EloRater thoughts
Posted by Andy on March 28, 2011
A reader pointed out something interesting about the EloRater, and I'm waiting for Sean to address it before I blog about it. But in the meantime, here are some things that interest me about the current rankings:
Batters
- Barry Bonds is currently 12th. I have a feeling he is going to be sentenced to jail pretty soon. I wonder what will happen to his ranking then. I also think 12th is too low to rate Bonds. Among batters, I feel comfortable in the assertion that he is the greatest player to come after Mike Schmidt and before Albert Pujols (with apologies to Alex Rodriguez).
- Mickey Mantle fifth and Yogi Berra 44th jump out to me as two Yankees rated too highly. (Lest you think I am anti-Yankee, though, I have no issue with Babe Ruth and Lou Gehrig in the top 3.)
- Interesting that A-Rod and Griffey are just a tick apart. When Griffey came on the scene, it seemed as if he was going to be the greatest post-WWII player. By the time A-rod got into his groove, though, Griffey was already on the decline. Many would have bet that A-rod was going to have the far superior career. Here they are all those years later, though, side by side. (Although A-rod of course still has a few years to go, and he's my pick for AL MVP this coming season.)
- Robin Yount at 36--I think this is an appropriate ranking, but I am always stunned to see Yount get his due. I always felt like he was largely ignored by playing in Milwaukee and yet got elected to the HOF quite easily. I suppose it's just me that is surprised by this--perhaps I need to adjust my thinking and realize that everybody does realize what a truly awesome player he was.
- Manny Ramirez is down at #43 and has the lowest W-L% of the top 56 players. This to me suggests a bit of Manny-hating. I sort of feel like most people acknowledge Manny as one of the great hitters of his era, but are happy to put him on the losing side of almost any comparison to a top player from any era. I guess that's what's going on with his rating.
- Brooks Robinson and Eddie Murray, two guys who defined the Orioles over certain years, are consecutive at #49 and #50. I like the symmetry.
- Entries 61 through 80 are all interesting. Many of these guys are fringe HOFers (some are in, some are out, and some come up for vote soon): Jim Thome, Craig Biggio, Larry Walker, Willie Stargell, Bill Dickey, Ron Santo, Jim Edmonds, Andre Dawson, Luke Appling. Fascinating.
- Ozzie Smith at #83...appallingly overrated? He's actually 82nd all-time in WAR among position players, so maybe not. I find it so tough to rank these guys with huge defensive components.
- Carlton Fisk and Gary Carter right next to each other at #87 and #88
- Tim Raines (#97) ahead of Gary Sheffield, Mark McGwire, Alan Trammell...
- Scott Rolen at #104?!?!?! First I'm thinking it's way too high, but then I see he's in the top 80 for WAR among position players. But he's ahead of Miguel Cabrera...but but but....
Pitchers:
- Mariano Rivera is 14th all-time. OK so he's not the Greatest Pitcher of All Time, but a ranking that high for a closer is astounding.
- Roy Halladay is 16th, a tick ahead of Pedro Martinez and well ahead of Bob Feller.
- I noticed that active players have way more ratings than non-actives. Active players have 400-500 ratings while most inactives have 100-200. I suppose that's because a lot of readers skip rating players from years earlier that they may not be familiar with. (I've done that myself.)
- Jim Kaat and Catfish Hunter are nearby at #61 and #63. That's not going to help my confusion!
- Tim Lincecum is 88th. I like Timmy, but isn't a bit early to put him in the top 100? His W-L% is insanely high at .649. Of course, one of the cool things about this player rater is that there are no rules. You can vote for or against a player for any reason you see fit.
OK I'm out of time...but please share some of your observations below.
March 28th, 2011 at 10:21 pm
The fact that both Bonds and Clemens have failed to crack the top 10 has been driving me crazy. I've assumed--incorrectly--that folks like me, who frequent BR and are dorky enough to fall into the rabbit hole of the EloRater, don't give a rat's ass about steroids. Just looking at the numbers, you have to put both those guys in their respective top 5s.
March 28th, 2011 at 10:21 pm
Why is the all-time hits leader 56th and how odd is it that he and shoeless Joe are so close together? Must be the cheater section of the list.
March 28th, 2011 at 11:22 pm
Bonds isn't going to jail.
Lincecum's (and any pitcher's) W-L is meaningless. Did you miss the whole Felix Hernandez winning the Cy Young last year thing?
