Very cool, Tango. What the Grizzlies have done after the beginning of the year is very, very impressive. Got to give Hubie the Undead (thanks, R2K) mucho credit. What the Rocks show is more complicated... if we hadn't had that one, long losing streak of 5 games at the end of January, you could point to similar results for the Rockets record vs the Griz, but without the long winning streaks. Like Jeff mentioned, that's the maddening inconsistancy we fans go through on a day-to-day basis. Still, without that one losing streak, our model would look much more like Memphis. I think it points out our struggles as a team to figure out how to play with Yao as a dominating center. I don't think our players truly bought into that idea until recently. And it coincides with Yao figuring out what he can do at the same time. How we do for the next 20 games will give us a much better snapshot of the Rocks. We have a tough schedule. A period of .600+ ball going into the playoffs will have potential opponents coaches staying up nights, because of the one thing we have that they and almost no one else in the League has... that dominating center. And the half court game in the playoffs is tailor made for a team centered on someone like Shaq or Yao. Look at what the '86 Rockets accomplished. We had a 2nd year, dominating center, a 3rd year, All Star MVP dominating center who got pushed out of his position by the other guy, and a bunch of role players. Yes, we had Ralph, who hasn't a comparison on the current team, but we'd lost our backcourt to suspensions, we had some injuries (if I remember correctly) and we still ended up eliminating the Showtime Lakers of Pat Riley in 5 games and took what may have been the best team ever, the '86 Celtics, to 6 games before losing in the Finals. Sure, things are different now. Talent is far more diluted, fundamentals are often lost on players who learned more on the playground than college because they barely went to college. But one thing stays the same. If you get into the playoffs with a dominating center who's at the top of his game, surrounded by a decent team of role players and at least one other "star player", you're gonna make some noise. And I'm looking forward to it.
simply cause its trying to find an equation for acceleration of improvement (wins avg accelerating away from .500) this makes me want to go home and graph out SF's shooting % this year game by game.... would be fun practice.
The question I had in mind is: 1) the fits do not look very compelling, I am not sure about the R-square and the p-value of the regression coefficients. 2) if you throw in more regressors, such as the cubic term, then most likely you will get a better fit, and the curve may take two changes and tells a different story. So unless you can argue a quadratic form is the right form because of certain logic, it remains questionable what to take from these regressions. What you post does raise a point. Basically, we are analyzing momentum here, and you can argue it is similar to acceleration notion in physics, which has quadratic form. (I am more an economist than a scientist so forgive my talking physics) Still, if there are two big momentum changes in a season, you will find it hard to find a good fit in quadratic form.
Coming from a former financial analyst who made his living with charts and graphs and Pivot Tables . . . Nice work, Tango.
SWT, every player who JVG was forced to keep, Cato, MoT etc, is playing better this year than last year. Even those who are not in the rotation like Boki is playing better than last year, so there is no reason for me to believe that Posey would not do great here. In fact he is as much a JVG type player as anyone and Posey is the closest JVG would have gotten to Spree who JVG adores. You should also keep in mind we are all in the speculation game. As far as I am concerned if JVG can jump up on a codfish box and beg the world to appreciate CD for getting Spoon then we are entitled to let the world know that a most improved player in the league candidate was foolishly let go. Deuce, if you look at Memphis' rotation you will see that it is more about a philosophy rather than the availabiltiy of talent. I believe if Hubie Brown was coaching this team, Boki and Pike would both be playing. Sorry to speculate again, but I am just trying to make a point.
Very nice analysis. I think the points raised about the fit of the line with data is a good one (i.e., r-square, cubic terms, etc.), but the graphs do give us a general sense of "momentum." Data like this is typically "messy," and its often the case that the best fitting line doesn't "look like" its fitting the data very well, but again additional statistics would be required to assess that question. I think what we're tempted to do here is infer, based on these graphs, whether any given team will 1) make the playoffs, and 2) does "momentum" translate into post-season success. That's a more complicated question. Maybe if I have time I'll look at some recent teams in the past that caught fire in the playoffs, and see if that is related to "momentum" in the latter parts of the regular season (e.g., the Rocket's last championship).
see your are talking about RECORDwise, yes if we play .500 ball it would be a flat line on the graph. but we are talking about improvement in the team,not the record. example would be kings play .700 ball, then if all ofa sudden they start playing .500 ball, the line would be flat, so no improvment?? no that would mean the kings are playing like crap, because they arent a .500 team. WE are not a .500 team either, we are a .600 team(currently) so diverging from that slope would indicate improvement or playing worse.
Interesting, but it does not take into account opponents and strength of schedule. However, as we have seen in the last week, that is hard to quantify for....it looks like Memphis has a better ceiling, and Houston is steadily getting better. I would be interested in seeing the graph for last year for the Rocks to compare and contrast the 2. DD
according to JVG, you are what your record says you are, so, the more games we are above .500 the better we are. this is not rocket science, pardon the pun.
