One month ago I had the idea to look at nba players in a different way, so I started writing a a program which would tell me pure team statistics while a specific player is on the court. It's not important to know which player did what, but something like "how is *detroits defensive rebound/opponents offensive rebound ratio* while ben wallace is playing" is..I finished the program a couple of days ago and started analysing the playoff games (so far I have only games of the first round) I ended up with alot of interesting and stunning statistics: Did you know dallas's FG% was 20% higher when tracy mcgrady was playing? did you know robert horry has a better impact on the spurs than tim duncan? Did you know that detroit is alot better with mcdyess on the court than with rasheed wallace? Did you know joe johnson ups phoenix's effective FG% by 17%? and triples their blocks? Did you know the spurs shot 47% when steven hunter was on the court, and 72% when he wasn't?(hunter also lowered memphis's FG% by 10%) Makes you wonder if the so called "stars" really deserve their status...
I think random stats like this can be very misleading sometimes. We have to take into account that while a reserve, such as Steven Hunter, is in, he is actually also playing against the other team's reserves. He isn't really going up against Tim Duncan the whole time. Also, McDyess is not better than Rasheed, it's just that when teams are going into their bench, Detroit has the better bench. McDyess is doing better vs Miami's bench than Rasheed is vs Miami's starters. To me these stats just show which team has the better bench players. Yes, they may be an excellent indication of who the difference-makers are in terms of benchplayers, but that doesn't mean they're better than the team's actual stars.
of course some of those mentioned get alot of minutes in so called garbage time(of which there might not be alot in the playoffs)...but since you mentioned duncan...he is actually worsening S.A's FG%.. judging hunter after only 2 games vs S.A provides too few data you're right...i just thought it was an interesting stat since he's been doing the same to memphis too and his oncourt/offcourt stats are just huge atm
^he might be making the Spurs' FG% lower than it might be if Horry was on the floor(was htat the comparison you made?), but Duncan brings so much oto the floor, that it outweighhs his minor drop in FG5, I mean, its not like he isn't hitting any shots, so thatargument, I think is very...invalid to say the least.
the thing is, it doesn't matter who hits any shots, just the team. even if horry didn't hit any shots at all, it wouldnt matter, maybe the ball just always falls in the right hands when he's oncourt hypothetical scenario: when duncan plays most attack go through him, he hits 48% which is alright. when horry plays the attack go through parker and ginobili...they hit alot more than 48%(effective FG%). thus horry gets the better(FG) numbers. duncan might get better numbers if he passed more to parker and ginobili but he doesn't, so in the end it's better for popovich to bench duncan a little more. you guys shouldn't get too hung up on the horry/duncan argument, he didn't seem to be in best health when they played denver...his stats look alot better vs phoenix so far(especially the rebounds)
First off, Tim Duncan is still not 100% recovered from his ankle injury and it will linger probably for the rest of the playoffs. I think that is affecting his rythym, because when I watch him he isn't as aggressive and his post moves aren't as good right now. Horry also had a few good shooting nights lately after being left wide open too much by Phoenix, so he will appear statistically better atm. Hunter may improve the defensive FG%, but what are the offensive stats when Hunter is on the floor?
they improve too...(phoenix's offensive stats when hunter plays) here are some stats for phoenix (all their playoff games combined) marion: bad stuff:FG drops by 6% opponents FG goes up 8% good stuff: less fouls for phoenix and more for opponent ups blocks by 30% ups offensive rebounds by 30% and what makes him absolut top: cuts opponents offensive rebounds in half! stoudemire: bad: ups opponents FG% by 9% lowers rebounds and blocks ups defensive fouls alot good: ups assists and he's good in the single category i've seen all *stars* being good at: causing alot of opponents defensive fouls hunter: ups FG% by 4, ups rebounds and blocks lowers opponents FG% by 6
<div class="quote_poster">Quoting back2newbelf:</div><div class="quote_post">One month ago I had the idea to look at nba players in a different way, so I started writing a a program which would tell me pure team statistics while a specific player is on the court. It's not important to know which player did what, but something like "how is *detroits defensive rebound/opponents offensive rebound ratio* while ben wallace is playing" is..I finished the program a couple of days ago and started analysing the playoff games (so far I have only games of the first round) I ended up with alot of interesting and stunning statistics: </div> 82games.com keeps track of those type of stats for the regular season. I think there's some value in them if you look at a large sample, and also take into the level of competition played with and against. For instance, its virtually worthless to look at +/- stats for players who only play in garbage time. Can your program show how the stats for 5-man units (both on offense and defense), versus just individual players? That would add some context to the numbers, and provide some more information.