March 28th, 2011 at 11:22 pm
Roy Halladay is a mighty fine pitcher, but c'mon, people. Pedro Martinez and Bob Feller should be in the conversation for best right-hander of all-time. I don't know if anybody has ever pitched better than Pedro did in the late '90s and early '00s, and Rapid Robert is one of a kind.
March 28th, 2011 at 11:23 pm
Albert7, the W-L% he's talking about has to do with the head-to-head ratings in the EloRater.
March 28th, 2011 at 11:31 pm
Apparently, in three days, Derek Jeter has gone from #1116 (look back at Steve's post from the 25th, down in the comments) to #52. That seems somewhat stark. How did that happen? Don't get me wrong - I'm glad to see that it did. I just don't understand how a jump of that magnitude can be made so quickly.
I understand that Halladay's great, but right now, his career totals are worse than Pedro. And his best seasons are worse than Pedro's. So WHY is he ahead of Pedro? (#4, I wrote this while you were posting, but just decided to go ahead and post it anyway, because I agree.)
That's maybe being too picky, but as Andy points out above, there's a definite slant towards current players, and it just doesn't sit right with me.
March 28th, 2011 at 11:36 pm
Also, Andy, the more I've thought about it, who is Manny Ramirez really behind? Well, among his contemporaries, he's behind:
Bonds
Pujols
A. Rodriguez
Griffey
C. Jones
F. Thomas
Bagwell
Frankly, that's pretty good. Actually, I'd consider Jeter and Biggio above him, as well. He's been just an awful defensive player, although great with the bat. Is he really THAT much better than Gary Sheffield (about whom I'm still waiting on a Hall of Fame post since you mentioned it, wink wink)? I'm just fine with Manny.
March 29th, 2011 at 12:15 am
The Clemens rating is just whacko. Clemens can be directly compared to his one-time teammate, Tom Seaver, and the comparison isn't even close, but Seaver is in the top 10, and Clemens is not. Somebody has a serious issue with steroids leading to unjustifiable ratings.
March 29th, 2011 at 12:52 am
As long as people vote against Clemens and Bonds (and maybe some others) because they're "jerks," the whole thing is a joke.
For a while there, Mays wasn't even cracking the top 10.
March 29th, 2011 at 1:17 am
@6, Dr. Doom: Somebody fixed the Derek Jeter problem today. I sent Andy an email pointing out that Jeter was ranked so low and asking if it could possibly be that there are so many Jeter haters in the world.
Suddenly, he is in the top 60. I think they found a problem or a hacker or something. I'm sure Andy will post something to explain what happened.
March 29th, 2011 at 1:28 am
I just wish the system wasn't automatically set to start with actives every time. Actually gives current players a bit more of a bias in the system.
March 29th, 2011 at 1:35 am
The fact that Honus Wagner is not in the top 10 (most likely becaue he played so long ago that casual fans don't know who he was) means this list is more about popularity than ability.
March 29th, 2011 at 1:50 am
@12 Ron: The current top ten are, all of whom started playing over 50 years ago:
Current Top Ten (Batters):
1. Babe Ruth (2732)
2. Willie Mays (2673)
3. Lou Gehrig (2670)
4. Ty Cobb (2664)
5. Ted Williams (2662)
6. Mickey Mantle (2655)
7. Hank Aaron (2644)
8. Stan Musial (2639)
9. Rogers Hornsby (2629)
10. Jimmie Foxx (2612)
Wagner is 13. Who would you kick out to Wagner in?
Maybe Mantle? Foxx?
March 29th, 2011 at 1:56 am
@12
I agree... They should rename it the "EmoRater."
I do disagree with people having a problem with Clemens and Bonds being left out of the top ten. They do not belong there. Their numbers may be better than some of the players being rated ahead of them, but this is how baseball fans compensate for their PED use. It is sort of how we compensate for the Coors Field effect when rating players... except playing in Coors Field is not illegal.
March 29th, 2011 at 1:59 am
...and the players who get an advantage by playing in Coors Field do not say, "I never played in Coors Field." or "If I played in Coors Field it was without my knowledge."
March 29th, 2011 at 2:36 am
Bonds in jail. I couldn't get past that thought. I thought the idea of jail was to lock up people who are dangerous to society? Its just ludicrous.
March 29th, 2011 at 6:05 am
@14: Steroids weren't against the rules at the time. No, being on the commissioner's list of banned substances doesn't count; that had no actual force in baseball since the players didn't agree to the work conditions. Plus, the whole league was on steroids, so penalizing a few is silly; their numbers are against other steroid users.
