Baseball for the Thinking Fan

Login | Register | Feedback

btf_logo
You are here > Home > Baseball Newsstand > Baseball Primer Newsblog > Discussion
Baseball Primer Newsblog
— The Best News Links from the Baseball Newsstand

Saturday, August 23, 2014

MGL: Which teams are optimizing their lineups?

New order: Everyone everywhere.

Let’s recap the mistakes that managers typically make in constructing what they think are the best possible lineups. Again, we will ignore player preferences and other “psychological factors” not because they are unimportant, but because we don’t know know when a manager might slot a player in a position that even he doesn’t think is optimal in deference to that player. The fact that managers constantly monkey with lineups anyway suggests that player preferences are not that much of a factor. Additionally, more often than not I think, we hear players say things like, “My job is to hit as well as I can wherever the manager puts me in the lineup.” Again, that is not to say that some players don’t have certain preferences and that managers shouldn’t give some, if not complete, deference to them, especially with veteran players. In other words, an analyst advising a team or manager should suggest an optimal lineup taking into consideration player preferences. No credible analyst is going to say (or at least they shouldn’t), “I don’t care where Jeter is comfortable hitting or where he wants to hit, he should bat 8th!”

Managers typically follow the traditonal batting order philosophy which is to bat your best hitter 3rd, your slugger 4th, and fast, scrappy, good-bat handlers 1 or 2, whether they are good overall hitters or not. This is not nearly the same as an optimal batting order, based on extensive computer and mathematical research, which suggest that your best hitter should bat 2 or 4, and that you need to put your worst hitters at the bottom of the order in order to limit the number of PA they get per game and per season. Probably the biggest and most pervasive mistake that managers make is slotting terrible hitters at the top, especially in the 2-hole. Managers also put too many base stealers in front of power hitters and hitters who are prone to the GDP in the 3 hole.

Finally, managers pay too much attention (they should pay none) to short term and seasonal performance as well as specific batter/pitcher past results when constructing their batting orders. In general, your batting order versus lefty and righty starting pitchers should rarely change, other than when substituting/resting players, or occasionally when player projections significantly change, in order to suit certain ballparks or weather conditions, or extreme ground ball or fly ball opposing pitchers (and perhaps according to the opposing team’s defense). Other than L/R platoon considerations (and avoiding batting consecutive lefties if possible), most of these other considerations (G/F, park, etc.) are marginal at best.

...Looking at all these “optimal” lineups, the trend is pretty clear. Bat your best hitters at the top and your worst at the bottom, and do NOT put a scrappy, no-hit batter in the two hole! The average projected linear weights per 150 games for the number two hitter in our 4 best actual lineups is 19.25 runs. The average 2-hole hitter in our 4 worst lineups is -20 runs. That should tell you just about everything you need to know about lineups construction.

Repoz Posted: August 23, 2014 at 09:29 AM | 73 comment(s) Login to Bookmark
  Tags: sabermetrics

Reader Comments and Retorts

Go to end of page

Statements posted here are those of our readers and do not represent the BaseballThinkFactory. Names are provided by the poster and are not verified. We ask that posters follow our submission policy. Please report any inappropriate comments.

   1. JJ1986 Posted: August 23, 2014 at 09:51 AM (#4777657)
I'm surprised the Brewers don't rate as more optimal. They seem to be the one team who completely ignores the stereotypical #2 hitter and put one of their best guys there.
   2. bobm Posted: August 23, 2014 at 10:07 AM (#4777661)
The 4 least optimal lineups would gain 5-12 runs per 150 games under this system. Why risk pissing off your players for at most 1 win in theory?
   3. The District Attorney Posted: August 23, 2014 at 10:08 AM (#4777662)
I'm surprised the Brewers don't rate as more optimal. They seem to be the one team who completely ignores the stereotypical #2 hitter and put one of their best guys there.
The Angels hit a decent guy 2nd. ;)
   4. PreservedFish Posted: August 23, 2014 at 11:09 AM (#4777679)
I LOVE that this article is a zillion words and buried in the middle is this:

The kicker, however, is that the “optimal” lineups, on the average, only slightly outperformed the actual ones, by only 2/3 of a run per team. Essentially there was no difference between the lineups chosen by the managers and ones that were “optimized” according to the simple rules explained above.


So, basically, we can just ignore everything else?
   5. Cargo Cultist Posted: August 23, 2014 at 11:25 AM (#4777684)
It boils down to this: as a manager, do you want every edge that you can get, or not?
   6. BDC Posted: August 23, 2014 at 11:49 AM (#4777696)
The Rangers aren't among the best or worst, but they show one common feature of bad lineups: they bat a scrappy no-hit guy second AND he is one of their better hitters :)
   7. Ziggy Posted: August 23, 2014 at 12:05 PM (#4777706)
I think the answer to #2 is: because an extra win is worth several million dollars.
   8. McCoy Posted: August 23, 2014 at 12:14 PM (#4777708)
So if the Cubs add one more win to their total this year they'll make several million more dollars? If the Royals had won 87 games last year instead of 86 they would have made several million more dollars?
   9. Misirlou's been working for the drug squad Posted: August 23, 2014 at 12:23 PM (#4777712)
So if the Cubs add one more win to their total this year they'll make several million more dollars? If the Royals had won 87 games last year instead of 86 they would have made several million more dollars?


All else being equal, how much more (on average) does a FA projecting to 4 WAR make vs a guy projecting to 3 WAR?
   10. Ray (RDP) Posted: August 23, 2014 at 12:32 PM (#4777715)
Odd that he uses "lineup" without talking about pitchers.
   11. Misirlou's been working for the drug squad Posted: August 23, 2014 at 12:34 PM (#4777716)
And to address your idiocy directly, 1 more win would have made a huge difference to the 2013 Rangers, the 2012 Rangers, the 2011 Red Sox and Braves, the 2010 Padres, and the 2008 Mets. IOW, on average one team every year for the past 6 years.
   12. tshipman Posted: August 23, 2014 at 12:59 PM (#4777727)
Isn't it odd that he finds three lineups in the NL (LAD, SFG, WAS) that are outperforming his sim by a significant margin?

