The Tank 11'12 thread.

[floppymoose]

The Lakers traded for Kobe (or for the Hornets pick) at 13. They did not get terrible to draft Kobe. that one goes solidly in the don't-tank column.

Right. What I think this list indicates is that you need to hit a draft homerun to get to 50 wins. If you can get a hall of famer at pick 13, then you don't have to go through <30 win seasons.

But a lot of these teams, most of them even, needed that 20-something win season to get their later success.

[floppymoose]

And you looked at the last 10 years.
The article looks at every 10 year period going back to 1978.

Usually more is better when it comes to data. In this case, though, I'm really not sure. I wasn't an nba fan before the late 90's, so I don't fully appreciate the differences. If teams more properly rate draft prospects now, for instance, then things that may have been possible before the 21st century may not be possible now.

[Sleepy51]

And you looked at the last 10 years.
The article looks at every 10 year period going back to 1978.

Usually more is better when it comes to data. In this case, though, I'm really not sure. I wasn't an nba fan before the late 90's, so I don't fully appreciate the differences. If teams more properly rate draft prospects now, for instance, then things that may have been possible before the 21st century may not be possible now.

I haven't seen any evidence over the last 30 years that teams are any better or worse at drafting than they used to be. If anything, I'd say teams make slightly worse risk assessments in the top 10 because of the early entrants, but that's purely an anecdotal suspicion.

[Sleepy51]

The Lakers traded for Kobe (or for the Hornets pick) at 13. They did not get terrible to draft Kobe. that one goes solidly in the don't-tank column.

Right. What I think this list indicates is that you need to hit a draft homerun to get to 50 wins. If you can get a hall of famer at pick 13, then you don't have to go through <30 win seasons.

But a lot of these teams, most of them even, needed that 20-something win season to get their later success.

The value of the piece to me was about probabilities. And the amount of data available suggests a significant disadvantage to becoming a 50 win team if you sink to 20 wins. Draft home runs, lotto order, David Robinson's injury, these are all elements of fortune that are part of the data in the article. In the end, over a significant sample, losing did not correlate with meaningful winning more than winning did.

[floppymoose]

And yet the last 10 years does support the "get bad to get good" story. It's not the only way, but being smart with high draft picks is still the main way, lately.

[Sleepy51]

And yet the last 10 years does support the "get bad to get good" story. It's not the only way, but being smart with high draft picks is still the main way, lately.

I don't know that one 10 year period is enough to declare a trend when we have a lot more data than that.

[floppymoose]

Speaking of tanking, Gallo is reminding me of Dirk a bit in these playoffs.

[EvanZ]

Speaking of tanking, Gallo is reminding me of Dirk a bit in these playoffs.

He was third most improved player this year according to my A4PM rating.

[Sleepy51]

The other element of the discussion that should be noted is that the article does not specifically address data on "tanking." It addresses data regarding "being bad."

A good team like the Spurs beset with one real and one exaggerated injury will certainly benefit from adding a top draft pick. They know how to draft and develop as evidenced by the fact that they didn't really suck in the season where they orchestrated suck for draft purposes.

I'm much more interested in the solution for a genuinely bad or even mediocre team. It would be a massive cherry pick, but I'd want to throw the Spurs out of any conversation about tanking in regards to it's relevance to actual bad teams.

[floppymoose]

I think there are a lot of factors at work here.

There are teams with bad ownership, bad front offices, bad coaches, bad luck.

There are teams that have good versions of the above and also get lucky.

I'm really trying to isolate for a particular question: how did 50+ win teams get to where they got? And I don't think Berri's data comes remotely close to answering that question. And I wouldn't trust any data analysis he did, just based on my past experience.

That doesn't mean he isn't right, because even bad analysis will get the right answer more often than not. It just means I want to start over, specifically identifying good teams and looking backward from there to see how they got there. The teams that failed to get there, I don't care about. They are cluttering the data. If almost every 50+ win team had to get bad to get good, then I don't care if a bunch of other teams got bad and didn't get good.