March 29th, 2011 at 6:12 am
Sean told me the Jeter issue was a bug that has now been fixed, and it was not anything nefarious. (There isn't really a way to hack or stuff the ballot box on this).
I do believe Bonds is going to jail. The judicial system takes grand jury testimony very seriously. If people who committed perjury are allowed to walk, then the whole system breaks down. There are plenty of examples of people who didn't commit any serious crime but did go to jail over perjury in this sort of situation. Scooter Libby comes to mind. I suspect Bonds will receive at least a token jail sentence.
March 29th, 2011 at 9:04 am
Taking for granted that Libby didn't commit any crimes that before perjury, that the perjury was covering up (not even having criminal knowledge of a conspiracy in which he may not have been the leader) is generous, to say the least.
March 29th, 2011 at 9:18 am
I think Andy's problem with Mantle stems from (or at least is in accord with) what I think is an unfortunate notion of "peak". When looking beyond career totals, what I want to know is how good was the player at his best. Not how much did he combine luck with skill to prduce a couple of phenominal seasons a decade aprt from each other. I want to know (as best as we can tell) if you could magically add a year in the middle of his career, how well could we *expect* the player to perform,. based on skill alone. By that criteria, Mantle is one of the best players of all time. #4? Well, peak isn't the whole story, but if it were, then #4 would be about right. In any case, I think I'd rate him above Gehrig.
March 29th, 2011 at 9:34 am
This thing is purely subjective, which makes it a mere popularity contest to me. It's one of those fun little things that br.com does and can't be taken all that seriously. It's just to spur debate, which it's doing. But who's first, second or 1679th doesn't mean anything in the real world.
By the way, Williams should be #1.
March 29th, 2011 at 10:29 am
I tend to agree that Williams should be higher. In fact I'd probably put Ruth at #1, Williams at #2, and Bonds at #3.
March 29th, 2011 at 10:35 am
@12, @13 - I would put Honus Wagner ahead of everyone except Ruth and Mays. I can see a good argument for Cobb, and weak arguments for Williams, Mantle, Bonds, and Aaron (only if you mostly ignore defensive value). For everyone else, I see no argument at all.
I know that (for example) Stan Musial has better lifetime offensive totals, but the only way I could see anyone rating Musial over Wagner is to:
-mostly ignore the offensive context
-ignore the difference in position
-ignore the defensive value.
I'm not knocking Musial; he's certainly one of the all-time greats and belongs in the Top Dozen, just not ahead of Wagner.
One of the real fun things abour EloRater is the real-time change in the Top-10. I voted for Cy Young over Steve Carlton, and _instantly_ saw Young go to #1, by one point over Walter Johnson, and also Carlton drop OUT of the Top-10 pitchers. I can't tell you how cool that was...
I hope this thread does not devolve into a steriod discussion, we've had enough of those...
March 29th, 2011 at 10:41 am
I notice that if you sort by all the active batters by "#" (times rated) they are ranked 1 through 224. Retired players are rankled 225-1726. Every active batter has been rated over 300 times, but no retired batter has been rated more than 250.
The distribution for pitchers is not as perfect, but still very similar.
Does the randomizer favor active players, or is this a by-product of allowing users to skip an opportunity to make a choice?
March 29th, 2011 at 10:43 am
#24 I mentioned the same at the end of my original post. I presume it's because people skip older players.
March 29th, 2011 at 10:48 am
Will active players be added during the season if they cross the
# 1,100 Innings Pitched
# 3,000 Games + Plate Appearances
thresholds?
Will a player who has not crossed these thresholds be deleted if their WAR drops to 7.9 or lower?
If players are added, will they begin at the bottom or at a default rating towards the center?
March 29th, 2011 at 10:55 am
Andy is there a way to include a players ranking twice? I am wondering if the current ranking could somehow be pinned to the player after a sort.
i.e. Currently Vlad Guerrero is ranked as the 50th best player. Could that #50 stay with Guerrero as we sort for "Wins" "Losses" or other categories?
March 29th, 2011 at 10:57 am
I hope this thread does not devolve into a steriod discussion, we've had enough of those...
I've got bad news, Larry, this thread started as discussion about steroids, which of course is in no way as pressing, current or important as the relative merits of Honus Wagner and Stan Musial.
It's always fun when someone walks into a room 30 minutes after a conversation has started.
