The Rotisserie system is not without its faults, which is why we feel impelled to tinker. But there is one thing that we rarely tinker with. The entire system for evaluating teams in a Rotisserie league has become nearly sacrosanct. The ranking of statistics, by category, and then the assigning of ranking points to determine the standings, is a fundamental component of the game's design.

And probably the most deceptively inaccurate component of the game.

This method of ranking teams -- or the ranking of anything -- distorts the truth. Tiny differences in the numbers can be magnified; huge variances can be minimized.

Let's take a simple test league as an example, a 4-category, 8 team league:

HOME RUNS          RBIs
Team   No. Pts     Team   No. Pts
====  ===  ===     ====  ===  ===
ONE   200   8       THR  905   8
TWO   198   7       ONE  856   7
THR   196   6       FIV  840   6 
FOU   187   5       FOU  831   5
FIV   152   4       TWO  821   4
SIX   151   3       SEV  807   3
SEV   150   2       EIG  806   2 
EIG   149   1       SIX  766   1

STOLEN BASES       BATTING AVERAGE        TOTAL
Team   No. Pts     Team   BA  Pts       Team   Pts
====  ===  ===     ====  ===  ===       ====   ===
TWO   179   8       ONE  287   8        ONE     27
SIX   170   7       THR  286   7        THR     26
FOU   165   6       SEV  282   6        TWO     22
THR   160   5       FOU  277   5        FOU     21
ONE   151   4       FIV  276   4        FIV     17
FIV   150   3       TWO  272   3        SEV     13
SEV   120   2       EIG  271   2        SIX     12
EIG   118   1       SIX  270   1        EIG      6

A nail-biter in this league, right? Sure was. Team ONE eked out the league title, backed by narrow victories in the HR category and finishing just ahead of Team FIV in stolen bases. Meanwhile, Team THR gave a valiant effort, falling just short in HRs and batting average while crushing the opposition in RBIs.

And look at poor Team EIG, dead last and the furthest distance that any two teams are separated in the standings. But that's deceptive. With just three more HRs, two more RBIs and three SBs, EIG could have conceivably picked up six points on their own, while dropping the two teams ahead of them enough to sneak into sixth place. A net gain of about nine points -- a 33% swing in ranking points -- all for the want of one extra HR and one extra SB every two months.

So this particular set of standings reflected an exciting race. Here's another set of standings...

HOME RUNS          RBIs
Team   No. Pts     Team   No. Pts
====  ===  ===     ====  ===  ===
ONE   232   8       THR  883   8
TWO   188   7       ONE  880   7
THR   186   6       FIV  828   6 
FOU   182   5       FOU  822   5
FIV   160   4       TWO  820   4
SIX   159   3       SEV  817   3
SEV   156   2       EIG  792   2 
EIG   144   1       SIX  790   1

STOLEN BASES       BATTING AVERAGE        TOTAL
Team   No. Pts     Team   BA  Pts       Team   Pts
====  ===  ===     ====  ===  ===       ====   ===
TWO   179   8       ONE  290   8        ONE     27
SIX   175   7       THR  281   7        THR     26
FOU   164   6       SEV  280   6        TWO     22
THR   163   5       FOU  278   5        FOU     21
ONE   151   4       FIV  277   4        FIV     17
FIV   143   3       TWO  276   3        SEV     13
SEV   130   2       EIG  271   2        SIX     12
EIG   108   1       SIX  270   1        EIG      6

Um... same standings. Same number of total league HRs, RBIs and stolen bases. Anything but a close race.

Team ONE crushed the opposition in HRs and BA, held large mid-standings leads in the other categories, and fell just short of first place in RBIs.

Meanwhile, TWO's performance was a lot closer than the four points they fell short of second place would suggest. And Team EIG? Deservedly in last place.

How can two sets of identical standings describe two completely different realities?

Therein lies the problem with this type of evaluative system. The inherent deceptiveness when you use ranking points makes it more difficult to accurately evaluate teams, or anything that looks at multiple choices.

What's worse, writers often use this type of system to support arguments in other analytical venues. You will see fantasy player draft lists constructed using ranking points. You will see companies publish comparative analyses of different services and use ranking points designed to support their promotional hype. Heck, the Elias Sports Bureau feeds Major League Baseball ranking points to determine whether players are Type A, B or C free agents.

And all of them are feeding you garbage.

They say that numbers can be manipulated to reflect the particular reality you want people to believe. Ranking systems like this are pure “razzle dazzle.”

The most amazing part about all this is that the remedy to the problem is exceedingly simple. If we were to scale the ranking points based on each team's relative productivity in each category, we would create a far more accurate picture of how the standings shape up.

For the league above, we'd give the first place team in each category 8 points, the last place team 1 point, and then scale all the other teams in between. This would be done by first calculating the spread between first and last place (say, 51 HRs). Then we'd take the distance each team is from first place, divide it by the total category spread, and multiply by 7 (the marginal difference between finishing in first versus last). The resulting value would be subtracted from 8 to determine the true ranking point value. Thusly...

HOME RUNS      
Team   No. Pts   Calculation         Pts
====  ===  ===   =================   ===
ONE   200   8                        8.0 
TWO   198   7    8 - (2 / 51) * 7 =  7.7
THR   196   6    8 - (4 / 51) * 7 =  7.5
FOU   187   5    8 - (13 / 51) * 7 = 6.2
FIV   152   4    8 - (48 / 51) * 7 = 1.4
SIX   151   3    8 - (49 / 51) * 7 = 1.3
SEV   150   2    8 - (50 / 51) * 7 = 1.1
EIG   149   1                        1.0

This provides a much truer view of the spread in the standings.

For the truly ambitious, you could even place a weight on the relative value of each category and in this way, make sure that stolen bases are not as important as home runs, for instance.

Let's see how our tight race above would have shaken out had we scaled the standings points...

HOME RUNS          RBIs
Team   No. Pts     Team   No. Pts
====  ===  ===     ====  ===  ===
ONE   200  8.0     THR  905   8.0
TWO   198  7.7     ONE  856   5.5
THR   196  7.5     FIV  840   4.7 
FOU   187  6.2     FOU  831   4.3
FIV   152  1.4     TWO  821   3.8
SIX   151  1.3     SEV  807   3.1
SEV   150  1.1     EIG  806   3.0 
EIG   149  1.0     SIX  766   1.0

STOLEN BASES       BATTING AVERAGE       TOTAL
Team   No. Pts     Team   BA  Pts      Team   Pts
====  ===  ===     ====  ===  ===      ====   ====
TWO   179  8.0     ONE  287   8.0      THR    28.9
SIX   170  7.0     THR  286   7.6      ONE    26.3
FOU   165  6.4     SEV  282   5.9      TWO    21.3
THR   160  5.8     FOU  277   3.9      FOU    20.8
ONE   151  4.8     FIV  276   3.5      FIV    14.3
FIV   150  4.7     TWO  272   1.8      SEV    11.3
SEV   120  1.2     EIG  271   1.4      SIX    10.3
EIG   118  1.0     SIX  270   1.0      EIG     6.4

Amazingly enough, Team ONE does not even finish in first place using this method. Their more distant finish in the two categories they did not win was enough to push them down the standings. But this is a more accurate view.

It's bad enough that we've been married to such a system to determine our Rotisserie results. But you need to raise that red flag high when you see this method being used for more critical decision-support purposes.