[Sleepy51]

Don't let Berri bias get in your way. That's not Dave Berri's article or data. Some dude named Arturo wrote that. I am no Berri fan. This isn't WP mumbo jumbo. But if you are starting from a premise that the data is going to be wrong because it happened to be posted on Berri's blog then no one can argue with you unless they are willing to reproduce all of the work. That's not a very reasonable standard.

[Sleepy51]

The teams that failed to get there, I don't care about. They are cluttering the data.

That doesn't make sense. You already acknowledged that there are other ways to get there. If there are other ways to get there then the probability of success for a particular route matters. The data seems to say that the other ways of getting there have a higher probability of success, with the exception of the "already be a 50 win team" path for the one 10 year period you singled out.

That's like saying stocks are a bad investment because they happened to be a bad investment for the one 10 year period that ended in '09. (That may or may not be true, but it's not because of what the data said in one point-to-point 10 year snapshot.)

[migya]

Those stats mean nothing. A team that improves does so by getting one or more very good players that fit in and make that team much better. A very good coach can make a pretty big difference in some situations as well. Bogut, might really make a difference here. Maybe that addition of a pseudo-dominant Center, who plays very good defense and rebounds very well (both the team's biggest weaknesses) can make the team that much better. I don't think his addition can make the team a contender, at least one more very good move is needed, but he could make the team at least a 50 win one.

[floppymoose]

The teams that failed to get there, I don't care about. They are cluttering the data.

That doesn't make sense. You already acknowledged that there are other ways to get there. If there are other ways to get there then the probability of success for a particular route matters.

I think it's a bit more complicated than that. If 75% of the teams that got good had to have a stud lotto pick to do it, but 100% of the teams that didn't get good had high draft picks, that doesn't mean that you don't need the stud draftee to get good. It may just mean that the teams that didn't get good squandered their draft pick.

The different variables in this aren't independent. (In particular, getting a stud draftee is not independent from having a high draft pick which is not independent from being a bad team.) If they were your statement above would be correct.

[Sleepy51]

The teams that failed to get there, I don't care about. They are cluttering the data.

That doesn't make sense. You already acknowledged that there are other ways to get there. If there are other ways to get there then the probability of success for a particular route matters.

I think it's a bit more complicated than that. If 75% of the teams that got good had to have a stud lotto pick to do it, but 100% of the teams that didn't get good had high draft picks, that doesn't mean that you don't need the stud draftee to get good. It may just mean that the teams that didn't get good squandered their draft pick.

The different variables in this aren't independent. (In particular, getting a stud draftee is not independent from having a high draft pick which is not independent from being a bad team.) If they were your statement above would be correct.

If you are willing to disregard the majority of the data, sure you're right. But once again, I'm not ready to concede that the previous 23 years and the multitude of rolling 10 year samples contained in that data are less meaningful that the ONE point to point 10 year period you are basing your analysis on.

We obviously aren't going to convince each other on that point, so we're probably at an impasse until another 23 years roll by.

[floppymoose]

Well, I'm willing to look at more data than what I did above, it's just that I have to understand and trust that data. I have neither understanding or trust of the data in the charts, but I do for my own view of the 2002+ years.

I absolutely *do* think that the way to approach this is to look at what teams were good, and determine how they got there. If a bad team tried to do the same thing but didn't do it right, that doesn't invalidate the approach - at least not to me. I'm predisposed to believe this isn't all luck; I think there is some skill involved for a lot of these teams in how they got good.

[Sleepy51]

Hmmm . . .

Maybe there's the answer to the data. The "plan" to get bad (and eventually recover) may indeed require certain skills. But, the execution of that plan and the implementation of whatever skill is involved is pretty obviously dependent upon the luck involved in the availability of a transformational player when you get that bottom draft pick, and the luck involved in not getting leapfrogged in the lotto. This may not all be luck, but there sure as heck is a luck element and that luck element seems to hold back bad teams, despite their best efforts to skillfully execute the "lets get bad to get good" strategy.

[floppymoose]

I absolutely agree with that.