Wouldn't that indicate that some of his premises are wrong?
   13. JustDan Posted: August 23, 2014 at 01:28 PM (#4777736)
Maybe, but the Washington lineup he posted doesn't include Desmond, who has played more games than anyone else for the Nationals. So I'm not sure what he is doing.
   14. PreservedFish Posted: August 23, 2014 at 01:46 PM (#4777742)
Also, "here are the top 5 and bottom 5 teams," followed by a list of 6 teams.

He also has the Mariners as outperforming his optimized lineup despite the fact that they had mediocre hitters hitting at #2 and #4.
   15. Ziggy Posted: August 23, 2014 at 01:54 PM (#4777744)
He addresses these points in the post. The "optimal" line-up, he says, doesn't factor in interaction effects. Some players' skills complement each other, and when they do, line-ups that have players who complement each other batting in positions in which they *will* complement each other can out-perform a line-up that would be optimal without interaction effects.
   16. PreservedFish Posted: August 23, 2014 at 02:08 PM (#4777748)
The "optimal" line-up, he says, doesn't factor in interaction effects. Some players' skills complement each other, and when they do, line-ups that have players who complement each other batting in positions in which they *will* complement each other can out-perform a line-up that would be optimal without interaction effects.


Seriously? I'm calling it a sample size quirk.
   17. Ziggy Posted: August 23, 2014 at 02:32 PM (#4777751)
It could be, but that's the explanation that MGL offers.
   18. McCoy Posted: August 23, 2014 at 02:41 PM (#4777752)
Maybe, but the Washington lineup he posted doesn't include Desmond, who has played more games than anyone else for the Nationals. So I'm not sure what he is doing.

How does one go about doing this study anyway? It would seem like there are way too many variables at play and that it would be virtually impossible to come up with a realistic number for each variable to pull off this kind of study with any degree of accuracy that anyone would find usable.
   19. bobm Posted: August 23, 2014 at 02:47 PM (#4777754)
No one is saying that 1 win isn't valuable. Rather, the question is: does optimizing the lineup do more harm than good if the players don't or won't buy in? Actual managers factor this concern in.

NYT (2010): Tony La Russa and the logic behind the lineup

Jerry Manuel, the Mets’ manager, said it [batting the pitcher 8th] makes sense for La Russa because of the overwhelming offensive presence of Albert Pujols in the No. 3 spot. In 1998-99, Mark McGwire was a similar presence for La Russa. By using a position player ninth, La Russa ensures that Pujols bats in the first inning and then has three hitters in front of him in subsequent at-bats. [...]

Some of La Russa’s innovations over 32 seasons as a major league manager have been widely adopted by others. His use of specialty relievers and a one-inning closer (an idea for which he credits the pitching coach Dave Duncan) is now common. The so-called Crazy-8’s lineup is not.

“In total, I like it,” said La Russa, who has a theory on why more teams don’t copy this strategy. “Well, it’s one less distraction. You’re doing something traditional. Players don’t ever have to be asked about why the pitcher’s batting eighth. And to me, every less distraction gives your team a better chance to be itself. But we’ve talked about it enough now that if I’m going to do it, they know what the answers are. And I usually wait till we struggle.”
   20. OCF Posted: August 23, 2014 at 02:49 PM (#4777756)
The Angels hit a decent guy 2nd. ;)

Some other historical cases of batting a "decent guy" 2nd include the 1982 Brewers and the 1996 Mariners.
   21. McCoy Posted: August 23, 2014 at 03:13 PM (#4777761)
And to address your idiocy directly, 1 more win would have made a huge difference to the 2013 Rangers, the 2012 Rangers, the 2011 Red Sox and Braves, the 2010 Padres, and the 2008 Mets. IOW, on average one team every year for the past 6 years.

My idiocy? So 1 win is valuable for one team a year and that makes me an idiot for saying that on average 1 win isn't worth millions? Um, ok.
   22. Rickey! trades in sheep and threats Posted: August 23, 2014 at 03:22 PM (#4777762)
Maybe, but the Washington lineup he posted doesn't include Desmond, who has played more games than anyone else for the Nationals. So I'm not sure what he is doing.


He also seems to argue that Phil Gosselin is a worse hitter than BJ Upton, which begs the question of whether or not he's watched either of those two play baseball this year.
   23. BDC Posted: August 23, 2014 at 04:07 PM (#4777780)
historical cases of batting a "decent guy" 2nd

I seem to remember George Brett batting second quite a bit, but it's selective attention. He did so for about a season's worth, mostly in 1975 (Jack McKeon's idea) and 1979 (Whitey Herzog's). I think I was reading box scores most obsessively in 1979 :)
   24. cardsfanboy Posted: August 23, 2014 at 04:10 PM (#4777782)
So much wrong with this article, from tone to conclusions to simplicity.

I agree with MGL, I obsess over lineups also even though, as he points out, it's relatively a minor overall difference when everything is said and done. But I obsess in some opposite points. To me a set lineup is extremely stupid. If a manager has the same lineup against the same handed pitcher three games in a row, I just think he isn't trying.

When I ran my simulations above, swapping the pitcher and the 8th hitter in the NL lineups. the resultant runs per game were around 2 runs worse (per 150) than with the traditional order. It probably depends on who the position player is at the bottom of the order and perhaps on the players at the top of the order as well.


He talks about using the pitcher in the 8th spot in the NL and he somewhat misses the point. His numbers is based upon the lineup being the same for all nine innings of the game. In that case of course the lineup with the pitcher batting eighth is going to be inferior, the two points of batting the pitcher 8th is 1. put more men on base for your best hitters(number 2, and 3...but of course his lineup construction removes the best hitter from the 3rd slot--he acknowledges that at least) and 2. in the 7th and later innings, your better bat bench players are available more frequently over the course of the season.




Of course this being MGL, he misses a lot of the human aspects of lineup construction. A player batting in front of an elite batter is less likely to be pitched around. A struggling player can enjoy the benefits of batting in front of an elite player, with the knowledge that the other teams pitcher isn't going to try and throw too many pitches way out of the strike zone.

He also doesn't once mention the word double play in the article and whether players who don't hit into a high double play opportunity should be in certain spots or behind certain players etc. I personally would have the high walk average speed guy batting first and then the high speed guy batting second just to reduce dp possibilities. And of course he completely ignores setting up a lineup to break up left and right handed batters to counteract loogy's which are used in high leverage situations.


