This is a tough one. 

It's always a tough one. Well, sometimes it's a tough one. Every so often, it's a tough one. This time, let's say, it's a tough one. Back to this, again.

The hardest thing about this award is divining and trying to separate people's expectations for certain players with their biases, or lack of knowledge about a particular player. Not everyone has the time to be an NBA nerd, not everyone has the time to sift through per-minute stats or game discs. So you have to inform them about what, exactly, could be deceiving them. And people don't like that.

Not to pick on the guy, but take Scoop Jackson. He selected Roger Mason Jr. for his Most Improved Player candidate, and he probably has an actual vote. And not to pick on RMJ, but this is a ludicrous choice.

To Scoop, it probably makes sense, because it's obvious that he doesn't watch much of the NBA. But there is absolutely nothing that has "improved" about Roger Mason Jr. in 2008-09. Now, there are some different things about the Spurs guard. For instance, he's been on national TV a lot more in 2008-09, and he's playing nine more minutes per game in 2008-09.

And for Roger, this is an improvement. He no doubt loves being the hero on several nationally televised contests, and every player loves more minutes. And as a result of these minutes, his points per game average has shot up by about three per contest. Improvement, right?

Well, no. He's playing more minutes, but he's doing less with them. Per-minute, his scoring is down from last season. His rebounding? Up a little. His assists? Down. His blocks? Down. His steals? Down. Shots per-minute? Down. His shooting percentage? Down. Usually a decrease in shots results in an increase in shooting percentage. His turnovers? The same.

"Most Improved Player," or, "a Guy Scoop Jackson Didn't Really Pay Attention to Until December of 2008, Despite What He'll Probably Tell You"? You make the call.

From there, you have to convince people that per-minute, pace-adjusted stats are the way to go. Per-minute, because those numbers don't tend to fluctuate wildly year-to-year, or team-to-team, or role-to-role. History has proven that.

Pace-adjusted? Well, that makes it easier to compare the relative value of performance between a guy like, say, Danny Granger (who plays on an ultrafast team that runs at every given opportunity), or, say, Brandon Roy (a guy stuck on a team with the lowest number of possessions per game).

Ignore pace-adjusted stats at your own peril, lest you think Stephen Jackson had a better year than Paul Pierce, or that Chris Duhon had a better season than Jason Kidd. Without pace adjustment, both Pierce and Kidd come out a step slow, and that isn't fair. Nor would it be correct.

And to me, the trickiest part is expectations. And it's the reason why Kevin Durant is not my Most Improved Player.

There was nothing about Durant's rookie turn, at the age of 19, that wouldn't have suggested the way he's playing in 2008-09. I don't mind for a second handing the award to a second or third-year guy if he makes a huge, unexpected leap from year to year, but a kid with a 15.8 PER, 20 points, and 4.4 rebounds in about 34 minutes playing at age 19 should well be expected to produce a 20.6 PER, alongside 25.3 points and 5.5 rebounds in 39 minutes a game at age 20.

That's a big leap from year to year, but it's not an unexpected leap. It is improvement, but it's nothing that you wouldn't expect. For a star of Durant's caliber, this is just gradual, year-to-year improvement. This is his production level's — based on his rookie year — idea of "gradual." He's that good.

This might be hard to fathom, but pretend Michael Jordan didn't break his foot early in his second NBA season. If he would have dropped 33 a game while playing 40 minutes, up from 28 and 38 minutes the year before, would he be an MIP? And because he wasn't as young as Durant in his first two seasons, his jump wouldn't have been as severe.

LeBron James' PER jumped about seven points in his second season, but I think that was to be expected. It's the player he is. Kevin Durant was like this last year, he's just growing at a rate you can expect. A brilliant, brilliant rate.

So, to me, we have to take it to the players who turned into something they weren't, previously.

This means getting rid of all the expected jumps, and getting rid of all the "improvement" that really came in the form of increased minutes.

David Lee? You improved this season. Your points shot up by over five a game, and your rebounds by nearly three a contest. But you also played six more minutes a game. Your per-minute numbers did improve, but not in a way that couldn't be explained away by your gradual, year-to-year, improvement.

You'll improve at this rate for the next few years until you hit your prime, so be happy with that. And be happy that your 35 minutes a game are a nice two fingers to the face of the people that argue that certain players are only good for 27 or 28 minutes a night "... because, well, that's what the coach plays him. Are you saying you're smarter than the coach? What have you ever coached?"

Charlie Villanueva? You made a solid, somewhat unexpected jump after your numbers declined for a couple of years. Nice turnaround, appreciate the consistency, but it wasn't enough.

Joel Przybilla? Great year, but check out his per-minute numbers (hell, check his per-game numbers) from his first two years in Portland. They don't hand out awards for returning to 2004-06 form.

LeBron James and Chris Paul? You improved, big time, but not to a level that wasn't suggested by your still-overlooked turns in 2007-08.

