What is more important in fantasy drafting? (1 Viewer)

[Lott's Fingertip]

I see a lot of stats thrown around when talking about players, and one thing that bothers me a bit is when people quote season stats for a player that played in less than 16 games against a player that played in 16 games.

Is 1000 yards in 16 games more valuable than 900 yards in 12 games, for example. Is it better to draft a guy that'll get you 12 ppg for 12 games, or a guy that'll get you 9-10 ppg for 16 games?

Obviously there are different situations for different players to take into account, but IN GENERAL what do you weigh more heavily when you are drafting?

PPG stats

or

Full season stats

 

[MAC_32]

Points per game, I would rather have the dude that produces 900 yds and 8 TD's in 10 games than the dude that puts up 1200 yds and 10 TD's in 16. I am confident I can find more than 300 yds and 2 TD's in those other 6 games.

 

[Couch Potato]

PPG, because you get backup points when he doesn't play, not zero points.

If I have a guy who puts up 15 PPG for 12 games (180 FP), and have to plug in a guy with 10 PPG for the other 4 games, I get 220 FP. That's more valuable to me than a guy who gets 210 FP over 16 games.

The 180 FP guy > the 210 FP guy. This assumes an adequate bench of course.

 

[JamesTheScot]

I'm not sure what you are asking.

Do I prefer drafting a guy who has great per game scoring potential but will likely miss some games over a guy who won't score as much per game but has very low injury risk? As long as he doesn't miss too many games, yes. But it's not cut and dried. If the guy that only had 10 games is a risk to only play in 4 games this season, I might take the 16 game guy so I can devote a draft spot to depth at some other position. I probably have to take a RB3 earlier than I normally would if I take Westbrook so I might pass on him and take someone who hasn't missed much time and can be taken later. By doing that, I can wait later to take a backup and I can pick a higher rated player earlier at another position.

Do I consider a player's per game stats for the next year's projections? Of course. Who wouldn't?

 

[King of the Wolfies]

season stats are meaningless in many ways. I look for players that are consistent. Players who miss a lot of games or are hit or miss means that you need backups at near the same quality or you are going to lose the games they miss or do poorky in.

 

[Riffraff]

Depends on how many games you are talking about for PPG. If they rock the world for 2 games and are out for the year, the PPG is useless.

 

[stevegamer]

80% PPG - 20% season totals

Extreme outliers in PPG are due to few games, and those guys help for their game or two. Last year, a DL1 in PPG in one of my leagues played in one game. He's uttelry useless, since you needed to know when to play him - and since he got DNP's other weeks (and I think got cut) he's unplayable.

Guys with elite season totals will usually also be elite in PPG, so using season totals to balance out the vagaries in PPG makes sense.

 

[Jason Wood]

Per game numbers are more relevant when trying to compare players, situations, different seasons, etc...

Of course, you should also adjust to league norms for the metrics, too, but that's a process most don't/can't bother with.

But I think ignoring some assumption of injury risk is foolhardy. It counts too, just not as much as a normalized, per game output.

 

[Lott's Fingertip]

Are you asking in terms of starts during the season or in terms of predictive value for an upcoming draft?

Predictive value for an upcoming draft.A good example would be Steve Smith last year... we knew he was missing 2 games due to suspension.Other examples include Marshall last year, Lynch this year, Brandon Jacobs every year.Guys that you either know are missing games, or assume will miss games due to injury, but who also put up very good stats when they play.Players that seem injury prone are probably the best examples, because you can think they will get injured, but they may play 16 games. Like AJ going into last year.

 

[Lott's Fingertip]

Per game numbers are more relevant when trying to compare players, situations, different seasons, etc...Of course, you should also adjust to league norms for the metrics, too, but that's a process most don't/can't bother with.But I think ignoring some assumption of injury risk is foolhardy. It counts too, just not as much as a normalized, per game output.

This about wraps up my view very nicely.

 

[lyon812]

Yes, take the 900yds in 12g over the 1000yds in 16g. That's just clearly better production. You can't predict injury and shouldn't obsess over it when drafting unless a guy has clear indicators that he'll miss far more than a couple of games (for example, recurrent hamstring problems, or if his name is Deion Branch). PPG is never completely consistent. Every player has ups and downs across the course of the season. In fact, many of the studs are far more mercuric from week-to-week than people realize. It isn't like the 12ppg guy is scoring 23, 1, 23, 1, 23, 1, and the 10ppg guy is scoring 10, 10, 10, 10, 10. There's much more variance.

