ProFSL: Pro Fantasy Sports Leagues
Fantasy Leagues => Franchise GM: Rules Changes => Franchise GM: History Books => Franchise GM => MLB Leagues => Franchise GM: Clarifications & Discussion => Topic started by: Colby on August 31, 2010, 11:58:04 AM
-
It is in my opinion, and others, that the defensive statistics in our scoring system have not been worth their while this season. Per much of the discussion in the Jamey Carroll thread (http://www.profsl.com/smf/index.php?topic=8836.10), position eligibility plays a crucial role with defense stats because assists and putouts are awarded differently for each position. Even if we tweaked our position eligibility rules (which I am slightly against), there would always be outliers that enable people to take advance of the scoring system. Also, add in the fact that a big problem persists that prevents us from moving players from one offensive position to another (non-Utility) in the middle of the week. The last one is a huge problem. I feel that the defense stats aren't worth all the hassle. It was agreed upon the RC last year that we may remove defense stats if they did not prove to be valuable.
I would like to hear from all GMs on this matter before I make an executive decision.
-
i vote 150% to remove them
-
I favor their elimination... while I like the effort put into scoring system to realistically include defense, there just isn't enough to work with. Setting the lineups has been a pain whenever an injury or call-up happened that made me want to change my position players in the middle of the week.
-
Sometimes the easiest path isn't the right path. I have come to know Franchise GM as a league that prides itself on being the most realistic league out there. I believe with the removal of defensive stats, we would take a step back. Defense is a REAL thing, Bobby Cox has to worry about it everyday just like everyone else involved with baseball. I know some owners along with me have made some moneyball trades that didn't necessarily land us a big name player, but anaverage player who played solid offensive and defensive. These players are valuable to Major League Clubs and Franchise GM has done a good job at making them valuable in the league. I feel that it would be unfair to move such a drastic ruling so quickly when in the past we have allowed a season to adjust before implementing.
~MTK
-
Feel free to offer suggestions at how to fix it. Perhaps the position-dependent part of the defensive statistics should be removed?
-
i understand what aubrey is saying. its not that im in favor of not counting defense. I am not in favor of using defensive stats if guys like carrol hairston jr and infante and headley are ranked so high on OF that A. you cant extend them to be a roster filler that gives you decent production and B. goes to the OF where they could just rank the stats. I could play omar in the OF but i choose not too. same with dunn. i choose to be realistic and keep them where they play in real life. If fantrax would allow you to only count stats if you played the position in the game that you were in on your fantrax lineup that would be great, however that is not an option. thats why i am totally for removing them. And mike is right, it sucks that you cant move guys around when someone gets hurt.
-
The ideal solution would be to count our defensive stats for the position that the player played in his game. I asked Fantrax about this a while ago, and it is something that cannot be done for the foreseeable future.
-
I am a 100% for removing of defensive stats. Assists and putouts do not refflect a player's defensive ability at all. Even UZR suffers from serious sampling issues.
Removing defensive stats quickly will hurt some teams in the short run but it's for the benefit of the league and it's not the without precedence.
-
I am a 100% for removing of defensive stats. Assists and putouts do not refflect a player's defensive ability at all. Even UZR suffers from serious sampling issues.
Removing defensive stats quickly will hurt some teams in the short run but it's for the benefit of the league and it's not the without precedence.
The 3rd option would be to increase the position eligibility. I would be for doing away with defensive scoring for the sake of being able to move around players during the week. However, to keep the game as realistic as possible maybe Option 3 is viable.
-
The 3rd option would be to increase the position eligibility. I would be for doing away with defensive scoring for the sake of being able to move around players during the week. However, to keep the game as realistic as possible maybe Option 3 is viable.
The infringement on lineup flexibility mid-week is what prompted me to discuss the elimination of defense. I will add more to my case for removing the stats...
We have no home-field advantage in our scoring system. This cancels itself out over the year. Neutral-defense can also cancel itself over the year as players will do worse, defensively, in other ballparks.
-
Get rid of them. Taking everything in totality, they're a larger negative than a positive.
-
Sometimes the easiest path isn't the right path. I have come to know Franchise GM as a league that prides itself on being the most realistic league out there. I believe with the removal of defensive stats, we would take a step back. Defense is a REAL thing, Bobby Cox has to worry about it everyday just like everyone else involved with baseball. I know some owners along with me have made some moneyball trades that didn't necessarily land us a big name player, but anaverage player who played solid offensive and defensive. These players are valuable to Major League Clubs and Franchise GM has done a good job at making them valuable in the league. I feel that it would be unfair to move such a drastic ruling so quickly when in the past we have allowed a season to adjust before implementing.
~MTK
I second this - changing the scoring rules should not be taken lightly as it controls the value of a player / every decision made. We should explore the options of making them better before completely trashing the idea.
