2. I see three clearly poor games by Game score, another average one I guess. Again with that kind of sample size I'm not sure you should be making any definitive conclusions. By having a super poor performance he just ensured a loss, by the way. The kind of loss Boston had recently by having a player shoot 0-13, where going 5-13 would have been different. If over a span of various Finals/playoff series this trend shows up, maybe you would have more to stand on. Giving credit for a team stat is something I'm still dubious about.

UB- In spite of his dislike for LA, he's the type of writer that reads too much into short term performance and intangibles. This is a nuance I just wanted to note.

The points I was trying to make with the Starks example were:
1.When remembering (or rewarding) a player's performance in a 6-7 game series people tend to attach, whether rightly or wrongly, greater significance to the last few games and often the end of those games also. Starks 1994 Finals is a clear example of both.
2.The average statistics for a series in isolation do not always give a good overview of how the player contributed to each game. Starks 1994 Finals is an extreme example of this. Do you really think that Starks had a terrible series? He had a bimodal series for sure, with 5 positive contribution games in the good-very good category and 2 negative contribution games in the historically bad category. But his 27 points on 9-18 game 6 only affects the outcome of one game and his 2-18 game 7 only affects the outcome of one game. I'm sure he'd love to average that performance over two games and win the series in game 7. Averages alone don't tell the story of a series.

Regarding crucial/important games, I think its an interesting idea but you'll see from my contribution in the previous thread that I'm not committed to the truth of the concept and am trying to stir up some different ways to think about equal versus unequal importance of playoff games.

True? Who knows.

But my point was simply that, to back-up the mere possibility that playing with Bryant has improved Gasol... To quote, "there can be a PERCEPTION that one player inspires others to greatness."

Thanks for the series of analysis in the playoffs though, Neil.

I'm not sure what the Starks example proved. He did have a terrible series, it doesn't seem like a huge surprise that he failed in some games to that degree.

"who carried the team through rough periods, the timing of key (both positive and negative) contributions that were made which influenced the outcomes of games and series etc. It is why I'm a big proponent of looking at distributions in addition to measures of central tendency. "

Nothing Neil has gone over has shed more light on over and under playoff performers. Ray Allen had a case for Finals MVP in 2008, because he had some decent numbers ( Game Score).

http://www.basketball-reference.com/blog/?p=6132

I just did a quick calculation of LeBron's game score against the #1 defense in the NBA, playoffs games 5-7 the past 3 seasons. Guess what I got? 22.12 (clink the link for his average crucial game score). This topic has been addressed already, and it is just a vague theory as it is that games 1-4 don't matter to an individual's legacy. Most important games in a 7 game series are team stats, dependent on teammates. You can't even get to that position without performing in games 1-4. There's an intrinsic flaw in your ideas, imo.

Neil has sufficiently eliminated a lot of myths about the league. I'm a proponent of large sample sizes, or at least a series breakdown. "The advanced measures such as ORtg, TS%, eFG%, AST% and Game Score all paint the same picture." They already have painted a clear picture in recent blog posts here. Nor should we let players take Games 1-3 off. It is just a hypothesis that we should believe in "crucial" playoff games as it is.

"Also, don't we get a better picture of Starks series performance by breaking it down distributionally (though I've presented it here with some aggregation and averaging used) rather than just his overall series averages?"

You're also ignoring just the random variability of performance. On average in crucial game situations, the same over-performers in the playoffs will usually show up. I think this is mostly a waste of your time but I'll still address this anyway. You forgot about players that under perform against bad defenses as well. Playing terrible against a porous Phoenix Suns defense in 2007 should also go against your resume.

The reason this came to mind is that, based on observation, I thought that Ray Allen had a legitimate case for Finals MVP in 2008. If I've understood your Game Importance post correctly the most important games in that series turned out to be Game 6 (Home up 3-2) and Game 4 (Away up 2-1). In terms of GameScore, Allen put up 25.8 and 20.5 in these games while Pierce put up 16.6 and 15.1.

As you mentioned in the original post it would also be great to see the importance-weighted Finals performance (with the game importance probabilities as the basis of a weighting scheme) to compare to Finals MVP. A low weight would be assigned to the least important game in that series, which turned out to be Game 5 (Away up 3-1), in which Allen had his worst GameScore of the series at 6.8 and Pierce had by far his best at 27.2. This would be negated to some extent by Game 3, the second least important, featuring Pierce's clear worst (-3.1) up against a good Allen game (21.9).

I also think that its the reason Neil has taken the time to devote a number of posts to defining "crucial" games, evaluating player performance in crucial games and determining from a probability based standpoint the influence winning each game has on winning the series. If one believes that every recorded statistical contribution a player makes is equally valuable regardless of game, score and clock there would be no motivation to define any game or contribution to a game as more or less "crucial" than another.

Also, I'm not referring only to evaluating and summarising a Finals performance numerically, but also understanding the way that it is both recalled and perceived. John Starks '94 Finals performance, particularly in Game 7, was brought up a few weeks back in a post on crucial games:

http://www.basketball-reference.com/blog/?p=5656

Here is his game log for the year:

http://www.basketball-reference.com/players/s/starkjo01/gamelog/1994/

Sticking to traditional statistics for the moment, do John Starks' series averages of 17.7ppg,5.9apg,3.1rpg on 36.8% shooting provide a good summary of his overall Finals performance? In Games 2 through 6, in which NY went 3-2, Starks averaged 21ppg, 7.2apg, 3.6rpg on 48.6% shooting. These are outstanding contributions considering the Knicks averaged only 89.2ppg during those games on 42.4% FG shooting. I would still suppose that most remembered moment is Starks' 3 being blocked by Olajuwon at the end of Game 6.

In Games 1 and 7, both very winnable games, Starks shot 3-18 and 2-18 respectively and Game 7 burnt a lasting image in viewer's minds. In these two games he averaged 9.5ppg, 2.5apg, 2rpg on 13.9% shooting. Wow. Did this dude just did this? Flat out terrible. The advanced measures such as ORtg, TS%, eFG%, AST% and Game Score all paint the same picture.

Moreso than Starks' series averages or his great stretch of basketball in Games 2-6, it was his 3 pointer getting blocked in Game 6 and his horrid Game 7 performance that came to define his main contribution to the series as people reflect on it today. Is that entirely fair? No, but in close 6-7 game series the crucial game moments and performances often seem to carry added weight. Also, don't we get a better picture of Starks series performance by breaking it down distributionally (though I've presented it here with some aggregation and averaging used) rather than just his overall series averages?