The only caveat to this is dependent upon your league's scoring system, primarily IDP. If some players score more because of big/rare plays--let's say sacks, and John Abraham--then I feel it's good to pair them with a more consistent person (tackles) to help stabilize your overall points per game. I think having only sack artists or only tackle guys will limit you.

 

[perry147]

Each player is different but for the most part for me PPG is a better.

 

[Tanner9919]

season stats are meaningless in many ways. I look for players that are consistent. Players who miss a lot of games or are hit or miss means that you need backups at near the same quality or you are going to lose the games they miss or do poorky in.

conversely, you also don't players like Jamal Lewis, who ran for 200+ yards in two games ( each) against Cleveland and Cincy a few years ago, but avg'd much less against the rest of the NFL..I'll take the season-long stats over PPG any day of the week.Portis gets great season-long stats.he's consistently good, but not great..and, he plays in 16 games nearly every season..so you get a consistent, long slow burn from him, vs. a flame on a candle, like Gore or Westbrook...a PPG flop like Westbrook is great for ,what, 6-7 weeks, then he's a game time decision/doubtful type of player the remainder of the season...? give me Portis over Westbrook/Gore every time..Slaton wasn't a PPG stud as such, but he was a long,slow,consistent burner for the entire 2008 season.

 

[Snotbubbles]

season stats are meaningless in many ways. I look for players that are consistent. Players who miss a lot of games or are hit or miss means that you need backups at near the same quality or you are going to lose the games they miss or do poorky in.

conversely, you also don't players like Jamal Lewis, who ran for 200+ yards in two games ( each) against Cleveland and Cincy a few years ago, but avg'd much less against the rest of the NFL..I'll take the season-long stats over PPG any day of the week.Portis gets great season-long stats.he's consistently good, but not great..and, he plays in 16 games nearly every season..so you get a consistent, long slow burn from him, vs. a flame on a candle, like Gore or Westbrook...a PPG flop like Westbrook is great for ,what, 6-7 weeks, then he's a game time decision/doubtful type of player the remainder of the season...? give me Portis over Westbrook/Gore every time..Slaton wasn't a PPG stud as such, but he was a long,slow,consistent burner for the entire 2008 season.

I prefer consistent players. I remember playing in a H2H league and I ended up with the most total points but missed the playoffs because I had inconsistent players (Corey Dillon comes to mind). Score 40 points one week and then go 2 or 3 getting me 3 points. Same goes with an injured player. You get good scoring for the weeks he's in. But then what do you do if he's a GTD. It's automatically assumed that when Player X misses a game you insert Player Y, but it doesn't always happen that way. I remember one year I had Brian Westbrook and he got scratched from a MNF game in the pre-game warmup so I never had the opportunity to start a bench guy.

 

[bicycle_seat_sniffer]

I for one have always been skeptical about people trotting out the PPG card, people often use it to justify grabbing oft-injured players

 

[lyon812]

Portis gets great season-long stats.he's consistently good, but not great..and, he plays in 16 games nearly every season..so you get a consistent, long slow burn from him, vs. a flame on a candle, like Gore or Westbrook...a PPG flop like Westbrook is great for ,what, 6-7 weeks, then he's a game time decision/doubtful type of player the remainder of the season...? give me Portis over Westbrook/Gore every time..

I don't mean to pick on you specifically, but this is an excellent time to examine the conception of consistency and demonstrate the danger of misperception of things such as "Portis is consistent and Westbrook isn't". This isn't an uncommon view by any stretch. The perceptions of certain players are greatly skewed. I intend to demonstrate that Portis, Westbrook, and Gore are essentially equal. I decided to analyze multiple years of data in order to establish an average PPG, and their variance from them by X amount of points. This was a three-step process.First, I took the last four years of data (2005-2008), and sorted the point values. Second, I grouped them by deviation from the mean, so that we'd know how many of the individual games were within a certain range. Third, I split them into hi (meaning good, or above) and lo (meaning bad, or below). Thus, in the chart below, you'll see that Portis had 6 hi and 8 low within 2 points of average. That means he had 8 games from 12.6-14.5, and 6 games from 14.7-16.6. The results were this:

Portis		Westbrook	 Gore	Years						   05-08		 05-08		 05-08total points					817.7		 923.1		 742.6	games						   56			56			59	average PPG					 14.6		  16.5		  12.6	 games within 2 points of avg	14(25.00%)	15(26.79%)	13(22.03%)						 hi	  6			 5			 3							 lo	  8			10			10	games within 5 points of avg	27(48.21%)	27(48.21%)	30(50.85%)						 hi	 10			11			 8							 lo	 17			16			22	games within 8 points of avg	40(71.43%)	34(60.71%)	40(67.80%)						 hi	 14			13			14							 lo	 26			21			26	games > 8 pts difference	16(28.57%)	22(39.29%)	19(32.20%)						 hi	 10			10			11							 lo	  6			12			 8	Total games above average	   24(42.86%)	23(41.07%)	25(42.37%)Total games below average	   32(57.14%)	33(58.93%)	34(57.63%)

The first thing you notice is that Westbrook has a high incidence of flameout games--12 that are 8 points or more below average. However, 6 of these 12 happened in 2008, which skews the results considerably; it's an outlier season (conversely, only 2 of Portis' were in 2008). Likewise, I'll bump Gore up a season so he won't be penalized for his rookie year compared to the others in the prime of their careers. This will provide a more representative look at their careers when things were more equal on an age/situation/injury level. We'll reduce the sample to 3 years, but if I had more than a couple of hours, I would have considered extending it to all seasons instead. Someone else can do the legwork for that.

Portis		Westbrook	 Gore	   Years						   05-07		 05-07		 06-08	   games						   40			42			45		   average PPG					 14.8		  16.8		  14.5		games within 2 points of avg	 4(10.00%)	12(28.57%)	 6(13.33%)							 hi	  0			 5			 3		   						 lo	  4			 7			 3		   games within 5 points of avg	20(50.00%)	25(59.52%)	18(40.00%)   						 hi	  8			11			 7			  						 lo	 12			14			11			  games within 8 points of avg	28(70.00%)	30(71.43%)	32(71.11%)	 						 hi	 10			12			13			 						 lo	 18			18			19			games > 8 pts difference	12(30.00%)	12(28.57%)	13(28.89%)								 hi	  8			 6			 7		  						 lo	  4			 6			 6			  Total games above average	   18(45.00%)	18(42.86%)	20(44.44%)   Total games below average	   30(55.00%)	24(57.14%)	25(55.55%)

As you can see, there's virtually no difference between them. In fact, Westbrook has the greatest amount of consistency +/- 5 points of his average PPG. Both Portis and Gore "catch up" to Westbrook in the 5-8 point variance range, as you can see by their ~70% numbers. However, now that we've happened to pick three very similar backs numerically, we should examine some other long-term dominant backs. To keep it contemporary, I'm going to include three other backs with runs of recent success. Looking at the Historical Data Dominator from 05-07, we find this ranking:1. Tomlinson: 1058.60 points2. Johnson: 767.70 points3. Westbrook: 705.30 points4. Jackson: 694.40 points5. James: 626.70 pointsI'll choose Tomlinson--who is truly dominant and an outlier in himself--Jackson and James. I don't consider Larry Johnson because his two years were really looking like the exception more than representative of a career of excellence. FYI, during that time period Portis ranks 8, and Gore 9, after Alexander and Parker. I considered doing them, but I'll run out of space if I add too many. =p Btw, I excluded playoffs.

Portis		Westbrook	 Gore		 Tomlinson	 Jackson	  JamesYears						   05-07		 05-07		 06-08		05-07		 05-07		05-07games						   40			42			45		   48			43		   47average PPG					 14.8		  16.8		  14.5		 22.1 (!!)	 16.1		 13.3games within 2 points of avg	 4(10.00%)	12(28.57%)	 6(13.33%)	4(8.33%)	  7(16.28%)   11(23.40%)				   						 hi	  0			 5			 3			1			 7			3	 						 lo	  4			 7			 3			3			 0			8	 games within 5 points of avg	20(50.00%)	25(59.52%)	18(40.00%)   10(20.83%)	20(46.51%)   28(59.57%)												 hi	  8			11			 7			3			13		   11								 lo	 12			14			11			7			 7		   17		games within 8 points of avg	28(70.00%)	30(71.43%)	32(71.11%)   19(39.58%)	28(65.21%)   42(89.36%)												 hi	 10			12			13			7			16		   17	   						 lo	 18			18			19		   12			12		   25		games > 8 pts difference	12(30.00%)	12(28.57%)	13(28.89%)   29(60.42%)	15(34.88%)	5(10.64%)													 hi	  8			 6			 7		   13			 7			3	   						 lo	  4			 6			 6		   16			 8			2	  Total games above average	   18(45.00%)	18(42.86%)	20(44.44%)   20(41.67%)	23(53.49%)   20(42.55%)					   Total games below average	   30(55.00%)	24(57.14%)	25(55.55%)   28(58.33%)	20(46.51%)   27(57.45%)