-
I have to agree as well that taking out defense would rather hurt this league. Besides it would be unfair to all those managers who built their teams around solid defenses. By the way, a change of this magnitude should be backed by an overwhelming majority of managers and not just by simple majority.
-
For those opposed to getting rid of defense, I pose the following questions:
How are assists and errors meaningful measurements of a player's defensive ability?
-Assists and errors don't take into account the range of field a player covers.
-Outfield assists are a very rare occurrence, there is a lot of luck and randomness involved.
-There is a lot of bias with how errors are assigned to players
If assits and errors are not meaningful, it doesn't really matter how position eligibility thresholds are changed, we are still using non-realistic stats.
Is there another stat that could be used to measure defensive value?
UZR? Highly volatile IMO.
Did you REALLY build your team on defense?
There is so much randomness involved with assists (especially in the outfield), I highly doubt getting rid of defense stats will set your team back.
-
I have to agree as well that taking out defense would rather hurt this league. Besides it would be unfair to all those managers who built their teams around solid defenses. By the way, a change of this magnitude should be backed by an overwhelming majority of managers and not just by simple majority.
A 2/3 majority is necessary to pass a decision. Every GM has a vote... your statements will help me determine if a majority rule should be passed.
-
The best solution would be to tweak the way we are scoring defense. I agree that OF's assists are a mostly random statistic, even considering that OFs with better arm strength tend to have the most assists. Maybe we can eliminate OF assists to solve the issue, but errors and infielder assists, while not the most accurate method, actually do represent to some level a player's defensive progress. Just like POs represent an OF's and 1B's quality fielding to some level. The more range in the OF that you cover the more PO's you'll have, for example. You can see Franklin Gutierrez and Casey Kotchman's stats to support this point. Besides, the total impact defense has on scoring is not too big and the Jamey Carroll issue can be easily corrected by taking simple measures.
-
The defensive scoring system is rather erratic to me. As others pointed out, it has caused lineup setting issues and weird effects like that of Jamey Carroll.
-
I second this - changing the scoring rules should not be taken lightly as it controls the value of a player / every decision made. We should explore the options of making them better before completely trashing the idea.
I agree with Ben and a couple others about implementing this radical change for 2012. Defense is about 1/6 of our scoring system. Is there anything that we can do to fairly fix it for 2011? I will bring up another poll about innings/games played/games started requirements for players to obtain position eligibility. IMO, this is something that can be implemented for 2011.
-
The lineup problems have been a big issue for me. If defense is the culprit then can we agree to replace it somehow and not have position dependent stats?
-
The lineup problems were a huge pain.
-
Finding a way to fix the lineup problems via defensive adjustments should be the priority here. One thing we could do is eliminate the positional-based scoring for defense and replace it with positive points for putouts and negative points for errors. All this does is reflect fielding percentage which will generally be worse for infielders. Thoughts?
-
Sounds good, but since defensive stats won´t be position-dependent, assists should also be part of the scoring system so that MI's don't lose value due to the high amount of errors and small amount of POs that they usually post. Also, whoever is playing UT should not have defensive points.
-
Sounds good, but since defensive stats won´t be position-dependent, assists should also be part of the scoring system so that MI's don't lose value due to the high amount of errors and small amount of POs that they usually post. Also, whoever is playing UT should not have defensive points.
Agreed.
-
Sounds good, but since defensive stats won´t be position-dependent, assists should also be part of the scoring system so that MI's don't lose value due to the high amount of errors and small amount of POs that they usually post. Also, whoever is playing UT should not have defensive points.
Good idea.... A + PO - E = ?
-
Good idea.... A + PO - E = ?
This will be across the board for all positions and players with multiple positions can be switched throughout the scoring period,right?
-
This will be across the board for all positions and players with multiple positions can be switched throughout the scoring period,right?
I can guarantee it would be across the board and players can be switched between positions if we do it for UTIL. If we don't do it for UTIL then there may be issues switching between UTIL and a position. However, I think Fantrax's issue was only if you switched players mid-week between C, CI, MI, and OF. I was able to switch a player between a position and UTIL without a problem.
-
I was hoping that if all positions used the same defensive stats it wouldn't matter where they play as long as they qualify. Personally, having the versatility to move players around is the only reason I would change or eliminate defensive ststs
-
I was hoping that if all positions used the same defensive stats it wouldn't matter where they play as long as they qualify. Personally, having the versatility to move players around is the only reason I would change or eliminate defensive ststs
:iatp:
I believe this is where our discussions are headed. I can dig up the old defensive stats spreadsheet that the RC used last year - it is here on the board.
-
I never had any problems switching players with the UT position either, so that shouldn't be a problem. That formula sounds ok, but I would propose PO plus A minus 10 times the errors (to make a significant difference between someone who makes 1 error and someone who makes 20).