To me getting bad to get good, is giving yourself the chance to get that draft luck. Chicago basically stayed bad until they got that luck. I don't think that was by design, as they got unlucky (Jason Williams, Eddy Curry) and also screwed up (giving up on Chandler, drafting guys like Fizer).

Then there are teams like the Raptors that have been a bit unlucky (got the #1 pick in 2006, a weak draft at the top, instead of in 2007, when Durant was available), but really their main problem was that they had less draft eval talent than the average Joe Sixpack yelling from his couch in tighty whiteys.

Then there are teams like the Thunder that both got lucky and were very complete in their efforts to rebuild through the draft - making the tough choices like trading a prime all-star for a draft pick.

[Sleepy51]

I absolutely agree with that.

To me getting bad to get good, is giving yourself the chance to get that draft luck. Chicago basically stayed bad until they got that luck. I don't think that was by design, as they got unlucky (Jason Williams, Eddy Curry) and also screwed up (giving up on Chandler, drafting guys like Fizer).

Maybe it's again important to parse out our terms here. A 30 win team is a bad team. A 40 win team is a bad team. Being a 20 win team is Uber-bad and that seems to be the proposition that the probability tables discourage over the long view. Uber-bad teams get leapfrogged, uber bad teams develop players poorly, uber bad teams probably deal with players sitting out more games due to injury and guys bolt in free agency because they hate your stupid organization. All of these same things can happen to 30 or 40 win teams as well, but being a 20 win team doesn't seem to be reducing their likelihood enough to make it a good bet. You don't have to scuttle the ship and max out your lotto balls to be in the lotto. 30 win teams are still in the lotto. 40 win teams are still in the lotto in the west. We've been in the lotto as a 48 win team. I think maybe the significance of this article is that given the lotto structure and that the worst team only has like a 25% chance of getting the top pick, and it's maybe a 15% chance of that top pick, and/or the player you choose being transformational anyway, that you are playing a suckers bet by trying to rack up ping pong balls.

Maybe it's Warrior experience bias that has us believing that a 30 or 40 win team can not make progress. It appears that our experience is contrary to what the data says. Something has apparently kept the 30 and 40 win teams improving more consistently than the 20 win teams over 33 years of data. I'm betting it's the opportunities that present themselves from successful asset management. You bottom the boat, and you diminish your players value and ability to be negotiated in beneficial trades. Maybe its the odds of winning a trade that tilt in your favor when you keep the 30 and 40 win teams progressing over the long view?

Or maybe the Bucks/Warriors malaise is real, but not inherent. Maybe the Bucks, W's, Hornets, Wolves, Nets, Cats & Kings are the CAUSE of of a decade of really bad NBA management (leading us to he recent lockout and fire sale on NBA franchises) that has created the lotto dependent dynamic we see today? Maybe it is operator error and mal-investment by a specific set of incompetent actors that that explains the shift you perceive in the recent history? Maybe the Warriors experience is the CAUSE rather than the effect? I'm just spitballing at this point, but I think it's something to think about.

[floppymoose]

Maybe it's Warrior experience bias that has us believing that a 30 or 40 win team can not make progress. It appears that our experience is contrary to what the data says. Something has apparently kept the 30 and 40 win teams improving more consistently than the 20 win teams over 33 years of data.

For what it's worth, I'm pretty sure the tables and graphs in that article are "double counting", in that 44 win teams that started out as 24 win teams but got to 50 win teams are being counted as a success for both the 24 team case and the 44 win case.

Here is a year by year look at the data, given a particular starting point:


Look at 7,8,9,& 10 years out. The crappy teams have an edge on the 40-44 win teams. It's a small edge but it's there. So the question is, is that edge really that small and thus not worth pursuing? Or is it just one part of a bunch of factors, and so it's being overwhelmed in the data a bit by the sum total of those other issues: luck, coaching, trade smarts, having a good plan, draft smarts...

Another interesting row is 4 years out. The 20-24 win teams had an edge on the mediocre teams.

One things does seem pretty clear from the data... you probably don't need to win fewer than 20 games. You want to be bad enough to get a top pick, but not so bad that there is zero talent on your squad to build around, or to use as trade assets.