Here's where we start to see some variance. Tomlinson is such an extreme that he really has no comparison. An amazing 60% of his games are more than 8 points above or below his average--and he's almost equally split. That being said, 8 points below his average is 14.1, which would put him right around an average game for Gore, Portis, and James. In short, Tomlinson was worth owning because even his crappy days were good. He only had 5 outings below 8 fantasy points, which is almost identical to everyone else (Portis 4, Westbrook 6, Gore 6, Jackson 9). James had 10 below 9 points, but his average was significantly lower as it is. Let's take one more factor into account. Let's look at all variances greater than 5 points.

Portis		Westbrook	 Gore		 Tomlinson	 Jackson	  JamesYears						   05-07		 05-07		 06-08		05-07		 05-07		05-07games						   40			42			45		   48			43		   47average PPG					 14.8		  16.8		  14.5		 22.1 (!!)	 16.1		 13.3games within 5 points of avg	20(50.00%)	25(59.52%)	18(40.00%)   10(20.83%)	20(46.51%)   28(59.57%)												 hi	  8			11			 7			3			13		   11								 lo	 12			14			11			7			 7		   17		games > 5 pts difference	20(50.00%)	17(40.48%)	27(60.00%)   38(79.17%)	23(53.49%)   19(40.43%)													 hi	 10			 7			13		   21			10			9	   						 lo	  4			10			14		   17			13		   10

Now some final rankings:Ranked in order of extreme consistency (>= 2 points of deviation)Westbrook 28.57%James 23.40%Jackson 16.28%Gore 13.33%Portis 10.00%Tomlinson 8.33%Ranked in order of moderate consistency (>= 5 points of deviation)James 59.57%Westbrook 59.52%Portis 50.00%Jackson 46.51%Gore 40.00%Tomlinson 20.83%Ranked in order of rough consistency (>= 8 points of deviation)James 89.36%Westbrook 71.43%Gore 71.11%Portis 70.00%Jackson 65.12%Tomlinson 39.58%Ranked in order of inconsistency (>8 points of deviation)Tomlinson 60.42%Jackson 34.88%Portis 30.00%Gore 28.89%Westbrook 28.57%James 10.64%Voila.Westbrook, for all his struggles last year, has been a very consistent elite back. Even within the 8 point range, Jackson isn't far off the mark. He's only 5% behind, which all things considered isn't too bad, especially because of the 65.12% of his games within 8 points of deviance, he is the ONLY back that has more HI than LO within 8 points of his mean. 16 high and 12 low. This means that although his inconsistency is greater, it frequently works in the owner's favor, not against it. In summary, you can roughly lump Portis, Gore, Westbrook, and Jackson in the same realm of consistency. This suggests that people who are worried about PPG variances are doing so needlessly.ETA: Think I fixed all of my code and cleared up some confusing bits. However, I created this on a 1440x900 desktop. Thus, my apologies if any of you have trouble reading it at lower resolutions.

 

[Lott's Fingertip]

...Voila.Westbrook, for all his struggles last year, has been a very consistent elite back. Even within the >=8 point range, Jackson isn't far off the mark. He's only 5% behind, which all things considered isn't too bad, especially because of the 65.12% of his games within 8 points of deviance, he is the ONLY back that has more HI than LO within 8 points of his mean. 16 high and 12 low. This means that although his inconsistency is greater, it frequently works in the owner's favor, not against it. In summary, you can roughly lump Portis, Gore, Westbrook, and Jackson in the same realm of consistency. This suggests that people who are worried about PPG variances are doing so needlessly.ETA: Think I fixed all of my code.

Nice work!