i decided to just do the playoffs because i thought(hoped) garbage time would be limited here. unfortunately i don't have 5 man units but it's something i will most likey work on in the future...
Exactly, I was wondering whether the Horry/Duncan impact stat was based on the impact Horry has on the San Antonio lineup with or without TD, and whether or not Duncan's impact was measured on the difference between SA's performance with him and their performance without him. Alright, that didn't make much sense.
Anyways, back2newbelf, you should keep at it with this program and see how it goes. Try to run it next season and keep us updated. Who knows, there might something important that we can deduce from these kind of stats.
Here are the +/- numbers for San Antonio's 5-man units, and also individual players. Unfortunately, it only shows the top five 5-man units (by minutes played). 5-man units: http://www.82games.com/playoffs/045PSAS2.HTM Individuals: http://www.82games.com/playoffs/045PSAS1.HTM From these numbers, I think the real difference maker for San Antonio in the playoffs has been Ginobili over Barry. When Ginobili is out, the Spurs have struggled. When Ginobili is in, they have been very good. It appears like the Spurs are at their best when Ginobili is on the court with Horry. But it difficult to take too much away from this without knowing who these 5-man units matched up against. If opposing teams went to a weaker lineup when Duncan went out, then that of course would be a huge factor.
i realized that not putting *who does a player play with and against* into the equation would lead to unsatisfying results so i took a different approach which is similar to what 82games does. what i'm gonna do is the following: i'll compare players only if they played with and against the exactly same players. example: to compare McDyess to Rasheed i'll look at: wade Djones Ejones haslem shaq VS billups hamilton prince rasheed Bwallace compared to wade Djones Ejones haslem shaq VS billups hamilton prince mcdyess bwalllace i didn't have time to compare single players yet but here are the 10-man units(if you wanna call it that) for the detroit VS miami series. i thought this alone would be interesting too: wade ejones haslem mourning dooling -- billups hamilton prince rasheed Bwallace 41-56 wade ejones haslem shaq dooling -- billups hamilton prince rasheed Bwallace 19-16 djones ejones haslem shaq butler -- billups hamilton prince rasheed Bwallace 29-30 djones wade ejones shaq mourning -- billups hamilton prince rasheed Bwallace 10-9 DJones wade EJones haslem shaq -- billups hamilton prince rasheed Bwallace 155-137 djones wade ejones shaq mourning -- hamilton rasheed bwallace arroyo hunter 13-5 DJones wade EJones haslem shaq -- billups hamilton prince bwallace campbell 15-18 djones ejones haslem shaq butler -- billups hamilton prince bwallace mcdyess 4-15 the first number is always the points the miami lineup scored, the seconds is the points the detroit lineup scored. you can see that miami's starting five did better than detroit's. i posted only those *10-man units* that produced more than 15 points together(because there are too many to post all)
Yeah alrite, but how do you count stats like 24 points. 15 rebounds and 7 assists, just adding on 5 blocks. It takes hard work from the individual to make those stats. Yeah assists, the other person has to make the basketball for it to count, but who gets the pass there? That player. If you have some lunatic passing that ball, he is more likely to hit you in the head, or send it back into the 15th row then get you the ball. Some line ups do work better then others. But unless you have that superstar in there, you will most likely end up like portland. (sorry every port fan)
