16 replies to this topic

[Red94]

New post: Winning, Losing and Turnovers

[thejohnnygold]

Good article. It has always been my view that turnover margin is what kills you. Regarding that, we aren't that bad overall. The league averages between 11.2 (nyk) - 14.7 (houston)...a 3.5 margin from top to bottom (remember, we have the highest # of possessions). I generally assume a 50% conversion rate on turnovers (it may be higher) so that is roughly a 4 point spread. What gets Houston in trouble are the nights the turnovers are excessively high and the conversion rate is higher so that a -8 turnover margin could very well be a +12 scoring bonus for the other team. This does not take into account the negative impact of us not getting a shot off. Presuming a 50% conversion rate there, we just lost 8 points for a total game effect of -20 for Houston. Ouch. (I recognize there are other variables, but for neat and tidy math this will do)

I have preached it before, and will do it again. When we have our turnovers under control--not eliminated, just minimized--the Rox are nearly unbeatable. (Ugh, I just spent 20 minutes looking for simple win/loss splits to compare turnovers and assists...if anyone can find them please post a link--thanks)

The times I can remember checking the stats after big wins the turnovers are way lower and the assists went way up. I draw the conclusion from this that a lot of our turnovers come off of passes--probably passes that, if they connect, result in an easy (high %) bucket. So again, this is probably something the Rox are willing to live with since I imagine it evens out in the end. The nights when the kick-out threes are falling help bump those assist totals as well.

EDIT: apparently the stats vary and Houston has 16.3 TO's per game on some sites. a tad larger, but still only relevant when compared to our opponent game by game.

[Kade]

I disagree

I have preached it before, and will do it again. When we have our turnovers under control--not eliminated, just minimized--the Rox are nearly unbeatable.

Disagree. OKC is second in most TO per game and Denver is third and both those teams have a legit shot at going to the Finals. The Rockets will be unbeatable when we get a coach who has shown he can succeed as a head coach and knows how to effectively use his players and we get a few more solid players.

http://www.teamranki...novers-per-game

[Sir Thursday]

I disagree


Disagree. OKC is second in most TO per game and Denver is third and both those teams have a legit shot at going to the Finals. The Rockets will be unbeatable when we get a coach who has shown he can succeed as a head coach and knows how to effectively use his players and we get a few more solid players.

http://www.teamranki...novers-per-game

How many turnovers OKC and Denver commit is irrelevant. Yes, it is possible to succeed in spite of committing large numbers of turnovers, but that doesn't change whether or not the Rockets could be doing better if they eliminated a few turnovers per game.

On the article itself, I think as a general rule when you're trying to do correlations between two statistics it would be better to use some kind of regression analysis, which reduces the skew effect of outlying results slightly. Also, it makes pretty graphs . Earlier in the season I tried doing a regression between number of turnovers and final margin but there wasn't much of a correlation...perhaps that's changed now we have more data though...

ST

[feelingsupersonic]

I disagree


The Rockets will be unbeatable when we get a coach who has shown he can succeed as a head coach and knows how to effectively use his players and we get a few more solid players.

I am pretty sure the Rockets, who are a young team that has about half a season under their belts as a unit, would not instantly become "unbeatable" with a "proven head coach," nor would they next year. I think the overwhelming majority of us would agree that the Rockets are overachieving or on schedule at the very least. Give us all a break with the coach talk, at least until the season ends.

[thejohnnygold]

Sir Thursday, you feel like updating that graph? I'd like to see it.

Kade, my point is that our turnover margin is what is relevant. Due to our higher number of possessions/game our total turnovers will be higher. Our opponents turnovers also increase in games against us due to this fact. Thus, a direct comparison of averages is largely irrelevant. Game by game what matters regarding us winning is the turnover margin--the difference between our turnovers and our opponents--in that exact game. If we turn it over 16 times, but they turn it over 15 the end result is negligible.

OKC and Denver are at the top of the list for both # of possessions and offensive efficiency. This allows them to easily overcome the high number of turnovers--which isn't actually as high as it looks. All three teams (Rox included) also get a lot of steals and points off of turnovers.

I think the turnover ratio is but one thing the rockets can, and should, improve upon that will ultimately bump their winning percentage higher. I'd also like to see more assists as assisted baskets tend to be higher percentage; although, I recognize that our offense is largely predicated on one player driving to the hoop which skews our numbers negatively. Asik's fumble fingers single-handedly skew all of our numbers as he drops 3-4 passes (turnovers) that would have been dunks/lay-ups (assists, points, efficiency) per game. If he can get that down to 1-2 a game we already receive a huge boost, statistically speaking. BTW, that's not McHale's fault.

EDIT: I have posted statistical data below...just click the link

[thejohnnygold]

I am pretty sure the Rockets, who are a young team that has about half a season under their belts as a unit, would not instantly become "unbeatable" with a "proven head coach," nor would they next year. I think the overwhelming majority of us would agree that the Rockets are overachieving or on schedule at the very least. Give us all a break with the coach talk, at least until the season ends.

FS, I saw a projection today (forget where...a smaller site I'd never seen before) that had Houston finishing with a .595 winning percentage...good enough for 5th place in the west according to their data. Talk about over-achieving...if that comes to fruition I 'd say nobody had us penciled in behind OKC, SA, LAC, and DENVER. I think in that scenario we would face Denver in the play-offs though....bad match-up for us.

[thejohnnygold]

chart post fail : (

[thejohnnygold]

Ack! I keep forgetting that spreadsheets won't post here. I made a link to a webpage that has the charts. Basically, I couldn't find the splits I wanted so I made my own. There are 2 charts: one has houston vs. opponent turnovers and assists in wins and the other in losses. I feel the numbers speak for themselves, but welcome any input, questions, or criticisms. Enjoy.