 

[CBower4545]

This is an excellent thread. I think as long as you have a large enough sample size from a situation that is somewhat comparable to their current one, PPG is much more relevant in deciding a baseline for your projections. I wouldn't include rookie seasons where players were hardly used and situations like that. Take S. Moss for example I would take his time as a #1 WR in Washington to get his PPG and baseline for stats. Since he has 60 games in Washington as a #1 WR it seems the most comparable to his current situation.

 

[lyon812]

This is an excellent thread. I think as long as you have a large enough sample size from a situation that is somewhat comparable to their current one, PPG is much more relevant in deciding a baseline for your projections. I wouldn't include rookie seasons where players were hardly used and situations like that. Take S. Moss for example I would take his time as a #1 WR in Washington to get his PPG and baseline for stats. Since he has 60 games in Washington as a #1 WR it seems the most comparable to his current situation.

I'm considering doing the same study for QBs, WRs, and TEs when I get the chance. I think it'd be interesting to see if different positions have greater variance. My initial inclination is to believe, based on my prior assumptions and the data so far, that most players fall within a certain acceptable range of consistency, and that only the extremely talented deviate significantly (The Tomlinson Effect).

 

[puckalicious]

This is an excellent thread. I think as long as you have a large enough sample size from a situation that is somewhat comparable to their current one, PPG is much more relevant in deciding a baseline for your projections. I wouldn't include rookie seasons where players were hardly used and situations like that. Take S. Moss for example I would take his time as a #1 WR in Washington to get his PPG and baseline for stats. Since he has 60 games in Washington as a #1 WR it seems the most comparable to his current situation.

I'm considering doing the same study for QBs, WRs, and TEs when I get the chance. I think it'd be interesting to see if different positions have greater variance. My initial inclination is to believe, based on my prior assumptions and the data so far, that most players fall within a certain acceptable range of consistency, and that only the extremely talented deviate significantly (The Tomlinson Effect).

Wouldn't be a lot easier to just show the Average and Standard Deviation for each player? Not sure what the point of all the "games > x pts above average" is meant to convey. Another way to look at the OP topic is to use MPP (median points per game) instead of average. This takes out some of the skew that can happen to the average from a few outlier performances. In general I do like the per-game performance of players when making projections, but injury risk (or other reasons for missing games) must also be weighed. The easiest way to do this is to compare season results.....Those of you here pimping the PPG aspect surely view SJax as a top 5 draft pick every year then, right? In reality you are downgrading him because you think he'll miss some time and draft accordingly, which really means you do take season results into account. As with most things in life (and FF) it's about balance.

 

[lyon812]

This is an excellent thread. I think as long as you have a large enough sample size from a situation that is somewhat comparable to their current one, PPG is much more relevant in deciding a baseline for your projections. I wouldn't include rookie seasons where players were hardly used and situations like that. Take S. Moss for example I would take his time as a #1 WR in Washington to get his PPG and baseline for stats. Since he has 60 games in Washington as a #1 WR it seems the most comparable to his current situation.

I'm considering doing the same study for QBs, WRs, and TEs when I get the chance. I think it'd be interesting to see if different positions have greater variance. My initial inclination is to believe, based on my prior assumptions and the data so far, that most players fall within a certain acceptable range of consistency, and that only the extremely talented deviate significantly (The Tomlinson Effect).

Wouldn't be a lot easier to just show the Average and Standard Deviation for each player? Not sure what the point of all the "games > x pts above average" is meant to convey.

I felt it valuable to know how often a player performs within a certain range of his average. This gives a better indication of varying degrees of consistency. For example, if you looked at Portis, Westbrook, and Gore and said, "how often are they within 8 points of their average", they'd be identical. However, the clear superiority possessed by Westbrook in performing within 2 points of his average (over twice as much) is quite significant. This, this adequately examines the larger picture(by which we can see that there's an inherent degree of inconsistency simply by being an NFL running back) and the more detailed picture, such as:1. Westbrook being more consistent within 2 points of his average2. Jackson's inconsistency having a more positive effect than other backs, not only within 8 points of his average, but overall3. Tomlinson's extreme inconsistency, suggesting that uberstud backs will be the most inconsistent of any4. The value of players like James who only very rarely explode for a lot of points and thus might be seen as more consistent, when in actuality their % of games above average are right in line with everyone elseThus, no, I feel that your suggestion of just using average and standard deviation would actually be detrimental and not illuminate nearly as many interesting aspects of this.