He addresses these points in the post. The "optimal" line-up, he says, doesn't factor in interaction effects. Some players' skills complement each other, and when they do, line-ups that have players who complement each other batting in positions in which they *will* complement each other can out-perform a line-up that would be optimal without interaction effects.


Which is what most managers are trying to do when they design their lineup. The fact that he acknowledges that the optimal lineup isn't actually optimal, while writing an article that is dripping with condescension of all managers is hilarious.
   25. bobm Posted: August 23, 2014 at 04:38 PM (#4777800)
   26. Lance Reddick! Lance him! Posted: August 23, 2014 at 04:59 PM (#4777802)
He also doesn't once mention the word double play in the article and whether players who don't hit into a high double play opportunity should be in certain spots or behind certain players etc.

Searched the page for "double play" and missed every instance of "DP" and "GDP," didja?
   27. cardsfanboy Posted: August 23, 2014 at 05:09 PM (#4777804)
Searched the page for "double play" and missed every instance of "DP" and "GDP," didja?


yep.

He focuses more on dp in the third hole than the 2nd hole which is where I'm usually more concerned about.
   28. Ziggy Posted: August 23, 2014 at 05:36 PM (#4777812)
Of course causation is hard to show, but on average each marginal win is worth millions. There is a nice correlation between wins and revenue. That's why teams try to win games.*

(And I explicitly disclaim any commitment to the "idiocy" comment. I know this is the internet, but let's aim for civility.)


* Except the Marlins, who don't, as an organization, try to win games. Because of revenue sharing, it is possible to be profitable without much in the way of revenue.
   29. McCoy Posted: August 23, 2014 at 05:47 PM (#4777817)
It isn't really each win but bunches of wins that increase revenue. If you go from 82 wins to 92 wins you might increase your revenue by 30 million dollars but that doesn't mean each win was worth 3 million dollars. In fact wins 1 through 5 might actually be worthless, meaning the team might very well have seen no real increase in revenue if they had won 83 to 87 games.

No team makes more money because they won 66 games as opposed to the 65 wins from the year before.
   30. PreservedFish Posted: August 23, 2014 at 05:53 PM (#4777820)
#29 - of course, but we're talking about averages. At the start of the season there are probably 20 teams that for all you know might end up in that area where 1 win can mean $10+ million dollars. So yeah, if you can eke another win out of your roster by tinkering with the lineup, it's worth considering.
   31. McCoy Posted: August 23, 2014 at 06:01 PM (#4777823)
I don't have anything against trying to maximize your wins. I'm against the notion that every single extra win is worth X dollars when it is not.

During the preseason what matters is what your customers expect*. If your team won 72 games last year and they add a player that is going to be on paper 1 win better than last year's options you aren't going to see an increase in ticket sales because you're expected to win 73 games instead of 72 games. If you add 3 or 4 players and are now expected to win 82 to 88 games you will see an increase in revenue.

During the season what matters is what you're likely to do based on in season past performance and your standings. If you go 20-40 in your first 60 games revenue will be sluggish from that point on. If you had gone 21-39 your revenue won't be slightly better from that point on either.


*It is one of the variables. There are many but in terms of wins on a year to year basis the greatest variables have to do with your fans and expectations
   32. Ziggy Posted: August 23, 2014 at 06:02 PM (#4777824)
Mr. Fish, I appreciate the help, and your point is well taken, but I'm not sure about the "of course" with which you began. The mapping from revenue to wins isn't linear, but it's not flat either. I'm sure we've got a graph around here somewhere...

Edit: flat until you hit the potential-playoff spot I mean
   33. BDC Posted: August 23, 2014 at 06:05 PM (#4777827)
There's a slight definitional disconnect here between what teams are willing to pay for a player one win better than another, and how much a win might yield them in revenue. And as several have pointed out, it's all on aggregate, and team situations might differ. Optimizing the Rangers' lineup this year would be like rearranging deck chairs on the freaking Titanic.
   34. McCoy Posted: August 23, 2014 at 06:13 PM (#4777830)
Vince Gennaro back in 2007 using 5 game swings at two different points 78 wins and 86 wins found that the value of those 5 wins is different. On average 5 additional wins at the 78 win mark would give a team an extra 4.2 million dollars and doing that at the 86 win mark would give a team an extra 11.7 million dollars.

   35. Walt Davis Posted: August 23, 2014 at 06:15 PM (#4777831)
In general, your batting order versus lefty and righty starting pitchers should rarely change

This makes no sense, by MGL's own system/logic. First he puts a ton of caveats on it, specifically around "L/R platoon considerations" and avoiding back-to-back lefties. So other than letting the handedness of the starting pitcher decide whether a player is in the lineup or not and other than letting the handedness of a pair of batters determine whether they hit back-to-back, the L/R issue should be ignored. OK.

But beyond that -- as MGL well knows, there are platoon advantages. It is quite possible that, oh I dunno, Adam LaRoche is one of your two best hitters vs RHP (and therefore should be in the 2 or 4 hole per MGL) but is not against LHP. Either it is optimal to put your best hitters 2 and 4 or it's not.

On the "interaction" thing -- I'm not sure I buy that explanation (sample size seems most likely) but it's a bit worrisome that MGL does yet continues to use the term "optimal" for an algorithm that doesn't include interaction effects.
   36. McCoy Posted: August 23, 2014 at 06:16 PM (#4777833)
Mr. Fish, I appreciate the help, and your point is well taken, but I'm not sure about the "of course" with which you began. The mapping from revenue to wins isn't linear, but it's not flat either. I'm sure we've got a graph around here somewhere...

Edit: flat until you hit the potential-playoff spot I mean


Neil deMause back in 2007 has a chart that is basically that. Flat until the playoff range and then back down to flat once you amass so many wins that a playoff spot is guaranteed. At the time I think his chart had it at something like 500,000 dollars in extra revenue per win until you get to around the 82 game mark and then peaks falls back to about 500,000 dollars once you hit the 100 win mark or so.
   37. PreservedFish Posted: August 23, 2014 at 06:16 PM (#4777834)
nuttin
   38. McCoy Posted: August 23, 2014 at 06:30 PM (#4777840)
Take a look at the Royals from 2008 to 2013. The Royals in 2009 actually lose 10 more games than they did the year before yet attendance rises by 200,000. This can be attributed to fan expectation. Customers thought the team was going to be better so they bought more tickets plus they started off somewhat decently (winning record in April and had a 6 game winning streak at the beginning of May) before the wheels came off their season. The next year they improve their record by 2 wins and yet attendance falls by about 180,000. In 2011 they increase their win total by 4 wins and their attendance goes up by 109,000. In 2012 they add one more win to their total and their attendance goes up by 15,000. In 2013 their win total jumps by 14 wins and yet their attendance goes up only by 11,000. This year they are on pace to surpass their win total of the previous year by 5 wins and they are on pace to increase attendance by 150,000.

