Guru:
Your latest Feedback letter and response were great. You're right about the stats being based on last year's performance (read: not necessarily indicative of this year's performance). However, in some cases, that's the exact reason why the tables are important.

SW seems to "charge" owners $200 per SWP scored last year. Players like Gugliotta and Olajuwon missed time last year and are easily underpriced because they did not score many SWP. That's a flaw with the SW pricing system. These are players whose performance last year will almost certainly not be indicative of their performance this year. Your tables expose these players immediately. - Gareth (1/28)

response from Guru:
This type of analysis can be helpful in identifying players to think about. But it has some very significant limitations.

First, you are presuming that the two centers should not cost more than $10 million. If you applied that rationale to all positions, you'd miss opportunities to combine an expensive player at one position with a cheap player at another. There's no reason to expect that all positions should have the same average cost.

Second, your approach would tend to "underspend", since each position would probably average a little less than the max. If you combined the excess cash from all three positions, you'd probably have enough loot to upgrade a player somewhere.

But all that aside, the biggest flaw is the reliance on last year's stats. I'll admit, I tend to do this as well, since I'm naturally analytical and last year's stats are what we have available. But even in a normal year, basketball players tend to change a good bit from year to year. In particular, it usually takes new players several years to reach their stride - especially with so many players now starting with less than four years of college hoops experience. And this is probably anything but a normal year. With training camps so short, many players will take longer to find their "A" games. And with so many free agents changing teams, many players will be assuming significantly different roles, which makes last year's stats irrelevant at best, and misleading at worst.

I have used a spreadsheet optimizer routine that actually assembles the best 10 player roster for $50 million, again based on last season's SWP/G averages. This at least avoids some of the first limitations I mentioned above. But it still has that dangerous reliance on last year's averages. And it produced a roster that has another potential flaw. The resulting roster turned out to be highly "barbelled", with four players costing less than $2 million, and 3 players costing more than $9 million. Even if the stats are correct, and this roster turns out to be the best at producing points for an aggregate cost of $50 million, it still probably isn't the best roster for the game, because the primary objective in the first few weeks is to generate gains in value, even if that results in the opportunity loss of a few SWP/G for the time being. And it is far from obvious that a barbelled roster will be the best at generating price gains. In fact, my intuition tells me that, all other things being equal (and, of course, they're not), a roster of players all priced closer to the average is likely to experience better price appreciation.

Finally, if you just look at my posted stats tables, sorted by SWP/G, and scan down the price column, you can pretty easily identify who the bargains are - based on SWP/G. Using an optimizer tends to pick out the most obvious of these players. But this type of optimization analysis is still fun to tinker with - especially when there aren't any games going on. Still, many of the best values this season won't be the ones who are mispriced based on last year's stats; they'll be the ones who are producing more points this year than last. And my stats tables don't indicate who they are. That requires a lot more insight, research, and luck. (Sigh.) - Guru

Name           Pos     Team     Price    SWP
Cunningham, R   QB      MIN     $9.34    350
Marino, D       QB      MIA     $5.87    116
Hoard, L        RB      MIN     $2.80    276
Murrell, A      RB      ARI     $4.36    192
Bates, M        RB      ARI     $1.66    178
Johnson, K      WR      NYJ     $5.83    396
Smith, J        WR      JAC     $5.64    318
McDuffie, OJ    WR      MIA     $4.27    218
Glover, A       TE      MIN     $2.72     98
Richey, W       K       SFO     $1.51     40
DENVER          DEF     DEN     $5.89   -137
Total                          $49.89  2,045