March 29th, 2011 at 10:57 am
#27 I don't think so. Your best bet is to put the data into Excel and then sort there.
March 29th, 2011 at 10:57 am
@25 Sorry I missed the original post Andy, I guess I have been too busy watching my my NCAA brackets disintegrate to put the necessary time into baseball.
I promise to be a better fan over the coming months.
March 29th, 2011 at 11:07 am
No specific observations of value- just want to say that the Elo Rater is an excellent time-waster and highly entrtaining. Thanks!
March 29th, 2011 at 11:12 am
In response to my own original question and #24, the powers that be tell me:
"When you first go to Elo.cgi, it automatically matches 2 actives. Also, if you skip, it automatically gives 2 actives"
March 29th, 2011 at 11:21 am
@17
Steroids were not against the rules at the time in baseball, so Bonds and Clemens should rate higher? They were against the law. Whether or not there was a rule against it in baseball does not matter. Back when all this got started, no one would have said that it would be OK for players to gain a competitive advantage by breaking the law.
Like I mentioned, playing in Coors Field was not against baseball rules either. And yet we compensate for that when we rank players who put up huge numbers there. And we should do that. The whole reason we do not need asterisks in the record book is because baseball fans are savvy enough to factor in things like lengthened schedules, grossly advantageous park effects, and players like Bonds, Clemens, Arod, and Manny – who cheat.
You say that the whole league was on steroids. The whole league was not on steroids. You say penalizing a few is silly. What few are you talking about? You are really worried about penalizing Clemens and Bonds when you should be worried about players for whom there is no evidence of PED use. Greg Maddux and Ken Griffey Jr. come to mind. The ones who did not use PEDs are the ones getting penalized here if we just decide to accept all of the numbers at face value.
I don’t know for sure if Maddux ever used PEDs or not. But there is evidence that Clemens did. And when confronted with this evidence, he has continually lied about it. I for one second will not rate Clemens ahead of Maddux given what we know about both players.
March 29th, 2011 at 11:31 am
I got tired of this because I don't want to rank active players since they are still active, but every time I skip active players it brings up more active players so I have to end up ranking some guy with 5 seasons before I get to a retired player and then it goes back to active guys pretty quickly. Very annoying so I stopped. I think this should not be set up to try to suck in the average fan who only watched ESPN and is only aware of active players and Babe Ruth. If someone does not know who Honus Wagner is they should not be voting on baseball players.
Also, you should start over and set this up without any names to try to reduce the popularity contest.
March 29th, 2011 at 11:56 am
@28/ Scott Ross - Yes, I did see your comment at the very start, but that does not mean we need to pulverize to dust the same old tired discussions about steriods, both defending and penalizing players, with no conclusion drawn.
I know that you are joking, but I for one _would_ much rather discuss the relative merits of Honus Wagner and Stan Musial.
{from the intro}:
"Entries 61 through 80 are all interesting. Many of these guys are fringe HOFers ... Willie Stargell, Bill Dickey,... Luke Appling..."
Andy, I wouldn't consider Stargell, Dickey, and Appling marginal HOFers. Perhaps by the standards of BBWAA inductees, they might be considered a little below average for their position, but they are all comfortably within the top dozen all-time for their positions (Dickey higher).
EloRater is a wonderful idea, and the rating of players can get quite addictive.
March 29th, 2011 at 12:13 pm
Lawrence, on other threads at least a couple of readers have voiced the opinion that Pops didn't deserve enshrinement. I didn't mean to suggest that those 3 guys didn't deserve it, just that they weren't necessarily slam dunks compared to some other players.
March 29th, 2011 at 12:21 pm
As someone who doesn't want to rate people when I all I know about them is the stastics, and as someone who no longer follows baseball but know a helluva lot of history, I find the skew toward active players very discouraging. I'd put in a lot mroe ratings where I *do* have an informed opinion if I didn't have to put in even more where I don't.
March 29th, 2011 at 1:01 pm
@8 Charles Saeger: The difference between Seaver and Clemens is one of sportsmanship and class, and you're right, there's no comparison and Clemens doesn't deserve mention in the same breath as Seaver. Tom Terrific didn't go around beaning and throwing broken bats at hitters who owned him, ESPECIALLY in the World Series! Plain and simple, Clemens was an a-hole with roid rage.
March 29th, 2011 at 1:05 pm
@36/ Andy Says: "Lawrence, on other threads at least a couple of readers have voiced the opinion that Pops didn't deserve enshrinement. I didn't mean to suggest that those 3 guys didn't deserve it, just that they weren't necessarily slam dunks compared to some other players."