-
I never had any problems switching players with the UT position either, so that shouldn't be a problem. That formula sounds ok, but I would propose PO plus A minus 10 times the errors (to make a significant difference between someone who makes 1 error and someone who makes 20).
I have taken a statistical approach to this problem. We have three basic defensive stat modifiers - errors, assists, and putouts. Our current scoring system does a very good job of splitting these up per position. I am running a simulation using three random multipliers times three weights (for example, a high negative weight is given to errors). I am then measuring the correlation between the current top 100 players' defensive fantasy points against this simulated system. Whatever gives the highest correlation could be a very good replacement for our old defensive scoring system.
-
Fielding percentage has an awful correlation (24.89%) with our defensive stats. I have been continuously running the simulation all day. I started out with random values and then after a good 5000 runs, I took averages and standard deviations of saved values (which ones with lowest correlation were continuously replaced with higher correlated values) and mixed them with random numbers to regress values toward a maximum value. This thing has been running all afternoon, and I have been able to achieve 72% correlation by using the following stat modifiers.
Errors: -28
Assists: 1.4
Putouts: 0.17
At-Bats: 1
I would prefer defensive innings played over at-bats, but there is obvious strong correlation between the amount of AB a player gets and how much they are on the field. This stat is thrown in there as a normalizing statistic.
The simulation is still going on, so I will update this at the end of the afternoon with some sample stats and final weights to leave it up to a vote. The end solution of all of this is to allow lineup flexibility during the week with call-ups, trades, and injuries. People are split on whether they want daily or weekly lineups. Our current defensive scoring system would not be a problem with weekly lineups.
-
The simulation appears to max out at 71.9965%...
Errors: -28
Assists: 1.4
Putouts: 0.17
At-Bats: 1
A lot of players are very close to their values. What happens in this uniform system is that the top defensive 1B lose some of their value (Pujols, Gonzalez) while the middle infielders gain back some value. That's not all that bad considering the middle infield has two of the most important positions on defense. The best part of a uniform system is that it prevent abuse.
The top 100 players, offensive, with their defensive points under this system, versus their points on the current system, and the resulting difference.
Pujols, Albert : 1119.84, 713.41, -406.43
Mauer, Joe : 566.4, 605.48, 39.08
Fielder, Prince : 705.44, 695.19, -10.25
Braun, Ryan : 745.6, 641.88, -103.72
Ramirez, Hanley : 625.12, 794.17, 169.05
Utley, Chase : 980.64, 865.87, -114.77
Teixeira, Mark : 678.08, 773.34, 95.26
Gonzalez, Adrian : 972.8, 754.48, -218.32
Dunn, Adam : 468.8, 239.01, -229.79
Youkilis, Kevin : 786.56, 641.2, -145.36
Lee, Derrek : 747.52, 679.2, -68.32
Howard, Ryan : 594.88, 575.62, -19.26
Cabrera, Miguel : 825.76, 767.19, -58.57
Zobrist, Ben : 505.6, 696.36, 190.76
Choo, Shin: 595.2, 456.29, -138.91
Helton, Todd : 981.6, 823.73, -157.87
Martinez, Victor : 520, 622.16, 102.16
Holliday, Matt : 633.6, 497.55, -136.05
Jeter, Derek : 1365.12, 920.68, -444.44
Bay, Jason : 820.8, 604.7, -216.1
Sandoval, Pablo : 222.4, 459.87, 237.47
Votto, Joey : 588.48, 491.9, -96.58
Werth, Jayson : 571.2, 478.41, -92.79
Lind, Adam : 699.2, 574, -125.2
Tulowitzki, Troy : 750.24, 905.75, 155.51
Suzuki, Ichiro : 572.8, 587.89, 15.09
Reynolds, Mark : 456, 275.98, -180.02
Figgins, Chone : 1020, 683.93, -336.07
Zimmerman, Ryan : 968.16, 608.89, -359.27
Rodriguez, Alex : 607.68, 455.22, -152.46
Berkman, Lance : 859.04, 642.59, -216.45
Abreu, Bobby : 537.6, 398.73, -138.87
Bartlett, Jason : 197.12, 443.16, 246.04
Morales, Kendry : 1024, 678.98, -345.02
Longoria, Evan : 990.24, 661.67, -328.57
Span, Denard : 536, 505.9, -30.1
Drew, J.D. : 625.6, 444.14, -181.46
Upton, Justin : 304, 245.58, -58.42
