I'm not making this Miller vs. BLANKY in terms of passing ability. You are. I was addressing a specific post by BoobNoMore.
No idea. I don't have any position on whether Miller is helping specific players. Just that he's producing more than Blake in general and has a higher Assist Rate (more of the possessions he uses result in an assist) so I think he's the better point guard.
Fair enough, you addressed it with stats that don't necessarily refute his position, then. You asked me to address the stats and I did. But, again, he stated Blake doesn't create easy opportunities. Would have been very simple, to prove that he does in your eyes, to show that he has assisted on 18 close shots, and 16 dunks, thus creating 34 easy opportunities. You made it a Blake Miller debate by mentioning Miller in your stats. When there was absolutely no need for it. If I argued that Outlaw doesn't score in the 4th quarter, showing me his scoring in the 4th quarter would be effective. Showing me his scoring alongside Webster's scoring in the 4th quarter would be unnecessary, and quite clearly an attempt at furthering some sort of Outlaw-Webster debate.
Correct. I added more than mere observation, however. Carry on with your imaginary debate about Miller vs. BLANKY. No need for us to communicate any further on it, since I agree with you.
I don't know EXACTLY what's being argued in this latest thread tangent (as I can't see, and don't want to see, all the posts), but it appears to be a disagreement on what consists an easy basket. Personally, I don't care how you define "easy". Is a contested inside shot "easier" then a wide open 15-footer? Depends on the player. Also, if the player receiving the pass makes a move to beat his defender, no assist is awarded. So, in a way, ALL assisted baskets are, by definition, "easy". I'll leave the arguing of semantics to those with nothing better to do. All I really care about is MADE shots. You only get an assist when the player you pass to makes the shot. And the fact is, and it's 100% backed-up by all the stats, Andre Miller has more assists in fewer minutes than Steve Blake. So, by that simple definition, Andre Miller creates more scoring opportunities for his teammates than Steve Blake does - and again, that's all I care about. This is also shown by the fact that Andre Miller has an AST% (28.6%) that is 36% higher than Steve Blake's (21.0%). Clearly, Andre Miller does a MUCH better job creating scoring opportunities for his teammates than Steve Blake. I'm not sure why anyone would argue that that's a bad thing. Miller also creates more scoring opportunites for himself - which is also valuable given this team's difficulty scoring as of late. But, still there are those who continue to defend Blake, and his 8.9 PER, poor shooting, lack of scoring, fewer assists per minute, inability to get to the FT line, poor FT shooting percentage and general subpar play. All my pro-Miller arguments are based on actual production - and both Miller's individual and team production are better than Blake's. To me, it seems obvious that the more productive player should get more playing time. I just can't figure out why some people stubbornly insist that the less productive player should be rewarded with more playing time. I guess there's just some things I'll never understand. BNM
PER is by definition a per 48 minute stat (or per minute, whatever). As for Miller, I think he's tons better than Blake and Miller should probably be in the game at the end of Q4. That said, he is doing it against 2nd string for the most part and I think that's by design. Scott Skiles wanted either Gordon or Nocioni as 6th man so he had some offense from the 2nd unit - maybe that's Nate's line of thinking here.
No it's not. It does adjust for PT and pace, but it is not a per 48 minute stat - as defined by it's creator, John Hollinger. I was merely pointing out the reason behind the discrepancy in PER values between those reported at 82games.com (PER/48) and those reported by Hollinger, basketball-reference.com, etc. (PER). This is why I like Miller as the starter and Bayless backing up both guard spots forf 25 - 30 MPG and Blake getting the table scraps (10 - 12 MPG). Miller is definitely better than Blake. Therefore, I think he should start (and finish). Bayless can generate his own offense - which we need on the second unit with Outlaw and Rudy out. BNM
See the 1/Min at the front? uPER = (1 / MP) * [ 3P + (2/3) * AST + (2 - factor * (team_AST / team_FG)) * FG + (FT *0.5 * (1 + (1 - (team_AST / team_FG)) + (2/3) * (team_AST / team_FG))) - VOP * TOV - VOP * DRB% * (FGA - FG) - VOP * 0.44 * (0.44 + (0.56 * DRB%)) * (FTA - FT) + VOP * (1 - DRB%) * (TRB - ORB) + VOP * DRB% * ORB + VOP * STL + VOP * DRB% * BLK - PF * ((lg_FT / lg_PF) - 0.44 * (lg_FTA / lg_PF) * VOP) ]
Yes, as I said previously, it accounts for minutes played, but it is NOT a per 48 minute stat - like the PER/48 reported by 82games.com. My point was, and is, the PER/48 values reported by 82games.com are higher than the straight PER reported by Hollinger and other sites. That's all. The statement was, and is, correct. BNM
I don't think you're right about WHY the values are different. Hollinger's are normalized for minutes played while I don't see a definition of 82games.com's PER as being normalized at a different rate... they say "PER*", not "PER/48" and I can't find a definition of what "PER*" is. To figure PER, there are league-wide stats that change every day, along with the individual player's stats. I would imagine that the sources just update their league-wide stats (like VOP) differently. EDIT: This page indicates they simply use Hollinger's PER equation, which seems to support my speculation. That page is OLD, though, so maybe they've changed things? 82games doesn't do a very good job of documenting what they have where. Ed O.