It looks to me that about the only time wins have much value is when you increase them by a good deal within the 85 to 100 win range. The Royals increased their win total by a massive 14 wins and yet it hardly registered as a blip in their attendance. Would they have drawn only 10,200 extra fans to the game if they had won 85 games? I don't think so.
   39. God Posted: August 23, 2014 at 07:40 PM (#4777855)
Both L.A. teams have one of the best players in the league batting #2.
   40. Walt Davis Posted: August 23, 2014 at 08:56 PM (#4777876)
#38: You hit on one of the complications -- it is more likely/common-sensical that last year's record is predictive of this year's attendance than any supposed real-time effects. This lag effect is sort of obvious -- after all, whether you win the last game of the season will have zero effect on your revenues of that season unless it's the one that puts you in the playoffs. That rather obviously extends back -- your record in the last week of the season won't really matter, your record in Sept will only matter if you're in contention, etc.

In-season performance effects are surely mostly limited to "are you playing OK when summer begins?" and "does that last through the whole summer?" That and September contention drive in-season sales, along I suppose with things like whether you've got a star player/pitcher.

Increases in pre-season expectations, clearly influenced but obviously not solely determined by last season's performance, are the likely cause of increased attendance from one season to the next. How much difference does 1 win make in pre-season sales? I don't have a clue but I'd be surprised if it's not substantially higher than the in-season effect.

All that said, yeah, I suspect any wins-to-attendance models, even lagged, will turn up minimal effects and they'll be non-linear around "500", "contention", "division win,best record". It's rather hard to see these direct links. Let's say Royals attendance is up 150,000 ... even at $80 revenue per person on average, that's only $12 M extra revenue. That's 2 wins on the FA market with no increase in profits and assuming the other marginal costs of the extra attendance are ignorable. Does 2 wins lead to 150,000 more people? That's Infante vs. Chris Getz ... kinda hard to imagine.

The other issue is whether it matters anyway. Ticket sales are still a big source of revenue but the fixed league revenue and the TV contract that the team may have signed 5-10 years ago are completely unaffected by in-season performance or pre-season expectations (unless maybe they negotiated a new TV contract that offseason). The question then becomes how much does an extra 1-2 wins this season impact on the next TV negotiation in 3-4 years? That's got to be close to zero I'd think.

Analysis: compare teams' April attendance to April of the previous year. This will almost entirely be due to the change in pre-season expectation although I suppose you have to be careful if the team offers a lot of discounted tix in one year but not the other. Does this ratio change in-season and, if so, how does that relate to in-season record to that point (and hot streaks, relative standings, etc.)?

   41. McCoy Posted: August 23, 2014 at 09:12 PM (#4777888)
This lag effect is sort of obvious -- after all, whether you win the last game of the season will have zero effect on your revenues of that season unless it's the one that puts you in the playoffs.

As usual Walt hits it right on the head. Winning or losing the last game of the season for almost all of the teams is meaningless yet people will insist that winning or losing that one game will gain/cost lots of money.

Heck, if you really want to get extreme think about a team that finishes the season with 10 games on the road and is sitting on 75 wins when they go on the road. At that point it doesn't matter whether they win none or all of those 10 games as their ability to gain revenue has already ended.
   42. Bug Selig Posted: August 23, 2014 at 09:55 PM (#4777903)
If your team won 72 games last year and they add a player that is going to be on paper 1 win better than last year's options you aren't going to see an increase in ticket sales because you're expected to win 73 games instead of 72 games.


The Detroit Tigers sold a record number of season ticket packages in the 24 hours after they signed Ivan Rodriguez. They had just lost 119 games.

I agree that the issue isn't nearly as simple as "1 win = $X million" but I dislike the groundswell of "Let's just throw our hands up 'cuz Presidentin' is hard." Not all hits lead to runs, not all runs lead to wins, not all wins lead to championships, not all prospects provide value - but the answer isn't to pretend that they are therefore without value or that their value is impossibly mysterious.

We have an agreement on a thousand issues here that things have aggregate value that has to be assigned discretely - why is that out the window in this case?
   43. TerpNats Posted: August 23, 2014 at 10:10 PM (#4777906)
Some other historical cases of batting a "decent guy" 2nd include the 1982 Brewers and the 1996 Mariners.
A "decent guy" batting 2nd, a certain catcher from New Hampshire, played a key role in the second-half surge of the 1983 White Sox.
   44. McCoy Posted: August 23, 2014 at 10:44 PM (#4777919)
We have an agreement on a thousand issues here that things have aggregate value that has to be assigned discretely - why is that out the window in this case?

What is being thrown out the window? I'm saying the marginal value of a win in a season is close to zero if not zero in most scenarios. To believe otherwise is to assume that a team with 73 wins going into its final game would make more money in that season if they won the final game instead of losing it. How is that possible? Are they betting on the game? Even ignoring the final game of the season issue why would a team make more money because they won 78 games instead of 77?

How much more money do all teams with 80 wins make as compared to all teams that win 79 games or 81 games?
   45. PreservedFish Posted: August 23, 2014 at 10:56 PM (#4777923)
I don't really understand McCoy's intransigence here. Your own quote:

At the time I think his chart had it at something like 500,000 dollars in extra revenue per win until you get to around the 82 game mark and then peaks falls back to about 500,000 dollars once you hit the 100 win mark or so.


Yeah, $500,000 per extra win. That's something.

If there are differences between 50 and 60 and 70 wins in a season then we are able to derive a per-win dollar amount. Even if that amount is low. This seems pretty simple to me. So yes, 73 wins does mean more revenue, on average, than 72 wins does. (To me the year that the money shows up is beside the point)
   46. PreservedFish Posted: August 23, 2014 at 11:01 PM (#4777925)
And anyway, I agree with the larger point, which is that most teams are probably not going to earn back the $6 million (or whatever) they've spent on increasing their win expectancy by a single win. Actually I think that most of the analysis we see in this community regarding $/WAR, the cash value of draft picks, even the motivations of GMs and owners is off base.
   47. McCoy Posted: August 23, 2014 at 11:02 PM (#4777927)
Yeah, $500,000 per extra win. That's something.