LINK HERE

EDIT: sorry for the bad formatting

EDIT 2: the info. on the side is for any figures that do not fit the norm. I found the stat with the greatest deviation from the norm to help possibly explain the anomaly.

EDIT 3: LOL, it is hard to read. focus on the avg's at the bottom of each. they are easier to see and tell the whole story.

[thejohnnygold]

Here is the TL;DR for above

In losses, Rox avg. 17.44 turnovers and a -2.92 differential while averaging 21.4 assists with a -2.36 differential

In wins Rox avg. 14.36 turnovers and a +.04 differential while averaging 24.71 assists with a +4.18 differential

The difference is roughly 3 fewer turnovers and 3.6 more assists...coincidence?

Last night GS win Rox had 11 to's and 32 ast's vs. GS's 13 and 25...+2 and +7 respectively.

Lin and Beverley combined for a tidy 15 ast's and 0 to's. Keep it up guys.

[ale11]

Thanks for the data, johnnygold, great read. And I agree with you about one very important thing: about 3 or 4 turnovers per game come from passes directed to someone very well positioned to make an easy bucket (usually, to Asik under the rim). Those 3-4 turnovers could easily be 3-4 assists with work (Asik developing some kind of "glue hands") and chemistry, understanding and recognizing where a teammate likes passes to be thrown (not throwing them at Asik's feet) and of course, a little luck regarding to opposite defenses, read where someone is distracted and leaves our player unguarded (several Parsons' cuts to the basket comes to mind right now).

It's not that bad to turn the ball over that many times, what's maddening is when our transition defense sucks. Also, if you manage to induce the opponent to a near/bigger amount of turnovers, then it's not that big deal.

[myjohnlai]

If you take risk in passing for a high percentage shot, the risk of TO goes up. But it's worth it because it's just like taking a shot.

[thejohnnygold]

Agreed. These are things I am willing to live with and I anticipate will diminish as they become more familiar with one another. I read somewhere earlier that quoted Chris Bosh. He was talking about the Heat's current offensive play and overall for the year. Basically, he cited the confidence that comes from knowing one's role, knowing the shots you will be taking, and having done it for a couple of years now. The heat are 2 years ahead of schedule than the Rockets in that regard...if the Rox keep this core group together (all signs point to yes) I think we can expect improved efficiency and, more importantly, consistency...which is scary for the league's highest scoring offense.

[Sir Thursday]

Sir Thursday, you feel like updating that graph? I'd like to see it.

Ask and ye shall receive!

The following graph plots the final margin of the game against Turnover Rate (% of possessions that end in a turnover):



As you can see, there does appear to be a reasonably strong correlation here. An online calculator measured the correlation coefficient at -0.499 (for reference, 0 would be no correlation, while -1 would be a perfect negative correlation). That's not cast iron, but it does show that the general sentiment of "If we commit fewer turnovers, it's likely that we'll do better" is a valid one (common sense, really ).

Looking at the gradient of the graph is illuminating. The slope implies that, on average, a 1% change in TOR correlates to an approximate 7 point swing. The Rockets average just under 100 possessions a game, so a 1% change is approximately equivalent to 1 turnover. Now that does seem quite high - after all, common sense dictates that the most difference a turnover can make is 6 points. So the implication is that there is a link between the number of turnovers and the final margin that goes deeper than just the individual play. In other words, if the Rockets are making more turnovers than usual, it is likely that there will be other facets of the game that are not working well either.

That finding is generally backed up by the eye test - when the Rockets are committing a lot of turnovers they often look sloppy. Often high turnover numbers are a symptom of fatigue, and with tiredness comes porous defence and shots being less likely to fall.

It should be noted however, that this does NOT mean "Every turnover is worth 7 points". As I said earlier, it is clearly not the case that the turnover itself causes a 7 point swing. What we can say is that fewer turnovers is a good sign that the team is playing effective basketball and are more likely to be outscoring the opposition.

ST

[thejohnnygold]

Nice. Thanks, ST!

It certainly does support the synergistic idea that everything is inter-connected and inter-dependent. In basketball, rhythm is important and turnovers hurt you while helping the opponent--in that regard--not to mention the obvious extra scoring chances vs. fewer for you, etc. I would say it even gets into the psychological realms of confidence, etc.

Perhaps this idea, or something akin to it, has influenced Houston's defensive schemes which seem to encourage taking chances on steals and getting lots of easy, fast-break, points. Both are dis-heartening to an opponent and over the course of a game could wear one down. Psychological warfare in basketball????

[Sir Thursday]

Nice. Thanks, ST!

It certainly does support the synergistic idea that everything is inter-connected and inter-dependent. In basketball, rhythm is important and turnovers hurt you while helping the opponent--in that regard--not to mention the obvious extra scoring chances vs. fewer for you, etc. I would say it even gets into the psychological realms of confidence, etc.

Perhaps this idea, or something akin to it, has influenced Houston's defensive schemes which seem to encourage taking chances on steals and getting lots of easy, fast-break, points. Both are dis-heartening to an opponent and over the course of a game could wear one down. Psychological warfare in basketball????

I didn't do this analysis for the Rockets' opponents (which might be an interesting next step), so it's probably not valid to extrapolate this trend to how other teams perform just yet. I could well imagine that there are some teams that don't have the same level of correlation between their overall performance and turnovers. Might make for an interesting study.

ST

[thejohnnygold]

Sounds like a job for......SIR THURSDAY!!!! you know, if you wanted to....