I've been thinking about this for quite a while before putting it into words... so bear with me and take the whole journey before you react.

"Confirmation bias" is a well-known and oft-cited phenomenon in the experimental sciences. It identifies the predisposition of the experimenter to prove his hypothesis with the data prematurely or inaccurately. It hits hardest those softer sciences that deal with surveys or psychological studies. For example, a devoted feminist might find preliminary data leading to suggest violence against women spikes on superbowl sunday (this is famously in error, btw). It might equally lead a fan who likes Sergio to see a Jack turnover as more evidence of a decision already arrived at.

Confirmation bias seems particularly relevant at this juncture in the season when we're starting to look harder at next year than this year. I see confirmation bias in every post that wonders why Jack was playing in the 4th quarter last night (who the heck else?). I see confirmation bias in every defense of Jack in the 4th quarter (for example, in Dave's recap he cites 3 assists--well, I watched the game and know at least one of those assists was a desperation pass after 20 seconds of dribbling against the clock where Outlaw bailed him out with a quick jumper; not the offense I think we want).

It exists in other places as well. Not to inject to much racism in the argument where it's not wanted but I often see someone suggest that we don't "need another white guy." The movie "white man can't jump" and Larry Bird famously suggesting that no white guy can guard him are other examples of a kind of confirmation bias: racial discrimination and prejudice.

So in summary for Part I, I think we--as fans--often come across confirmation bias based upon players who look like they can play, do things we like to watch, or confirm some kind pet theory we have of the way the game ought to be played (a fast pace or a half-court or heavy defense or whatever). We see that all the time here, and certainly saw it last night's press conference where someone with a confirmation bias against Jack and for Surge asked Nate the dreaded question, which only confirmed his confirmation bias about fans.

But that leads us to a more pervasive second kind of confirmation bias and what I really wanted to talk about: expert confirmation bias. Where the most brilliant minds in science suffer from this phenomenon (can anybody say Ptolemaic Epicycles?), certainly we should expect coaches, sportswriters, and even GMs to suffer from this as well. What does Nate see when he watches a game? How about Stephen A Smith or Bill Walton or Charles Barkley? Certainly, they know more about the game and their judgments are more considered. But it would be extremely naive to think that Nate doesn't see a defensive lapse from Surge differently from Jack's; or a failed free throw from Joel a little differently than Webster. Players playing for coaches have to bring more evidence to bear to defeat a predisposition than brand new player out of the draft. This is why I think Surge will probably always be a poor fit for Nate and why Joel and Raef might get limited minutes even should they badly outplay Oden. I think when we here "development" and "time" these terms are used to defend decisions made against evidence contrary to confirmation bias. Not that these aren't very relevant issues but, really, how much time and development is needed? Many times, I suggest, this time is longer for those at the positive end of the confirmation bias: players who look like they can play.

This brings us to Moneyball. For those of you who've read the work you know that it's about statistics overcoming confirmation bias. The book, chronicling the Oakland A's, is filled with examples of scouts and coaches and GMs who fail to assess players properly because they do or do not have that look." "He just *looks like a gifted athlete." Meanwhile, capable players are being passed because they look awkward or have an unimpressive physique. Granted, baseball doesn't command the kind of athletes that basketball does but the example is still extremely relevant.

So, we use statistical data to overcome confirmation bias, as any good fantasy owner knows. In this objective world, we can overlook glitzy performance and get at the brass tacks, no?

The basketball box score is nice because it's ubiquitous and universal. After that, I think we can all agree that it's a very poor assessment of a player's contributions. Animart1 has introduced us to the "fouls drawn" stat in Europe. We have steals that don't credit the deflection, rebounds that make no distinction between contested and uncontested balls, shots that are credited equally if wide open or not, judgment calls around assists, and no good measure of defense. Any time we use the box score to credit an argument, we're using a flawed system that can do nothing to contradict ingrained confirmation bias. That's why Oakland in Moneyball was successful: they used a more credible statistical system.

For all the controversey, the +/- point system is an attempt to get a more credible measure. I know people have lots of trouble with it, but I like very much the gesture behind it and what it's trying to do. Is there a better way?

That's why I'm so fascinated with KP's algorithm: presumably, he has a better way of assessing players. I often wonder what his system is. I often wonder about what his system has to say about Surge, Jack, Web and Joel. I wonder if his algorithm is in conflict with Nate, who seems a sure-fire kind of guy who trusts his instincts and gut while watching a player and would have little time for statistial nonsense. What does the system have to say about development and future potential? What is the score that "culture" brings to the argument?

And what do we do as fans? Well, to be good fans, I think we need to do what any good scientist does: look specifically for evidence contrary to your hypothesis. I try to watch for the good that Jack does and the defensive lapses of Surge. I think Dave and other Jack defenders need to look hard at not just the number but the kind of turnovers that Jack coughs up. (In reality, I'm in a place now where I don't think the Blazers have a good PG and it's pointles to speculate on one or the other.)

More to the point, we need to agree as fans what sort of evidence is required to recommend a firm conclusion. That's the conversation I'd like to have... can Blazer's Edge create a whole new Algorithm of assessing talent? It has worked in the baseball world and we have enough talent (and obsessive slavish devotion) to make it work. I think of the following:

1. Unassisted baskets. Good morning, Travis Outlaw! I want to know what players can command a basket when they wish and, more to the point, who operates outside of the flow of the offense.
2. Offense creation. What passes lead to open looks at the basket or free throws? Right now, assists is sadly lacking in generating these kinds of issues.
3. Careless / Dumbass turnovers. Sometimes, a TO is under pressure, sometimes the results of boneheadedness (I saw this last night a few times).
4. Pressured TOs. Who can handle the pressure?
5. Deflections. Something we already keep in some circles.
6. Fouls drawn (I love that... I'd also like to know if they're assisted or producede fully formed from the player himself.)
7. Box outs. That's at least as important as the rebound itself.
8. Help defense. Can we track that somehow?
9. Contested shots. Certainly this matters a great deal.
10. Contribution to an offensive set. Is it obvious when a basket is made as part of a play and can that be quantified?

So those are my thoughts. I'd like to see us in the BE community argue less about what we think we saw, less about the reported box score statistics, and more about what we should be looking for to come to firm conclusions. This is already a very intelligent site as blogs go... and I think it can get a lot better. In fact, if we can figure out a way to bring stats-keeping together on this website, we could change how the game itself is ultimately reported. It has happened in other places.