Ethier, Andre : 582.4, 455.83, -126.57
Hawpe, Brad : 504, 348.21, -155.79
Roberts, Brian : 798.56, 971.13, 172.57
Crawford, Carl : 659.2, 557.99, -101.21
Swisher, Nick : 528, 397.26, -130.74
Johnson, Nick : 515.68, 422.65, -93.03
Kubel, Jason : 707.2, 532.91, -174.29
Damon, Johnny : 526.4, 455.8, -70.6
Lopez, Felipe : 575.52, 755.91, 180.39
Pena, Carlos : 504.48, 469.75, -34.73
Pedroia, Dustin : 922.72, 1065.93, 143.21
Young, Michael : 678.4, 592.44, -85.96
Wright, David : 451.68, 336.49, -115.19
Morneau, Justin : 732.16, 683.84, -48.32
Kemp, Matt : 699.2, 633.69, -65.51
Victorino, Shane : 716.8, 660.32, -56.48
Scutaro, Marco : 1634.24, 890.5, -743.74
Coghlan, Chris : 473.6, 404.07, -69.53
Butler, Billy : 635.52, 649.24, 13.72
Ramirez, Manny : 390.4, 267.83, -122.57
Jones, Chipper : 268.64, 177.65, -90.99
Uggla, Dan : 666.88, 757.28, 90.4
Cano, Robinson : 858.72, 918.45, 59.73
LaRoche, Adam : 991.84, 819.45, -172.39
Ibanez, Raul : 636.8, 492.81, -143.99
Matsui, Hideki : 681.6, 456, -225.6
Markakis, Nick : 664, 542.86, -121.14
Hill, Aaron : 1132.32, 1215.79, 83.47
Ellsbury, Jacoby : 670.4, 634.29, -36.11
Konerko, Paul : 779.84, 740.78, -39.06
Cuddyer, Michael : 670.4, 527.56, -142.84
Hunter, Torii : 539.2, 478.16, -61.04
Granderson, Curtis : 672, 620.6, -51.4
Kinsler, Ian : 815.68, 931.22, 115.54
McCann, Brian : 278.4, 390.28, 111.88
Bourn, Michael : 657.6, 600.47, -57.13
Posada, Jorge : 308.8, 367.67, 58.87
Blake, Casey : 906.72, 590.37, -316.35
Branyan, Russell : 466.08, 415.59, -50.49
Lee, Carlos : 761.6, 602.81, -158.79
Rolen, Scott : 923.04, 675.27, -247.77
Callaspo, Alberto : 1297.76, 662.45, -635.31
McCutchen, Andrew : 454.4, 435.54, -18.86
Fukudome, Kosuke : 636.8, 501.85, -134.95
Willingham, Josh : 427.2, 273.41, -153.79
Montero, Miguel : 326.4, 381.61, 55.21
Cust, Jack : 657.6, 470.13, -187.47
Overbay, Lyle : 854.4, 656.56, -197.84
Escobar, Yunel : 565.92, 769.07, 203.15
McLouth, Nate : 587.2, 544.81, -42.39
Cameron, Mike : 587.2, 506.28, -80.92
Ortiz, David : 15.04, 528.54, 513.5
Cruz, Nelson : 492.8, 360.57, -132.23
Cabrera, Asdrubal : 636, 795.17, 159.17
Napoli, Mike : 291.2, 314.62, 23.42
Hudson, Orlando : 983.52, 884.85, -98.67
Beltran, Carlos : 324.8, 291.56, -33.24
Cantu, Jorge : 334.24, 454.55, 120.31
Pence, Hunter : 571.2, 493.12, -78.08
Byrd, Marlon : 604.8, 527.84, -76.96
Jones, Garrett : 624, 402.22, -221.78
Diaz, Matt : 516.8, 316.49, -200.31
-
Im only in favor of getting rid of the defense as a last resort. The most important thing is position eligibility. By making that more even to where your best OF arent infielders than it wouldnt matter. If you can come up with a new system that works better than thats pretty awesome as well. I like the defensive stats, just not when they're abused.
-
Im only in favor of getting rid of the defense as a last resort. The most important thing is position eligibility. By making that more even to where your best OF arent infielders than it wouldnt matter. If you can come up with a new system that works better than thats pretty awesome as well. I like the defensive stats, just not when they're abused.
We are trying to work on that but the thread has been rather light in conversation - http://www.profsl.com/smf/index.php?topic=9507.0 (http://www.profsl.com/smf/index.php?topic=9507.0)
-
Colby no matter how good your correlations are getting, the underlying assumptions that E, PO and Field %s are good metrics for defensive prowess are still false.
-
Colby no matter how good your correlations are getting, the underlying assumptions that E, PO and Field %s are good metrics for defensive prowess are still false.
Yeah, I know your position on our current defensive scoring system, but it does a much better job than anything else.
-
Yeah, I know your position on our current defensive scoring system, but it does a much better job than anything else.
If I understand it correctly, you are using Field % as your response variable? Should something like UZR be a better option?
-
If I understand it correctly, you are using Field % as your response variable? Should something like UZR be a better option?
UZR would be better. I don't have the stats readily available as of this moment.