since I brought stats into this, here’s my response. I like stats for two reasons: 1) human observation is unreliable, especially when it’s coupled with emotion, and 2) they show us things that cannot be seen by observation.

1) Take the Wolves, please (ba-dum ching!). Right now Wolves fans are, as I’ve said, flagellating due to the failures of the team. If you look around Canis Hoopus you’ll see plenty of people bemoaning that Jefferson isn’t a cornerstone for the team, that he’s worth less than Stephen Marbury, and that the Wolves should be considered rubes for all time. Of course, if you asked any non-Wolves NBA fan, they’d tell you that Jefferson is one of the best centers in the game. And that’s because they aren’t emotionally involved in team, because when you’re emotionally involved with a sports team, you tend to blame the leaders (on and off court) whether or not they deserve it. Additionally, people like the players that do things that they think win, like shoot, assist, and rebound. The problem is that out of those, only rebounding helps win games. Focusing on assists is a cart before the horse, since you could put Steven Nash on a team of Mark Madsens, and all his great passing isn’t going to result in a single assist. Also, shooting is vastly overrated. If a player just jacks up shots left and right, it’s almost as bad as a turning the ball over. That can actually shown by statistics. In any given season, the defense grabs roughly 70% of the rebounds. So missing 10 shots is as bad as turning the ball over 7 times. But one thing Berri shows is that players who score get paid more, and that is because we pay more attention to them. If you look at Top 20 in FGA this year, no less then 9 of the players are overrated, because while they are ball hogs, they don’t score. And yet everyone in the top 20 is considered a star to some degree.

2) Observation is incredibly unreliable at picking up two things: 1) incremental changes 2) nonactions. For incremental changes, an example from Moneyball (a very good book on the creation of modern baseball statistics) is the best one there is: for everyday baseball players, the difference between a .200 hitter and a .300 hitter is two hits a week. If you watch baseball every day you may, may, may be able to see the difference between Nick Punto and Joe Mauer at the plate, but probably not. Now, nonaction: One of the main reasons that Al Jefferson is a top 5 center in the league is that he doesn’t turn the ball over. To backtrack a little, I’ll start by pointing out that center is the most turnover prone position in basketball. Basically, centers get the ball and they either score or give the ball up. But Jefferson commits turnovers about half as often as the average center, and is below the average for any position in the game. That’s simply phenomenal, and something that you just don’t find in centers. But no one mentions it because not something that happens, it’s the lack of something. And the human brain just doesn’t notice those things. While not turning the ball over is critical, it’s not something we notice. And yet turnovers are vital to winning games. That’s why statistics are a better way to understand games.

Now, as to your specific points: One of the issues with the Raptors was that they lost an uncommon number of close games. Any given team should win half of the games that are decided by 5 points or less (long story short: “good teams win close games” is a fiction), and the Raptors only won 4 of 16. It’s essentially like a coin coming up tails 12 of 16 times. It’s certainly possible, but it’s an aberration. Also, the estimate Berri used to estimate the Raptors expected wins has a +/- of 3 wins. So if you take the aberration with the margin or error, the idea of the Raptors winning 49 last year isn’t unbelievable at all. We shouldn’t give the Raptors credit for not winning the games, but they win-loss record did not reflect the quality of the team. Also, stats disprove the GDP situation you mention, since all you need to do is look at the % of people in a third-world country living in poverty, and you can see the statement that GDP determines the health of countries is a lie. It’s a matter of knowing which statistics matter.

Which brings me to my final point: basketball stats are woefully incomplete. There is a lot of baseball statistics that are used with good cause, but baseball has 40 years of work behind it. By contrast, basketball statistics is less than 20 years old. Basketball also has no defensive statistics worth anything (personally, I would give both middle fingers for quality defensive stats), and the best we can do is by inference (like with +/- and adj. +/-). It’s a work in progess, but still very effective. And it’s certainly better than making assumptions based on watching half the games (like McHale seems to do).