 

[pollardsvision]

I think this is a tough question b/c I try not to predict injuries. This question makes sense in retrospect, but I don't think it does much good beforehand.

In retrospect, sure, I'd rather have a guy who will be great but miss a few games over a guy who will be mediocre, but play in every game, but there are very few times that you can predict that.

If I have a reason to believe that a player will get hurt and miss a significant portion of the season, I'm probably just not going to draft them (unless everybody else thinks they'll get/be hurt and there's good value).

If guys have an injury history or will miss a portion of the beginning of the season (the only time we can predict it) for injury or suspension, then I downgrade them within their own tier. I'll draft them after all of the other players who get a similar PPG are gone, but ahead of guys who won't be gamechangers.

So, my short answer is PPG b/c predicting who will miss games, how many, and when is usually impossible.

 

[puckalicious]

This is an excellent thread. I think as long as you have a large enough sample size from a situation that is somewhat comparable to their current one, PPG is much more relevant in deciding a baseline for your projections. I wouldn't include rookie seasons where players were hardly used and situations like that. Take S. Moss for example I would take his time as a #1 WR in Washington to get his PPG and baseline for stats. Since he has 60 games in Washington as a #1 WR it seems the most comparable to his current situation.

I'm considering doing the same study for QBs, WRs, and TEs when I get the chance. I think it'd be interesting to see if different positions have greater variance. My initial inclination is to believe, based on my prior assumptions and the data so far, that most players fall within a certain acceptable range of consistency, and that only the extremely talented deviate significantly (The Tomlinson Effect).

Wouldn't be a lot easier to just show the Average and Standard Deviation for each player? Not sure what the point of all the "games > x pts above average" is meant to convey.

I felt it valuable to know how often a player performs within a certain range of his average. This gives a better indication of varying degrees of consistency. For example, if you looked at Portis, Westbrook, and Gore and said, "how often are they within 8 points of their average", they'd be identical. However, the clear superiority possessed by Westbrook in performing within 2 points of his average (over twice as much) is quite significant. This, this adequately examines the larger picture(by which we can see that there's an inherent degree of inconsistency simply by being an NFL running back) and the more detailed picture, such as:1. Westbrook being more consistent within 2 points of his average2. Jackson's inconsistency having a more positive effect than other backs, not only within 8 points of his average, but overall3. Tomlinson's extreme inconsistency, suggesting that uberstud backs will be the most inconsistent of any4. The value of players like James who only very rarely explode for a lot of points and thus might be seen as more consistent, when in actuality their % of games above average are right in line with everyone elseThus, no, I feel that your suggestion of just using average and standard deviation would actually be detrimental and not illuminate nearly as many interesting aspects of this.

I see your point. I still feel that using 2 or 8 or any predetermined # could bias the results. Using the average and standard deviation as I first mentioned may also not be prudent since I doubt this follows a normal distribution.A good estimate of expected production will be a range about the median value. Normally I would look at the 1st and 3rd quartiles, but in terms of standard deviation it is more useful to consider the 16th percentile (instead of 1st quartile which = 25th percentile) and the 84th percentile (instead of 3rd quartile which = 75th percentile). The spread of data from the 16th to the 84th percentile represents the same population size as +/- 1 standard deviation, or roughly 68% of the total. This basically tells us what kind of point spread to expect for the majority of games, and means that there were an equal number of games scored below the lower bound as there were above the upper bound.Let's look at the game data over the years you show. I've compared the 16th, 50th (median) and the 84th percentile values for each RB. And since each RB is in a unique situation (not all running behind the same line during the same games) it will be useful to express their scoring spread as a fraction of the median. This allows us to compare the RB's variation to each other directly.

Player	 p16   p50   p84   VarPortis	 8.6  13.5  23.7  0.56Westbrook 10.0  15.4  22.2  0.40Gore	   7.4  12.5  22.5  0.60LT		 9.8  19.2  37.3  0.72SJax	   7.1  16.5  23.9  0.51James	  7.6  12.1  19.2  0.48

This tells us that LT's scoring varies the most, and Westbrook's the least. But when you compare the level of scoring you'll take LT since his greater variation happens above and beyond Westbrook's ceiling.I think we've drifted way off topic, but my general take is that if you're going to focus on PPG, use the median value, not the average.