Post-season Roster Evaluation of Mystery Cards

I've started to do a Post-season evaluation to determine if I truly got a "raw deal" or a "good deal" with respect to the luck involved in getting the better of the five mystery versions of players drafted and/or acquired after the draft.

At the end of the mystery league season I evaluated my 25-man roster and assigned a number from 1-5 to indicate how good their mystery card was. I used WHIP for pitching and OPS for hitting. For example, if Joe Blow's mystery card was from the 1968 season and 1968 was a better OPS (or WHIP) than 1970 and 1966, but, not as good as 1964 and 1969, you would assign a '3' because you got his third best mystery card.

Anyway, my total mystery card points calculation was 72/25 = 2.88. I probably should have included players waived, but the 2.88 gives me a general idea of what I had to work with. The calculation has nothing to do with evaluating whether I got the players I wanted from the draft. It only evaluates the players I ultimately had on my roster.

I'm not sure what to do with that statistic, but, it tells me that, on the average, my mystery cards were not "worse than expected" and not "better than expected". In other words, on the average, I got the third best mystery card for each player on my final roster.

Since I got to the finals, I'm thinking that my mystery rating of 2.88 means that it's not necessary to go crazy and waive players too frequently if you think you've only got their third best mystery card. As long as the average mystery rating of your final roster is 3 or lower, you can't really say you got "bad" cards. And if you didn't get bad cards, then, you should have a realistic expectation that you can compete for post-season play, unless you fared poorly at evaluating who you should draft, or if the competition simply got a mystery card calculation significantly better than 2.88 (using my example).

The bottom line is that if you excessively waive players and acquire new players in response to a belief that you've only got "third best" mystery cards, then, that could prove to be a high risk strategy. Of course, other factors, such as standard deviation between best and worst mystery cards of a particular player, should influence your decision on how aggressive to be at waiving your players who aren't producing to your liking.

Anyway, my total mystery card points calculation was 72/25 = 2.88. I probably should have included players waived, but the 2.88 gives me a general idea of what I had to work with. The calculation has nothing to do with evaluating whether I got the players I wanted from the draft. It only evaluates the players I ultimately had on my roster.

I'm not sure what to do with that statistic, but, it tells me that, on the average, my mystery cards were not "worse than expected" and not "better than expected". In other words, on the average, I got the third best mystery card for each player on my final roster.

Since I got to the finals, I'm thinking that my mystery rating of 2.88 means that it's not necessary to go crazy and waive players too frequently if you think you've only got their third best mystery card. As long as the average mystery rating of your final roster is 3 or lower, you can't really say you got "bad" cards. And if you didn't get bad cards, then, you should have a realistic expectation that you can compete for post-season play, unless you fared poorly at evaluating who you should draft, or if the competition simply got a mystery card calculation significantly better than 2.88 (using my example).

The bottom line is that if you excessively waive players and acquire new players in response to a belief that you've only got "third best" mystery cards, then, that could prove to be a high risk strategy. Of course, other factors, such as standard deviation between best and worst mystery cards of a particular player, should influence your decision on how aggressive to be at waiving your players who aren't producing to your liking.

That's an interesting formula you came up with, and it would be interesting to see if actual hard statistical analysis of possible WHIP and OPS variations for all seasons on your players cards supports its use. However, assuming it does, I'm not sure if it addresses the situations where managers actually vacillate on whether to release a player or not in the Mystery Game. As a manager who usually gravitates towards players with 3 or more good cards, I can definitely say that I almost never release a player if I have determined he is on his third best card unless another manager has released a better player who has hit on his third best card or higher. As to other managers, I would say that--outside of those ultra-gamblers who pack their teams with high-risk/high-reward players with only two good cards--most other managers would also hold onto players they have determined to be on their third best cards.

Therefore, a more relevant use of your formula would be if you applied it to teams--including their waived players who saw substantial playing time--who averaged a 2 or worse and see if those teams succeeded or failed. If either result tended to be the case, and if some harder statistical analysis supported your formula, you could possibly see if managers would be better off releasing or keeping players they've determined (or intuited) to be on their 4th or 5th best cards, players most managers are currently releasing or considering to release.