If there are differences between 50 and 60 and 70 wins in a season then we are able to derive a per-win dollar amount. Even if that amount is low. This seems pretty simple to me. So yes, 73 wins does mean more revenue, on average, than 72 wins does. (To me the year that the money shows up is beside the point)


Okay. Prove it. The only reason we're able to say that 72 wins brings in more revenue than 71 wins is because you can compare a 80 win team to a 70 win team and see that a 80 win team brings in more money. Well, duh. 10 more wins is a great amount and thus will alter the spending habits of potential customers. 1 win will not. The actual real world difference between a 72 win season and a 73 win season is practically zero if not zero.
   48. Ziggy Posted: August 23, 2014 at 11:03 PM (#4777928)
While I'm not sold on team record having little effect on revenues except when you're near a playoff spot (in fact the study cited in #34 appears to say that 5 wins for a 78 win team were worth almost a million dollars each seven years ago), but let's say that's right and see what it tells us about optimal strategies. (I'll note too that it's not just attendance that will change - it's also how much you can get for so-and-so being the official bank of your team, how much the signs in the stadium are worth, if you own your own RSN how much you can get for advertising, and so on.)

Anyway, let's say that team record makes almost no difference to revenue unless you're up around 90 wins or your cable deal is close to expiring. Getting to 90 wins is expensive and risky. The Yankees, for example, have to play .686 ball the rest of the way to manage it. If they don't manage it, they're paying McCann $18m, Teixiera $23m, etc., for almost no return. If I was an investor in a baseball team, that prospect would terrify me. If marginal wins have very little value (until you get up around 90), it MIGHT be a good idea to play only rookies until you hit the spot when your team record will make a difference to your next cable deal. (And once the deal is signed salary dump everyone.) Whether that's a good idea or not depends on the probability that spending $x will get you 90 wins, and the value of getting to 90, but it might actually be a good idea.
   49. cardsfanboy Posted: August 23, 2014 at 11:16 PM (#4777933)
While I'm not sold on team record having little effect on revenues except when you're near a playoff spot


I'm not sold at all. Obviously expectations (or population size)is the primary factor in driving attendance (and tv ratings which is also revenue geared) all of these arguments are focusing on either one year or at most two year revenue in relation to record, but that isn't the full story. The trust that is built between management and the fans will affect attendance for lean years. (there are exceptions of course, as Tampa has crappy attendance and has been a quality team for years, has done all the right things, said all the right things and still can't get fans to the stadium) but ultimately if a team shows a willingness to spend, shows a willingness to keep their players, and shows some level of committing to improvement, the fans will stick around even in less than perfect years. But it does take time for the fans to build up that trust with the owners, and it doesn't take much for the owners to lose that trust.

Anyway, my point is that attendance factors are almost entirely based upon what the fans perceived as a competitive team and whether or not the ownership could be trusted to try and compete each year.
   50. McCoy Posted: August 23, 2014 at 11:40 PM (#4777947)
While I'm not sold on team record having little effect on revenues except when you're near a playoff spot (in fact the study cited in #34 appears to say that 5 wins for a 78 win team were worth almost a million dollars each seven years ago),

The study says that winning 83 games instead of 78 games will net you about 4 million dollars on average. That doesn't mean that if you had won 79 games instead of 83 games you would have gotten 800,000 dollars extra in revenue.

People don't make buying choices based on that minute of a detail.

There have been 4 franchises since 1963 that have gone from 79 wins to 80 wins from one year to the next. Their attendance went up by roughly 220 tickets per game. Two of the teams had sizable decreases and two of the teams had sizable increases.
   51. PreservedFish Posted: August 23, 2014 at 11:45 PM (#4777952)
Okay. Prove it. The only reason we're able to say that 72 wins brings in more revenue than 71 wins is because you can compare a 80 win team to a 70 win team and see that a 80 win team brings in more money. Well, duh. 10 more wins is a great amount and thus will alter the spending habits of potential customers. 1 win will not. The actual real world difference between a 72 win season and a 73 win season is practically zero if not zero.

I don't really have any response to this. I have the same question as the one above that you quoted but didn't answer: We have an agreement on a thousand issues here that things have aggregate value that has to be assigned discretely - why is that out the window in this case?

When MLB teams project future revenues, they surely include some sort of $/Win factor. To do that you need to assign a value to each win. You can't just say "69 wins won't do anything, but 76 wins will." What if they win 72 or 74 games?
   52. McCoy Posted: August 23, 2014 at 11:54 PM (#4777962)
There have been 14 teams that have gained 1 win more than they had the year before between 1969 to 2008 in the 68 to 81 win range. Those 14 teams on average had a drop in attendance of about 106,000 per team. The average team in this group won 74 games in year 1 and 75 games in year 2. It looks like to me when you are a mediocre or worse team your fans will not tolerate you putting out basically the same product you put out the year before.
   53. McCoy Posted: August 23, 2014 at 11:58 PM (#4777970)
When MLB teams project future revenues, they surely include some sort of $/Win factor. To do that you need to assign a value to each win. You can't just say "69 wins won't do anything, but 76 wins will." What if they win 72 or 74 games?

And I'm telling you that if you do a $ per win ratio you'll find that it won't apply to 1 single solitary win. A team may get 900,000 dollars per win if they win 5 extra games but they won't get 900,000 extra dollars if they win 1 extra game instead of 5. In fact I'm willing to bet that whatever revenue boost or decline a team does get in that year will have almost nothing to do with winning 1 extra game. Fans simply don't make buying choices off of that small of a difference*. We simply cannot register that minute of a change and have it alter our buying habits.

*Well, it appears that if you put out the same crappy team year over year less fans will come out even if you do win one more game than you did the year before.
   54. McCoy Posted: August 24, 2014 at 12:09 AM (#4777981)
Using the same data group that I used to get post 52 I find that if you lose 1 more game as compared to the year before your attendance goes down by 82,000. There were 10 teams in the sample group*. Put the two together and you basically have little to no impact on attendance outside of a slight negative one. An extra win at this level of win totals amounts to a decrease of attendance by almost 28,000.