From there, to me, it comes down to three guys. Brandon Roy, Danny Granger, and Devin Harris.

Harris is the popular pick. He struck early, averaging 26 a game in November as everyone put the award aside for the lightning-quick Nets guard. But few seemed to notice as his points per game average dropped by 10 points in December. That averages out to about 21 a game, and that's what he finished the season with.

Now, and you'll have to trust me on this, had you told me last summer that Devin Harris (at age 25, turning 26) would score 21 a game for the Nets this year, I wouldn't have blinked. They were a subpar team, with or without Vince Carter (who I thought would be traded out West), he's steadily improving, he'll get a fair amount of shots, minutes, and freedom from Lawrence Frank.

The problem I have is that Harris had an 18 PER with Dallas last year, while being lorded over by Avery Johnson. He was underrated, he was injured, and then he came back a little too early with the Nets to finish out the year. So jumping from 18 to this season's 21.6 PER on a team that affords him all sorts of new opportunities just seems about right.

It is improvement, it's very good improvement, but has it been the best improvement?

He's shooting worse, the worst mark since his rookie year. His rebounds are the same. His assists are the same despite being given more chances to collect dimes on the Nets, instead of having to give the ball up early in a possession as he had to in Avery's offense. Again, these are per-minute numbers. Numbers that hold up best. Numbers that aren't clouded by a coach who coached well enough to be fired last season.

Steals the same per-minute, turnovers the same per-minute, points are up because he gets to shoot more. About three more shot attempts per 36 minutes, over four more shot attempts per game.

So, his points have gone up. He's been great, and worth our time and this breakdown and an All-Star spot. But has he been the most improved? I don't think so. He's up there, but he hasn't jumped enough.

Danny Granger? He's improved. His game is different. Just by watching this kid play over the last two years, you can tell.

He's moved from a standstill, Richard Jefferson-type prone to bad ball-handling and iffy choices, to an all-around offensive go-to guy. You might be in trouble if he's your best player, but he kept the Pacers in a ton of games this year, and was one of the league's top fourth quarter scorers without adjusting for pace.

Watching your PER jump five points? Indiana's face past doesn't cloud that. This is why I use the stat. His shooting percentages have remained the same, but with his improved dribble, he's also been able to free himself for more perimeter and interior looks. That's not just a function of Mike Dunleavy Jr. being out for most of the season. Granger has improved.

His rebounds have gone down a bit, but everything else has stayed the same. The biggest thing, besides the points jump? His turnover rate. His per-game turnovers have actually increased, but when you use turnover rate (the percentage of possessions he uses that actually end up in turnovers), you can tell how much he's improved at taking care of the ball. From 11.1 percent to 9.8. Don't sniff at that.

(And this is why we use that stat. Granger has had the ball in his hands way, way more in 2008-09 than he had it in 2007-08, so it's not fair to say that he's been even or worse in comparison because his turnovers per game jumped from 2.1 to 2.4 a contest.)

His points per game has jumped by 6.1 since last season. That's a ton, make no mistake, but he also plays on the NBA's third-fastest team. Lots of chances to score those 6.1 points when you're giving up 120 on the other end and shooting with 18 left on the shot clock. This isn't meant to demean Granger's significant accomplishments this season. This is just pointing out what happened. And the points increase is what he's leaning on the most.

And in the end, that's why I go with Roy.

You see, Danny Granger with a 16.7 PER in 2007-08? You tell me that he's going to jump to 21.6 PER in 2008-09? With Dunleavy out and more chances to score? That sounds about right.

Brandon Roy, a nice player, jumping to superstar-level 24.1 PER? That's a bigger jump. And that's not even taking into account his improved defense (he shuts down small forwards this year compared to last, while doing solid work on shooting guards), something PER can't document.

With Harris and Granger, increased responsibility and increased shots per game resulted in declining shooting percentages for Devin, and sustained (almost to the exact number) shooting percentages for Granger.

With Roy? With the same team? He's taken more responsibility, and earned more responsibility, taken more shot attempts, and earned them with big increases in shooting overall (45 to 48 percent ... for a guard!), three-point percentage (34 to 38), and free throw percentage (from 75 to 82). Though it must be noted that his free throw mark last year was likely a fluke, as he shot 83 percent in his rookie year.

And while only a four-point per game jump seems small in comparison to Granger's six points, remember that the Pacers are the third-fastest team in the NBA, while the Trail Blazers are far and away the slowest. Couple that with the fact that nobody on the Trail Blazers was lost for most of the season due to injury — Roy didn't have someone else's shots to take — and the fact that a jump to this sort of output in his third season seems wholly out of line with his contributions in his first two seasons, and you have my Most Improved Player.

It's not that interesting. It's not that overwhelming, and it's not that mind-blowing. Just a very good player who vaulted into the NBA's top tier, and just a step behind Kobe at his position. That's improvement. More than any I've seen this season.