I am puzzled by this discussion. Are you equating having the 3rd best card of a an All-Star Player with having the 3rd best card of bench role player ?

For your idea to be useful, you need to weight the value of the player ( similar to 'degree of diffculty' ), and weight the value of each possible year.

I am puzzled by this discussion. Are you equating having the 3rd best card of a an All-Star Player with having the 3rd best card of bench role player ?

For your idea to be useful, you need to weight the value of the player ( similar to 'degree of diffculty' ), and weight the value of each possible year.

I'm pretty sure he was primarily focusing on starting players, as It would be less successful against bench players who rarely have 3 "good" cards. I'm skeptical of such "averaging" myself, as I am of taking such a broad approach of evaluating my players; I always evaluate my players with the scrutiny you mention. However, like Paul following his "system," I rarely cut a starting-level player whom I've determined to be on his 3rd best card unless he's a high-risk/high reward card with a crappy 3rd one.

I agree with Strether's statements.

Example of Mystery Performance Calculation

Here's a more detailed example of a team of mine that won over 90 games and lost in the semi-finals:
Calculation: 49/18 for players on roster most of the year = 2.72 (compare to average of 3.0)
Partial Players: Giusti was dropped; Buskey added MPC net result = 0 (No MPC improvement)
Players not ranked are because they were bench players or did not see significant playing time.
The MPC factor is to the right of the player's year.
Pitching
Candelaria, John '80 5
Siebert, Sonny '68 1
Martinez, Dennis '81 3
Bibby, Jim '76 2
Drago, Dick '74 3
Lerch, Randy '77
Hall, Tom '73 L 4
Lavelle, Gary '78 4

Brandon, Bucky '71 not much playing time

---------------------------------
Drops and Adds
# Giusti, Dave '74 5
# Gossage, Rich 'Goose' 5
# Pina, Horacio '71
* Buskey, Tom '80 (replaced Giusti) 5

--------------------------------------
Offense
Martinez, Buck '79 1
Kendall, Fred '72
Colbert, Nate '71 3
May, Lee '75 3
Morgan, Joe '76 1
Beckert, Glenn '72
Reitz, Ken '76 4
Howell, Roy '77
Concepcion, Dave '75 5
Russell, Bill '70 3
Foli, Tim '79
Piniella, Lou '77 1
Cromartie, Warren '78 2
Lee, Leron '72 1
Howard, Wilbur '74
Monday, Rick '75 2
------------------------------------------
Drops
#Herrmann ''71 4
#Chambliss '79 4
#Muser '74
#Stanley '71 1

---------------------------------------------------------------------------------------------------------------------------
Conclusions:
1) my average MPC of 2.72 (Mystery Performance Calculation) was better than average. I can't complain about getting bad cards.
2) my average MPC for pitchers was slightly over 3. I could have done more to try to upgrade, I suppose
3) My attempted upgrades were few. I didn't accomplish much other than churn cash. Although, going from a Chambliss (4 MPC) to a Colbert (3 MPC) was possibly worth the money, except that I had to drop Stanley (MPC 1) in trying to get Colbert.

Obviously, there are other factors to consider, such as weighting your higher salaried players. Having a stellar Joe Morgan (MPC 1) card is more important than having a stellar mop-up pitcher of MPC=1.

My most reason 90s season, I was convinced I had the worst season for five different relievers ranging in salary from $2.35 to 6.19 million. It turns out that not one of them was on his worst card--not one was even close to his worst season. Looking back, I can see I was way too quick to pull the trigger. In my defense, I stayed with Arthur Rhodes too long last time I had him. Anyway, evaluating relief pitchers during the season is hell!

http://onlinegames.strat-o-matic.com/team/1126540

Everybody must have thought Dennis Eckersley was the 1991 version, not the 1988 version and decided to stay away after you let him go. It's like that old tv show, To Tell The Truth. All the panelists voted for 1991, but, it was the 1988 guy who was the real guy.