*I excluded the 2001 to 2003 Rangers as their attendance was warped by the presence of ARod.
   55. PreservedFish Posted: August 24, 2014 at 12:30 AM (#4777997)
And I'm telling you that if you do a $ per win ratio you'll find that it won't apply to 1 single solitary win. A team may get 900,000 dollars per win if they win 5 extra games but they won't get 900,000 extra dollars if they win 1 extra game instead of 5.


Ok. What about if they win 2 extra games? Or 3? Or 4? Obviously they don't just magically jump 4.5 million in revenue when they go from the 4th extra win to the 5th. You're too good for division and straight lines, so what does that graph look like? Is it a curve between those two points?
   56. McCoy Posted: August 24, 2014 at 12:40 AM (#4778001)
There have been 9 teams that have won 2 more games than the previous year. They won between 70 and 77 games in year 2. On average their attendance declined by just under 50,000 in year 2. I again excluded ARod's year 1 in Texas.

There was 5 teams that won 3 more games. Their range of wins was between 72 and 79. On average attendance went down by almost 100,000. 2002 Florida is doing a lot of the damage in this case. The other 4 only average a loss of about 8,700.
   57. Danny Posted: August 24, 2014 at 12:57 AM (#4778004)
In general, your batting order versus lefty and righty starting pitchers should rarely change

This makes no sense, by MGL's own system/logic. First he puts a ton of caveats on it, specifically around "L/R platoon considerations" and avoiding back-to-back lefties. So other than letting the handedness of the starting pitcher decide whether a player is in the lineup or not and other than letting the handedness of a pair of batters determine whether they hit back-to-back, the L/R issue should be ignored. OK.

But beyond that -- as MGL well knows, there are platoon advantages. It is quite possible that, oh I dunno, Adam LaRoche is one of your two best hitters vs RHP (and therefore should be in the 2 or 4 hole per MGL) but is not against LHP. Either it is optimal to put your best hitters 2 and 4 or it's not.

I think you're misreading. MGL is saying that your lineup against RHP should rarely change, and your lineup against LHP should rarely change--you shouldn't adjust them based on short term streaks/trends. He's saying that handedness platoon is--in general, given his caveats--the only valid reason to switch up the lineup.
   58. cardsfanboy Posted: August 24, 2014 at 01:00 AM (#4778006)
I am not really sure if I'm asking the right question or presenting the right point of view, but I wonder what the value there is for stability? or more accurately slow growth?
   59. cardsfanboy Posted: August 24, 2014 at 01:05 AM (#4778008)
I think you're misreading. MGL is saying that your lineup against RHP should rarely change, and your lineup against LHP should rarely change--you shouldn't adjust them based on short term streaks/trends. He's saying that handedness platoon is--in general, given his caveats--the only valid reason to switch up the lineup.


And someone who does this is an idiot who thinks computer simulations are more accurate than understanding human beings. Streaks absolutely happen, even a robot like MGL admit this, (his argument is the predictive ability on whether or not the streak will continue in the next game and not about whether they happen)

He wants to ignore the concepts of keeping human beings fresh by resting them day to day, by keeping the bench fresh by playing them relatively frequently, by tailoring a lineup not only based upon the starting pitcher, but based upon what the opposing team can bring in the 7th, 8th or 9th innings, basically, he wants to play strat-o-matic baseball and ignore the concept of day to day burnout, trends etc. he's a g-d computer who has absolutely no clue how human beings operate.

   60. McCoy Posted: August 24, 2014 at 01:12 AM (#4778010)
Cub fans are a pretty good example of fans not really caring about the amount of wins you have when your team is crappy. In 2012 the Cubs lost 10 more games than they did the year before but attendance stayed largely the same. In 2013 they won 5 more games than they did in 2012 and yet attendance cratered by almost a quarter of a million. This year the Cubs are on pace to increase their win total by another 5 games and yet their attendance figures are on pace to stay unchanged.
   61. cardsfanboy Posted: August 24, 2014 at 01:23 AM (#4778012)
Cub fans are a pretty good example of fans not really caring about the amount of wins you have when your team is crappy. In 2012 the Cubs lost 10 more games than they did the year before but attendance stayed largely the same. In 2013 they won 5 more games than they did in 2012 and yet attendance cratered by almost a quarter of a million. This year the Cubs are on pace to increase their win total by another 5 games and yet their attendance figures are on pace to stay unchanged.


The cubs are somewhat limited by their stadium, but having said that, I fully expect that after the relatively surprising positive year, that there will be a moderate increase next year.

But this does point out, a factor of fan loyalty that goes beyond record and ownership(of course it helps having a city with 8million people in it)
   62. PreservedFish Posted: August 24, 2014 at 01:25 AM (#4778013)
It's also strange that you keep ignoring the year's attendance/revenue lag.
   63. McCoy Posted: August 24, 2014 at 09:20 AM (#4778034)
It's also strange that you keep ignoring the year's attendance/revenue lag.

I do? Wait, I'll answer that. I don't. I specifically said that fan expectation during the offseason helps drive attendance. I will say that you seem to keep ignoring my point that 1 win doesn't mean anything to fans. People aren't going to buy more tickets because last year's team won 72 games instead of 71 games. I'll also add that the studies that look at revenue per marginal win are looking at revenue for that season so. . .
   64. BDC Posted: August 24, 2014 at 10:44 AM (#4778049)
slow growth?

I was thinking that at some point in this discussion the sorites paradox applies. One win isn't much, two wins isn't much, three wins isn't much … but twenty wins is.

No team sets out in the off-season to improve by a single win. Almost no team figures it's suddenly going to improve by 20. You have to get to twenty by small increments, and thus each small increment is pretty important.
   65. snapper (history's 42nd greatest monster) Posted: August 24, 2014 at 11:22 AM (#4778062)
I do? Wait, I'll answer that. I don't. I specifically said that fan expectation during the offseason helps drive attendance. I will say that you seem to keep ignoring my point that 1 win doesn't mean anything to fans. People aren't going to buy more tickets because last year's team won 72 games instead of 71 games. I'll also add that the studies that look at revenue per marginal win are looking at revenue for that season so. . .

McCoy, under your theory no team would ever pay more than league minimum for a 1 WAR player, yet they do all the time.