Andy, fair point, and I do recall that "Stargell: HOF-qualified?" discussion. It's just that when I hear the phrase "marginal HOFer", I think of the likes of Jim Rice, Rick Ferrell, and Dave Bancroft; not Stargell, Dickey, and Appling.
March 29th, 2011 at 1:20 pm
@21 Larry: It's more than a fun little thing. I bet it's the greatest idea that BBR has come up with to increase ad revenue!
Every time we vote or skip a vote, another ad gets served up, and the cash register goes, "cha-ching". There have been times that I must have looked at dozens of pages I would not have looked at because of the feature.
I'm not complaining, by the way. I think it's great they thought of the idea and will make more money. I'm sure some of it will get turned into content.
Keep up the good work.
March 29th, 2011 at 1:29 pm
I am not privy to the revenue details of the site, but I don't think anybody's getting rich off of ad revenue specifically. I don't think that's what motivates Sean to add new features like this.
March 29th, 2011 at 1:41 pm
I wonder if there's a way for Sean (or whoever else might be a 'powers that be' person) to customize the rater so that the default setting is as it currently is (ie. skewed towards active players to not make it daunting for the causal fan), but with the option to customize it to make it a series of truly random matchups. I agree with the sentiment expressed above that it'd be great to be able to skip matchups without being redirected to current players, as I have the same problems with rating someone who has played for 5 years only.
March 29th, 2011 at 2:07 pm
@20 Dvd Avins: Is it possible that when you look at peak value, you are underrating Gehrig? After all, he played during the same time as Ruth, and there were at least three seasons that Gehrig finished second right behind Ruth in WAR. In several other categories over a four or five year period, they were 1-2.
I guess what I am really asking is that if Gehrig had done what he did, but there was not Babe Ruth, would his peak value have been higher?
Or, would Lou Gehrig have been the Babe Ruth of his generation if not for Babe Ruth ? 😉
March 29th, 2011 at 2:32 pm
@33: Baseball isn't the arbiter of the law. To take this to the Nth degree, are you willing to rate O.J. Simpson as less of a football player because he killed two folks and kidnapped another? This is how they rate as baseball players. It would make more sense to fret about corked bats or spitballs, which were unabiguously against the rules when Babe Ruth and Gaylord Parry used them. And, let's face it, steroids never turned a bad player into a great player.
@38: Again, Seaver is a better person than Clemens, but not a better pitcher, and the evidence is absolute. Clemens had about 150 more innings pitched, and finished with about 40 more wins, 20 fewer losses, a better ERA in context, over a thousand more strikeouts and fewer home runs allowed despite pitching in a homer-happy era. If you're using this rater to say how much you like the guy personally, you're abusing it for the rest of us. Are you knocking down Cobb for being an even bigger jerk than Clemens, Bonds and Carl Everett combined?
@18: We do tend, as a society, to place a great deal of seriousness to the idea that "thou shalt not lie on the witness stand." There's some great testimony in Helter Skelter wherein one of Manson's followers says she would give her life for Manson but would tell the truth on the witness stand even if her telling the truth would send Manson to the electric chair. It's like adding sex to a crime; we go from righteous indignation to pure overdrive spazzing out.
March 29th, 2011 at 2:48 pm
@43/... Joseph Says: "... I guess what I am really asking is that if Gehrig had done what he did, but there was not Babe Ruth, would his peak value have been higher? Or, would Lou Gehrig have been the Babe Ruth of his generation if not for Babe Ruth ? "
Joseph, it's more like Gehrig being the Rogers Hornsby of his generation; Ruth cast such a large shadow, that not even all-time great contemporaries such as Gehrig and Hornsby could completely escape from it.
March 29th, 2011 at 3:18 pm
@44
OJ's alleged murder and kidnapping did NOT help him gain a competitive advantage. If kidnapping and murder inflated his 2003 yard season to 2,312 yards then your bringing him into the conversation would make sense.
I did not want to turn this into a debate about steroids. I simply was responding to the early posts on here, including yours, that implied that there was no legitimate reason to rank Bonds and Clemens out of the top 10. There IS a legitimate reason to do so. If you want to rank both players in the top 5 that is fine. But do not wonder out loud the why bulk of voters have not. There is a good reason.
March 29th, 2011 at 4:16 pm
@3
WP for a career does have some value. Wins (and WP) for a season are meaningless.
