Just realized I could start new thread now, yeah!:grin: Okay, this is my first time doing this kind of data analysis. I'm no Bball expert, just a nerd who has good affinity to numbers. I'm not trying to introduce new formula, but rather just try different visualization of number distribution, and hopefully see some trend with limited sample size. First, I've seen people talking about turnovers, so I pull out data for our 5 main players: Lin, Harden, Pasons, Asik, and Delfino. I grouped their each game's turnover number by win or lost games, and following box plot shown the distribution. A little explanation on how to read the graph: Y-axis is the turnover number per game. Each box shown the distribution of middle 50% of data point, and vertical line gives the span of all data point, whereas the white line in the middle of the box indicate the median of the data set, the number is shown in the table under the graph. Blue group contains data from lost, and red indicate win games. Individual dots are considered outliers. As shown in the graph, even though people complained about Delfino, and Pasons's TOV, collectively there is not much difference in games that we won vs. games that we lose. They just statistically will turn over that much, or that less. Harden and Asik on the other hand, better watchout their turnovers, as it does shown they tend to have more turnovers in games we lose. Jeremy is the only one among the group that have opposite trend. As quite some LF said before, Lin's high turnover game not necessarily linked with team lose, he actually tend to have more turnover in games that we win. The difference is small though, not sure whether it is statistically significant, but the trend is there. Maybe when Jeremy’s low turnover game was less aggressive, or less ball handling, at this time, maybe we should have less concern about Jeremy’s turnover. However, I do wish he break this reverse correlation someday, and control his turnover without limit his aggressiveness.
The second graph is by %USG. I don't really like %USG as that only take consideration of FGA, FTA and TOV, but not taking consideration of assist, or hockey assist, so it is not a true indication of ball handling. I'd take this number as indication of a player looking for his own shot. One thing obvious is that Harden have very high USG% among the team, which is normal. Lin though is the second, really don't have any separation among Pasons or Delfino, or Asik. I guess it is normal too? Interesting to see that in games we won, Lin had slightly higher USG%, suggest that he need to be more aggressive, and look for his own shot, and more interesting part is, Delfino is opposite.
Ooops, forgot to change the title, that does not make any sense...... However, it is hard to make much sense for the graph I'm putting here too, so, who cares.... All these number game is to make people confused, and they would just take my word, because I have all the numbers to support what I'm saying, haha....
Okay, whoever survived and remain conscious after previous two graphs are welcome to take the challenge of this one. This is a scatter plot of each player’s Offensive rating (ORtg) Vs. Defensive rating (DRtg) for each game. Again, win game labeled red, and lost game labeled blue. Basically the higher the ORtg, the better this player played offense in that game, and the higher the DRtg, the worse the player played defense in that game. I found James Harden had the most interesting scatter pattern. Basically his performance in win games is clearly separated from games we lose. If one draws a diagonal line, you could separate the two groups of red dots and blue dots. The diagonal line pretty much means whatever effort a player put on court, he has a 50:50 split of effort in offense and defense. And for dots that under the diagonal line, indicate that player have a bit more effort in D than in O, and dots that is on top of the diagonal line indicate in games where player have less effort in D than in O. So, for Harden, I guess we could all benefit when he play a bit more D. The other players do not have so clear cut trend in win games vs. lose games. Maybe Pasons have slight trend of games with better D tend to win more than he play less D. (A horizontal line across 108 (his average DRtg). And it has less trending with his offense. For Delfino, I could only say, he is a vet, his D and O is less affected (or affecting less) to game win or lose, basically he will play the worst D among these 5 players, and he will play unpredictable (high variation) O in each game. And for Omer, if he play poor D, and poor O, we lose for sure, and when Jeremy have good D and good O then we have high chance to win. And Jeremy has more winning percentage when he has better Offensive game, period. To my surprise, even though many people complained about Harden’s D, and saying Parsons had a reputation as a good defender, Harden’s D is still better than Parsons. And look at the distribution of Harden’s game, he does shown a super start’s trend with pretty consistent performance in both D and O (his data has the least scatter compare to others). On the contrary, the rest 4 players have much bigger swing from game to game in both D and O. Omer has the best D, Jeremy is second, followed by Harden, Parsons and Delfino is the last. In terms of O, Harden and Parsons are both good, Harden usually gives better impression in O, I think is due to volume, and also his consistency. Parsons has big swing in good game vs. bad game though.
okay, done with graphs, now come with numbers. Break down some of the performance by month for these 5 players. Delfino: slightly increased play time, but decreased FG%, and 3P% in Feb. Nice improving trend of offense, but big drop off in Feb, bad trend of DRtg. Chandler Parsons: Pretty steady performance with a slump in Dec. D in Feb no good. James Harden: out of his mind in Feb, FG% and 3P% both over 50%. Average offensive rating is the same as the first game he played as a Rocket. Remember how people got excited after his first game? He is now averaging that performance for a month. Ha, how fast people desensitized with Harden's super performance in O. And focus only on his D.
This graph clearly says that if everyone plays hard on D, we actually win less games. Weird. rarely lose when our defensive rating sucks. Maybe something about the way it's calculated
Jeremy: Impressive improvement in 3P%, suprisingly steady of FG%. Don't we all think he was struggling in shooting at the beginning of the season and get much better now? He was acturally shooting his career average FG% all along. think this way, with his improved 3P% and steady FG%, his 2P% is dropping, slightly though. But that is fine, according to Morey's money ball theory, he is shooting same amount of shot per game, and same average of 3s, with increased 3P% and slight drop of 2P%, he still got increased points per game, and increased his efficiency. Which is reflected by his ORtg. Omer: while CF surfers busy barking at LOF/LOH about Jeremy's 3P%, Omer Asik quietly improved his offensive game, with FG% improved from Nov 44% to Feb 55%, Offensive rating increased almost 30 points per 100 possession, wow! I surprised myself big time looking at these numbers. With butter fingers, this 7ft guy would shot bank shots, dribble drive, fast break, I will just be completed if one day he give me a 3P, so my data sheet would not mess up with all 0 3PA and 3P%. Okay, finally, a really bad trend I’ve seen, across the board, all five players have decreased D in Feb, yes, include Asik, and include my boy Jeremy Lin. Coaches better do something!
uh, my understanding is that they are bonding, and they will need double bond or triple bond to be stronger.......sometimes high pressure will help the double bond formation...... heat will help too...... do we still have game facing heat this season???:grin:
Okay, after all the intensive training in previous graphs, this one is so easy to see the trend. Here are the DRtg (deffensive rating) for each player grouped by game win or lose, and only the road game shown here. Obviously, defense, defense, defense, for ALL FIVE PLAYERS!!!! Interestingly, home game don't show this clear trend.
cooooool nerd party :grin: Your graph confirms that Harden is more consistent and with a better avg on D, but to be fair to Parsons, he does have more games deeper in the "Good D" zone as a sign of his capabilities. It would be nice to count the number of games with D rating below 95 for example. Omer even has 3 games below 80
So, home game is all about offense. This graph only have data from home games, and in order to have a home win, we will need Omer, James and Chandler to have a good offensive game.