At the beginning on the season, teams don't know where they're going to end up. So, if there's a 1 in 20 chance that 1 win will get them to the playoffs and generate $40M in incremental revenue (now and in the future) then the win is worth $2M.

The lineup change work the same way. If you invest in the "optimal" lineup you expect to gain 1 win, which will, on average, lead to a $2M gain. Even if it leads to no gain 95% of the time it still has value.

You're getting hung up on the probability of the win mattering. Let's look at it the other way.

What if optimizing you lineup had a 95% of generating no extra wins, and a 5% chance of generating 20 extra wins? That would obviously be really valuable right? You'd pay a lot for two Cuban defectors with a 1 in 20 chance of being Mike Trout.

This is exactly the same economically. You invest in the win, and there's a small chance the win is incredibly valuable.
   66. McCoy Posted: August 24, 2014 at 11:35 AM (#4778067)
So 95% of the time the win is meaningless and yet you think that proves my view wrong? Under your view 95% of the time an extra win has no value, not even 2 million dollars, and 5% of the time it is worth 40 million dollars. That's it. That doesn't mean every extra win in every scenario is worth 2 million dollars. It means simply what it means. That 95% of the time the win is valueless and 5% of the time it is worth 40 million dollars.

If you were a 72 win team the year before improving by a win is not going to increase your revenue by 2 million dollars (if the 40 million is accurate and it isn't) in fact the data appears to show that if you're a crappy team and you don't noticeably change your performance the next year your revenue is going to be hurt by a such an insignificant gain as 1 more win. There is no difference in a customer's mind between a 75 win team and a 76 win team.

McCoy, under your theory no team would ever pay more than league minimum for a 1 WAR player, yet they do all the time.

Supply and demand, probabilities/flawed predictions, and of course stupidity. I certainly agree, and said it earlier in I think the KC thread, that a team should be trying to maximize their wins when they can reasonably project themselves to be in or around the playoff entrance threshold. At that point you're kind of dealing with pot odds when you're handing out money to players. If you think you're a 85 win team +/- 2 wins and you think that player Y can increase your win total by 1 win and you believe that moves your playoff chances from 20% to 35% and getting into the playoffs and being that much thicker in the playoff hunt will net you 30 million more dollars you can "overpay" for that 1 win because of the increased odds to get into the playoffs.
   67. dave h Posted: August 24, 2014 at 01:31 PM (#4778105)
I'm not sure how McCoy is even typing his arguments, since every time he tries to move his fingers to type, they only make it halfway to the key, and then they make it halfway again, and halfway...
   68. Walt Davis Posted: August 24, 2014 at 06:39 PM (#4778254)
I find these criticisms of McCoy a bit odd.

He originally wondered why this extra win was so important and somebody responded "because each extra win is worth $5 M".

He then said "no it's not, the effect is non-linear so most of the time, that extra win is adding little/no revenue at all."

This has been countered, as one example, by the riposte "maybe that win is only worth $500,000 but that's still $500,000 extra revenue."

Which is a statement that McCoy was right. Person A claimed a win was worth $5 M; person B claims it's worth next to nothing; subsequent analysis puts the value at 10% of the original claim ... Person B is "correct" under those circumstances.

And that's leaving aside McCoy's look at actual cases of 1-2 win improvements experiencing attendance stagnation/decline.

Why dispute this aggregate number and not others? A fair question but the answer is mainly because somebody claimed that an optimal lineup is worth $5 M more than a "standard" lineup -- a very micro claim. If the claim had been "because you never know when an extra win is going to make a big difference and optimizing the lineup costs you nothing" then maybe the thread heads another direction.

I keep coming back to this -- plesae do the math. What is the average revenue per marginal fan? I tossed out $80 thinking that would be rather high but plug in whatever number you think belongs.

Now say you increase attendance by a million -- massive, a 50% or bigger increase for some teams, at least a 30% increase for almost everybody. That would be an increase of $80 M.

Depending on your $/WAR estimates, that's anywhere from 12 to 16 WAR. Again, that's assuming none of that extra revenue is taken as profit and that it has negligible impact on costs.

Is it really the case that going from 75 to 89 wins increases attendance by 1 million fans? That's probably where those 14 wins would have the most impact -- you past the 500, contention and possibly playoff thresholds.

Again, the Royals signed Omar Infante for 4/$32. We have to believe that 100,000 extra fans are coming out to see Omar Infante. Really?

The other reason "why question this" is because we in fact know virtually nothing about revenue generation in baseball. It is a complete mystery to us.

The $5M/win figure comes from an estimate that has nothing to do with revenue generation. That estimate is of course based on the marginal cost of a win in the FA labor market ... and it it then simply assumed that this cost must be balanced by a revenue increase of $5 M or more per win. The difference being we KNOW how much each player costs and we have very good estimates of how many wins each player is individually responsible for, so we have some confidence in that $/WAR estimate. We know virtually nothing about revenue generation, how an extra win here or there affects revenue, what the effect on future TV contracts is, what effect on attendance 1,2,3 years down the road, the effect on that "trust" between team and fans.

What we do know suggests the effect on attendance is quite non-linear. But we also know that attendance is a much smaller driver of revenues than it used to be -- TV deals (sometimes RSN ownership), stadium deals, shared revenue, competitive balance redistribution have greatly weakened the attendance-revenue link. TV deals and publicly built stadiums are, presumably, more reflective of that necessary "trust" and have nothing to do with any individual signing ... the shared revenue sources aren't affected at all by individual signings.

And that attendance revenue is not exactly shaped linearly. The key driver is presumably pre-season sales. Then you have the low months until school lets out. Summer "casual" sales are presumably affected by performance/contention throughout the summer (along with weather, whether the Yanks are coming to town, etc.) Then, unless you're in contention, you're back in the doldrrums for Sept ... when several teams are intentionally not fielding their "best" lineups as they try to give some experience to prospects.

Again, Sept makes it fairly clear that there is not any sort of strict win=$ conversion. The Royals presumably will see a very nice boost to their Sept attendance as casual fans jump on the bandwagon; Cub attendance might see a nice little spike if they play Bryant, etc.; the Red Sox will see a big drop in Sept attendance unless the vast majority of tix were sold pre/early-season; the Rays will probably see some decline. So Sept attendance is presumably a function primarily of last year's performance (Red Sox) and this year's performance through August.