Much of what makes wins so meaningless falls away over multiple seasons. As teams change, so to do the conditions under which a pitcher accumulates his wins and losses.
As to Lincecum, his career may be too short (so far) to really get much value from his career winning percentage, but it still does say something about him.
March 29th, 2011 at 4:18 pm
I have a question about how the Elo works. I've played with it more over at the b-ball site, but I assume it works the same. In my experience, guys had a bigger shift in their number if they were "upset". That is, if I picked a guy who was rated significantly lower than another, both saw big changes in their rating. If they were close together, then regardless of who I picked, the change seemed moderate. If they were far apart and I picked the leader, it was minimal. Doesn't this make it easier to knock top guys down then boost them back up? How will Ruth ever pick up more than a few points since he'll be rated higher than everyone he faces? Am I misunderstanding it?
Also, is it possible to see head-to-head W/L? For instance, can I see how Ted Williams fares against Joe DiMaggio? I doubt it, but that would be awesome!
March 29th, 2011 at 4:20 pm
@44
I think you're off-base on your steroids talk, but not every discussion needs to be about steroids, so...
Your point about voting down a player based on whether or not he is likeable is "abusing it for the rest of us" is silly. People can vote however they like. Do you honestly think people are objective when they cast these votes?
If they did, Bonds and Clemens would easily be in the top 10, and Mickey Mantle would not be in the top 5.
March 29th, 2011 at 4:22 pm
Also, any reason you don't compare pitchers and position players?
March 29th, 2011 at 4:25 pm
Ingres-
While I agree that, technically, there isn't a wrong way to vote, I definitely think there are wronger ways to vote and there is valid criticism for those which are.
If I value peak over longevity and you go the opposite way, it'd be hard to objectively demonstrate that one approach is more valid. If you value peak over longevity and I value blue eyes over green eyes, I think it is fair to say that my approach is far less valid and, as such, diminishes the experience for the rest of us. I don't think character matters, unless it somehow had an impact on the field. If a sizable enough portion of the voting public here was basing their votes on things like that, to the point that it skewed the results, I'd be less likely to participate.
March 29th, 2011 at 4:26 pm
BTW, not sure if it is deliberate or a function of my slower connection speed at work, but the matchup shows up about a half-tick before the rest of the page loads. This gives me JUST enough time to formulate a gut reaction before seeing the numbers. Fun!
March 29th, 2011 at 5:23 pm
The concept of the EloRater comes from crowdsourcing and crowd psychology techniques. Many of these can be found in James Surowiecki's book, The Wisdom of Crowds. The entire point of this kind of collaborative thinking is that the "formula" doesn't come from a single source.
The EloRater provides an alternative to the formula-centric culture of analyzing baseball players, where a formula attempts to quantify that Player X is better than Player Y by generating a number (or numbers) that supposedly proves worth of a player within a certain condition.
None of these formula, such as WAR, VORP, Win Shares, Linear Weights, etc. are infallible, and neither is the EloRater approach. In some cases, the collaborative thinking method can arrive at better results than a single or small group of experts working on a problem.
March 29th, 2011 at 7:29 pm
As much as it's frustrating for A-Rod, Bonds, and Clemens to get downgraded, it is refreshing that casual fans are voting, as opposed to "stat heads" dominating the polls.
March 29th, 2011 at 7:34 pm
Oh, one more thing- I really wish I could run this tool as a player comparison from the player page one day, or perhaps make it a player comp tool page by itself, so I could select my two players.
March 29th, 2011 at 9:02 pm
@55
I second that. I'd love to see the chart for Gehrig vs. Foxx or Mantle vs. Mays but it would take days of clicking before those would pop up randomly.
March 29th, 2011 at 9:26 pm
@18
Fair enough. But maybe Bonds should run for office. Apparently if you are a politician and lie to Congress you are off the hook.
March 29th, 2011 at 9:37 pm
@56
Agreed, but it would likely take forever for Mays v. Mantle to come up. Or almost forever for Mays v. anybody -- he's tied for the 11th fewest matchups of anyone on the hitter side. I've done hundreds of these and no luck so far on Mays, my favorite player.
Walter Johnson is equally low on the pitcher side. It's good for both of them they do so well in matchups.
Funniest matchup I've seen so far, the polar opposites: Rafael Furcal vs. Babe Herman.
And the most agonizing: Eddie Murray vs. Cal Ripken. Hard to choose between 2 other favorites.