The table headings in their "by position" section are clearly labeled as: Player 48-Minute Production by Position Opponent Counterpart 48-Minute Production Opponent Counterpart 48-Minute Production These are the stats Denny Crane originally referenced and are clearly normalized to 48 minutes. That explains why they are higher than the straight Hollinger PER rating for the same player. BNM
That still doesn't make any sense. PER is normalized per minute. If those were simply PER-per-48-minutes numbers, then on 82games.com they would be about 48 times what they actually are. I would imagine that it's listed as "PER*" because it's NOT a "per 48" stat. Ed O.
Then why do they label the tables as "48-minute production" and why are their values consistently higher than Hollinger's PER ratings? And no, PER is not normalized per minute. uPER (micro-PER) is the per minute version of PER. BNM
Are you ever going to address the actual assist statistics I gave earlier in the thread about BLANKY'S assist and "easy" baskets?
Are you referring to Hollinger's ratings at ESPN.com? If so the reason the per-48 stats are higher is because the table at ESPN.com on each player's profile page lists per-40 minute production. Secondly, PER is normalized ... it doesn't matter if it's per minute or per 48 minutes, because the PER number doesn't have any inherent value other than measuring a player's efficiency while they are on the court. Thus a player who only plays ten minutes a game can have an identical PER to a player who plays 38 minutes a night (assuming they are equally efficient)
Because it was the only reasonable place to put PER on the site, and they put the asterisk after PER to indicate that it's not per 48, because that makes no sense given PER is already a per-minute statistic. They didn't explain why they put an asterisk there, but it's a reasonable explanation. "uPER" is a sub-component of PER that Basketball-Reference.com uses. It is unadjusted PER, which does not take into account pace. Hollinger has not, to my knowledge, made that distinction. I'm quoting from Hollinger's 2003-2004 Pro Basketball Prospectus: Page 10, Column 3: "However, three more adjustments are necessary before the final PER is available. First, one must adjust for minutes. The PER is a per-minute ranking so after adding all the good and subtracting all the bad detailed over the last few pages, we need to divide the total by the number of minutes played." I added that emphasis. Also, if you look at the equation (which Denny posted), and if you read what I posted before, (1/minutes played) is a multiplier for PER, so it IS a per minute basis. If you want a reasonable guess as to why the numbers are different, please refer to my earlier post about league-wide statistics that are critical to PER determination. That explanation is not very precise and might not be accurate, but your assertion that PER is not a per-minute statistic is not one that is based on anything I've ever seen or read. Do you have any sources to back up your claim? Ed O.
OK, I may have been wrong about the way 82games.com reports PER, but the fact remains their PER ratings are different then the official Hollinger PER ratings. When I started this thread, I used Hollinger's official, current PER ratings for Blake (8.9), Miller (15.1) and Bayless (17.9). Those numbers were (and still are) current and accurate. I frankly have no idea how 82games calculates (or miscalculates) their PER ratings. All I know is they are not current, nor accurate. I prefer to go straight to the source and quote the official Hollinger PER ratings. Which, as I stated originally, show that both Miller and Bayless are MUCH more productive than Steve Blake, and if PT is based on production, they deserve more and Blake deserves less. I stand by the stats I quoted and the conclusions I reached based on those accurate stats. If somebody wants to contact 82games.com and ask them the source of their error, be my guest. The current Hollinger PER ratings are easily available. So, I will continue to use them BNM