Wins in Sept will have pretty much nothing to do with it for any of these teams except the Royals ... and the teams will act accordingly and use Sept to evaluate talent as much as to try to win games. That talent evaluation will hopefully help them perform better next year ... maybe even build a bit of offseason interest but most likely in hopes that the 2015 team will do better, increasing revenues for 2016.

Nevertheless, "winners" clearly draw better than "losers."
   69. PreservedFish Posted: August 24, 2014 at 06:55 PM (#4778260)
McCoy was certainly right about his larger point. He's just being stubborn.
   70. bobm Posted: August 24, 2014 at 08:08 PM (#4778283)
We know virtually nothing about revenue generation, how an extra win here or there affects revenue, what the effect on future TV contracts is, what effect on attendance 1,2,3 years down the road, the effect on that "trust" between team and fans.

What we do know suggests the effect on attendance is quite non-linear. But we also know that attendance is a much smaller driver of revenues than it used to be -- TV deals (sometimes RSN ownership), stadium deals, shared revenue, competitive balance redistribution have greatly weakened the attendance-revenue link. TV deals and publicly built stadiums are, presumably, more reflective of that necessary "trust" and have nothing to do with any individual signing ... the shared revenue sources aren't affected at all by individual signings.


I would say off-season free agent signings drive season ticket sales, as a signal to fans of the team's commitment. Other than the old NL real attendance data, it is hard to gauge the walk up effect over the course of a season.
   71. Bug Selig Posted: August 24, 2014 at 09:43 PM (#4778321)
Why dispute this aggregate number and not others? A fair question but the answer is mainly because somebody claimed that an optimal lineup is worth $5 M more than a "standard" lineup -- a very micro claim. If the claim had been "because you never know when an extra win is going to make a big difference and optimizing the lineup costs you nothing" then maybe the thread heads another direction.


I like this. It was my question and I'm totally cool with this interpretation of the context.

Been thinking about this, because you're right - this thread went a little off the tracks. I think it might be because the claim that a win doesn't matter is anathema to, well, sport. I don't expect that McCoy thinks so in a larger sense because he was responding to a very specific claim, but the insistence that it is true in even a financial sense is, taken out of context, offensive and kinda wacky-sounding. Again, not fair to project the larger issue onto a specific claim but I suspect that the reaction is somewhat visceral and based on that idea.

It is interesting, though, that we regularly discuss transactions, strategies, etc. that are worth much less than a win or a loss and absolutely shred managers and GMs who do things we find suboptimal. Hell, almost everything we discuss in terms of strategy exists in a realm of a win or less, and the "why bother" point of view is never or virtually never expressed. Lineup optimization, with caveats considered, is a way bigger deal than pinch-hitting Phil Coke for Miguel Cabrera with the bases loaded and we'd go absolutely banana-pants if anybody even proposed that.
   72. dave h Posted: August 24, 2014 at 09:47 PM (#4778325)
The fact that teams pay $6-7 million per marginal win is exactly the right measure of how much optimizing the lineup is worth. Let's say you're the Epstein-era Red Sox. They pretty much always seemed to target a 95-win projection, which made sense for maximizing the chances of making the playoffs without completely breaking the bank. You're Epstein, and you have deals laid out to get to 94 wins. You can either pay the going rate for another WAR in free agency, or you can instruct your manager to play the optimal lineup, for free.

This is assuming that you can really get 1 win out of the optimal lineup, but that doesn't seem to be the argument.
   73. McCoy Posted: August 24, 2014 at 09:48 PM (#4778326)
McCoy was certainly right about his larger point. He's just being stubborn.

Wait, what? That's what you take away from this?

You must be Registered and Logged In to post comments.

 

 

<< Back to main

BBTF Partner

Support BBTF

donate

Thanks to
Phil Birnbaum
for his generous support.

Bookmarks

You must be logged in to view your Bookmarks.

Hot Topics

NewsblogMets close season optimistic for next year
(51 - 8:05am, Sep 30)
Last: Lassus

NewsblogESPN: Ron Gardenhire out after 13 Seasons with Twins
(36 - 8:03am, Sep 30)
Last: Lassus

NewsblogMLB’s Biggest Star Is 40 (And He Just Retired). That Could Be A Problem.
(8 - 7:50am, Sep 30)
Last: simon bedford

NewsblogOT August 2014:  Wrassle Mania I
(262 - 7:49am, Sep 30)
Last: Rowland Office Supplies

NewsblogBaseball Past & Present: Vote: The 25 Most Important People in Baseball History.
(282 - 7:41am, Sep 30)
Last: Rennie's Tenet

NewsblogRemembering George ‘Shotgun’ Shuba, 1924-2014
(6 - 7:35am, Sep 30)
Last: Jolly Old St. Nick Is A Jolly Old St. Crip

Hall of MeritMost Meritorious Player: 2014 Discussion
(10 - 7:32am, Sep 30)
Last: bjhanke

NewsblogAttanasio discusses Brewers collapse, changes coming for 2015
(127 - 7:04am, Sep 30)
Last: Harveys Wallbangers

NewsblogRBI Baseball, 2014 Playoff Edition
(3 - 6:54am, Sep 30)
Last: Leroy Kincaid

NewsblogThe Calm-Before-The-Storm and Postseason Prediction OMNICHATTER, 2014
(82 - 6:33am, Sep 30)
Last: Dr. Vaux

NewsblogMadden: How dare the sabermetrics crowd and others try to diminish Derek Jeter’s greatness
(144 - 5:20am, Sep 30)
Last: cardsfanboy

Hall of MeritMost Meritorious Player: 1958 Ballot
(9 - 2:49am, Sep 30)
Last: bjhanke

NewsblogOT: Monthly NBA Thread - September 2014
(392 - 1:53am, Sep 30)
Last: Gold Star - just Gold Star

NewsblogOT: NFL/NHL thread
(8138 - 11:55pm, Sep 29)
Last: PASTE Thinks This Trout Kid Might Be OK (Zeth)

NewsblogThe Captain’s Log: Derek Jeter’s Lady-Killing Past, From ‘Yeah, Jeets!’ to Gift Baskets
(29 - 11:21pm, Sep 29)
Last: Walt Davis

Page rendered in 0.9070 seconds
53 querie(s) executed