March 29th, 2011 at 9:38 pm
@53 - the problem with the approach here is that there is no incentive mechanism in place when people vote. If you ask a whole bunch of people how many jelly beans are in a jar (or how much a bull weighs, etc.) and if they come closest to the actual amount they win a prize, then there is an incentive to reveal their best guess as to how many jelly beans are in the jar.
But if I get a match up of Jose Lind and Willie Mays, and I rank Willie Mays ahead of Lind or Lind ahead of Mays, what is my payoff? There is no incentive for me to act truthfully (even if my information is wrong - I'm still revealing what I know) and reveal the information that I know using the EloRater. And you need people to have the incentive to act truthfully in order for the wisdom of crowds to work. Otherwise the aggregation process is aggregating a bunch of nonsense.
It's not that I don't think the wisdom of crowds works - I use it all the time when playing college/pro football pick 'em leagues (Yahoo! reports the percentage of people who have picked a particular team) with friends and I even used it as a guide when filling out my NCAA brackets this past year. But the people in those "crowds" have some incentive to make picks according to their information - they want to win their leagues. With the EloRater, there is no such incentive, and so I will take the stats based approach over the wisdom of crowds approach in this instance.
Not that it's not fun to rate the players - I, like some of the others have mentioned, prefer to rate older players because that's how I can learn about those players.
March 29th, 2011 at 9:44 pm
@55, 56, and 58 - The problem with allowing people to pick their matchups is that then a hacker could really come in and destroy the rankings, if those matchups were counted in the overall tally.
As I mention in post 59, there is no incentive for anyone to act truthfully. So if I despise Derek Jeter all I have to do is pick Jeter vs. Rick Burleson and vote for Burleson. Or Jeter vs. Furcal and vote for Furcal. Or Jeter vs. Keith Miller and vote for Miller. Or Jeter vs. Knoblauch and vote for Knoblauch. I'm guessing anyone with a decent amount of programming knowledge and a little bit of free time could write code to do that.
March 29th, 2011 at 9:56 pm
People CAN vote however they want. If people want to pick David Eckstein over A-Rod because they think the latter is a jerk or 'cause he used steroids, they're welcome to.
But that still makes the whole thing a joke when it stops being about baseball.
Bonds and Clemens were both HoFers and Top 10 players/pitchers before anyone even accuses them of doing steroids, so whatever.
I hope people stop voting for Mays, Aaron, etc. cause they used greenies to keep themselves going too then!
March 29th, 2011 at 9:57 pm
Excellent discussion about crowd logic, folks. I've really enjoyed reading the comments on this thread.
March 29th, 2011 at 9:59 pm
I'm also gonna stop voting for people who have had Tommy John. COMPETITIVE ADVANTAGE. They should've all had to quit like the good ol' days.
AND they throw harder when they come back! It's almost like they're ROBOTS.
March 29th, 2011 at 10:10 pm
@44: Clemens was only "better" than Seaver because he used 'roids. He wouldn't have pitched the extra 150 innings or won the extra 40 games without drugs. And if there was any doubt he used 'roids, it was erased by incidents like his bat-throwing stunt in game 2 of the 2000 World Series. I was at the game, and I think all that saved him from being ejected was that it was at Yankee Stadium and the umpires feared it might cause a riot. And don't forget how he threw at Piazza's head and beaned him just before the all-star break in July, knocking Piazza out of the all-star game. What a hot-headed, unprofessional a-hole!
March 29th, 2011 at 10:13 pm
I don't even like Roger Clemens, but that's utter bullcrap.
Do you have ANY proof that Clemens wouldn't have done that? Can you guarantee he wouldn't have recovered from whatever injuries he used 'roids for and still done it?
No, of course not. Steroids help people heal. So does surgery. If your issue is that steroids are illegal, then take up with you the government, not baseball officials who aren't the police.
March 29th, 2011 at 10:36 pm
@60
55, 56 and I (58) weren't suggesting we should be able to call up players at will within the exercise. It would spoil the fun and, as you said, somebody could game the system.
It would be cool if, outside the rater, there was another page where you could display players you choose in the same format.
I do think some people here are taking this too seriously. Did they notice the EloRater is indexed under "Frivolities"?
March 30th, 2011 at 8:19 am
@crpls #65
I don;t have any "proof" for any of my picks. I vote subjectively based upon my observations, memories, gut-feeling and the stats offered.
And should I be given the choice of a Seaver-Clemens match-up, I will most likely choose Seaver.