Tango, here is what I think: try your exercise on other teams, say Heats and Pistons just for example, or other teams that went through several ups and downs. I bet you will find a quadratic form will not give you good fits at all. Then you will have to throw in cubic term or even higher power terms to get a fit. This raises two points: 1) a quadratic form doesn't necessarily tells the trend. 2) if you throw in other regressors, will you find the similar prediction on the three teams you studied? Another thing to do to test whether what you are doing tells anything is: use ONLY the results from first 41 games to predict the trend and see how that prediction worked on the 20 some games afterwards. If that prediction didn't work well the past 20 games, it is hard to convice anyone that it will work for the rest of the season, right? I know 41 games is not a long period, but you get the idea. Now, if you are serious about it (say you are a gambler), then you certainly want to consider other factors, opponents' performance for one. For example, if you did this exercise on Minny a week ago, your prediction on them will break the roof. But if you look at their schedule for the rest of the season, I really don't believe they can keep that momentum. Memphis has a tough tough April, I don't believe they will climb that rapidly at the end as your chart predicts.
These charts are only another way to tell the truth that Rockets have had winning records every quarter so far.
ok let me put it this way, a consistantly good team- kings, win its first game, 1-0, then 2-0, 3-0, 3-1, 4-1,5-1..... so on and so forth. they started at one game over .500, then 2 then 3, growing in number as the season goes on. the graph will be a upward sloping line, not nessicarily curved. THEY have NOT improved during the season, they start good, and ended good. but their record has improved through out. get it? that is the essence of being a .500+ team, u win MORE than u lose. that does not mean u are improving as a team. otherwise 3 quarters of the western conference is improving. i dunno how else i can put it, just trust me on this. heres another example, a 6-4 team is doing just as well as a 18-12 team, they are both .600. how ever one team is 2 games over .500 and the other is 6 games over .500. if u want to IMPROVE u increase the the ratio, not the number of games over .500. can anyone else back me on this, this is as clear as i can put it
Agreed. If your winning percentage stays the same, it does not show improvement it shows consistency. However, if your winning percentage goes up...then the team has improved. I think the Rocks need a winning streak to say they really improved. DD
there is no doubt the rox have improved, chart or no chart. the problem is that we don't exist in a vacuum and other teams improve as well. showing consistency, at least as far as the rockets go, is improvement. i think someone should make a better graph if they don't think tango's shows enough nuance instead of criticizing his. and yes, a little win streak would be great. i think i feel one coming on.
Tango, great work. Real nice graphs. Everything I thought has been pretty much said, but I will reiterate to DearRock that Posey is most definitely not the difference. Not by a longshot! haha
It'd basically be a flat line because they've been right around .500 (as in a game or 2) practically all year long.
Thanks for the compliments all. Lots of good points being tossed up for discussion here! Sorry it sometimes takes awhile to respond with posts. Consuming work and then daddy duties in the evening make me a habitual midnight oil poster! A few thoughts on all being discussed... R-SQUARED, CURVE FITTING DISCUSSIONS: canoner2002, ckahlich001, outseam - pretty basic stuff I'm doing here. I'm mainly using MS Excel to do the lifting and added the trend lines as an afterthought through excel and used the default setting for polynomial regression which is quadratic. Admittedly I didn't pay attention to trying to increase the r-squared fit but as mentioned by ck and outseam I think the quadratic eqn. gives us more of view of overall momentum. I played around with higher order polynomials (excel only lets you go up to x^6) and can get higher r-squared values but that tends to give us more "instantaneous" trends if you know what I mean vs. a view of the longer run. Interesting data as well though. I'd be happy to try a few other things and post them for comment regarding appropriateness. (I had to get help from my wife who was a former natural gas rate analyst to help me with this!) Canoner- good thoughts regarding best fit for "predicting" (and gambling!) but really that's not where I was headed. I was more interested in the historical vs. the future. An interesting thing that could be done as well is to try to assess the "riskiness" or "consistency" of teams much like analyzing stock volatility. STRENGTH OF SCHEDULE ETC: I've toyed with the idea of adding that in along with other possible variables but that'll be a project for some other time! CAVS, JAZZ, PISTONS, HEAT: I'll do those shortly here and post them. REGARDING "IMPROVEMENTS": I think this is a really interesting question and cuts at some of the fundamental differences in expectations of different fans on this BBS. What is improvement when referring to this sign post? I think how you answer that question gives insight as to what you expect from the Rockets. My opinions as to what the data tells us about the Rockets: (1) Even if the Rockets curve tends to be "straighter", in my book it does represent improvement not in the sense of the season itself but the fact that we are our 11 games over .500 for the first time since '99 - a great milestone demonstrating improvement. (2) The curve is not nearly as "exponentially" curved as the Grizzlies meaning that yes we've improved over the course of the season in winning .pct which is a great, but we're improving at a slow rate.
The clear thing about these graphs is that they are time sensitive - a few months ago the nuggets and grizzlies would have been reversed, the rockets would have been the same - which is odd, because the catchphrase you keep hearing from JVG and JJ and the media is that the team just isn't consistent enough. In reality, they're consistent, just not consistently great - consistently decent. There haven't been a lot of enormous highs and lows. The Grizz run of late has been impressive, but the problem is, few teams have ever been successful with more than an 8 man serious rotation in the playoffs. Depth is important to have in case of foul trouble, etc., but for some reason those longer rotations never work at the end.
