try to avoid league with too much tankers as it could corrupt the information.

Don't cop out and own it please. If I have to do the work, I'd like you to be fair.

Well it is strange to see you criticize my offer for 176 games study while basing first your demand on one extreme game lol. But I will take that as a victory that you change from 1 game sample to 176 games with a serious methodology

I criticise it because you talk about doing a statistical exercise and in order to do that, you'd need more information and more data points. 176 games is not a great sample. An NBA season has 1230 games. I never meant my points to have a statistical foundation, I always treated it as an outlier, but something it shouldn't happen in a game based on RNG where a 0 FT shooter has exactly 0 career FTM.

So once again, if the game is an outlier (i.e. Marin hasn't done anything in the last couple of seasons in respect of 3pt FG% or GDP effect) then it's one hell of an outlier. If it isn't then I think it's fair we know what changes have been made.

Just to be sure, you claim that both open 3s and contested 3s have increased in 3FG% or only contested 3s ?

I'm claiming that unassisted and mostly guarded 3s went in at a 80% rate through 3 quarters. I'm claiming that there is a very high correlation between assists and 3pt FG% in real life and it should be the same in BB. A team with very few assists shouldn't be able to score at that clip. 26% assist rate is about half of each and every NBA game you linked.

Last edited by Lemonshine at 11/16/2016 11:03:42 AM

Why y'all discussing RL probabilities? This is no NBA simulation. Your discussion is beyond useless.
I almost feel triggered to throw in my usual 0% FT or 0 assists rant.

However I feel that a rare occurence of streaky outside shooting does not hurt, although some boxscores have looked fishy to me. If this seems to be a bigger issue I might dig them out.