March 30th, 2011 at 10:44 am
@63/Crpls Says: "I'm also gonna stop voting for people who have had Tommy John. COMPETITIVE ADVANTAGE. They should've all had to quit like the good ol' days. AND they throw harder when they come back! It's almost like they're ROBOTS."
Crpls, wait till they have the great debate in 2058 as to whether robots should play in MLB - we'll forget all about our current steriod arguments!
March 30th, 2011 at 11:27 am
@68
What debate?? Robots in the AL, No robots in the NL, And during inter-league robots can play in AL parks.
Commissioner Wendy Selig wouldn't have it any other way.
March 30th, 2011 at 1:08 pm
@69 - Spartan Bill... LOL, glad that you are not taking this _too_ seriously.
March 31st, 2011 at 10:14 am
EloRater is ideal for stuff like this. I don't know any statistical method that can identify the impact of steroid use.
We know of some players that used PEDs and some people have opinions one way as to how much it affected them, some people have completely different opinions. No side can claim to be definitively correct on how much effect x PED had on a person's career.
In terms of a player's total subjective worth, EloRater will certainly be closer than any formula trying to weed out PEDs. If 80% of the raters impose harsh penalties on a player because they took PEDs, that's probably going to be reflective of the general populous. If only 10% impose harsh penalties, then I would imagine that would be reflective of the general populous.
In terms of absolute worth, it may be better as well. PEDs are far too "squishy" an issue to tackle via raw data analysis.
March 31st, 2011 at 10:38 pm
@43 I give Ruth extra credit for changing the game. Not only the way it was played on the field, but for popularizing it beyond what it had been. The sportswriters did their share, but Ruth was the icon and, for better and worse, bis being the center of attention for so long didn't change him in ways that made him less interesting. I also downgrade for level of competition a little. Ruth takes that hit and remains #1. Gehrig slips a few spots, but is still the best first baseman ever.
April 2nd, 2011 at 11:17 am
@66 That's right, Don. I just wanted the player comp format so I could easily list two players. That would really enhance baseball-ref if it would be implemented.
***
As far as this being a frivolity, I agree that it will always be unscientific because there are fans who will vote on popularity/infamy, their hometown favorites, WAR total only, or be biased against players they never heard of. And that's fine, since this is open to the public.
However, I do think the rankings do are significant and will continue to be refined as more people play this game.
***
Only one complaint- sometimes the advertisements are intrusive.
April 2nd, 2011 at 7:12 pm
If anyone is still reading this (and if anyone cares), I just went to elo-heaven. I've played around with it a bit here and there, going through 10 matchups or so when I have time, and it's always exciting when I get to compare some incredibly high-ranking players, although I enjoy the whole randomness.
But seriously, I had the matchup of Skeeter Newsome vs. Aaron Miles (both in the bottom 15 players of 1726), and then it pops up: 'who's better: Babe Ruth or Ted Williams'? Goddamn. I had to look at those insane numbers in the 'tale of the tape' for a long, long time.
This is a fun, fun time waster.
April 2nd, 2011 at 8:18 pm
I'm still reading...glad you're enjoying yourself!
April 6th, 2011 at 12:23 am
I have a question about how the Elo works. I've played with it more over at the b-ball site, but I assume it works the same. In my experience, guys had a bigger shift in their number if they were "upset". That is, if I picked a guy who was rated significantly lower than another, both saw big changes in their rating. If they were close together, then regardless of who I picked, the change seemed moderate. If they were far apart and I picked the leader, it was minimal. Doesn't this make it easier to knock top guys down then boost them back up? How will Ruth ever pick up more than a few points since he'll be rated higher than everyone he faces? Am I misunderstanding it?
I partially skimmed this but I don't think it was answered (and maybe no one's reading it anymore anyway). If Ruth is ranked higher than (almost) everyone, no, he won't get more than a couple points at a time. But presumably he will win most matchups (that's why he's at the top), so he's getting those couple points almost every time.
Related, someone referred to picking Jose Lind over Willie Mays. Assuming this is like the hoops one, that's not going to happen. Matchups are chosen among players within X points of each other. So Mays never faces Lind, he faces Ruth or Mantle or Pujols or guys like that. So you can't game the system (maybe some deranged hackers can, but not most folks). So if you think Mays was better than Ruth, fine, pick Mays and he gains 5 (or so) points. Or go vice versa and Ruth gains 2 (or so) points. But if everyone picks Mays, than Ruth will fall behind, and then when he is chosen he will gain more points.