Open question for discussion: value of need
I don’t understand something in the current discussion and the many many comments about how the M’s “need” more power. I’m hoping someone can enlighten me on the theory at least, if not on the practice.
Say there’s an average team with no first baseman and no internal options. They have two options in free agency for exactly the same price: Doug the Defensive Guy, who would be five runs below average hitting and a wondrous +15 above average on defense. And they could sign Mike the Masher, to get a +15 on offense, -5 defense guy.
Everyone would argue it’s a coin flip in value, right? You’d start to look again for extremely fine differences like whether the team’s pitchers would particularly benefit (or cover for Mike), or if the park suits one or the other. But say you do all of that and those numbers are indeed the same. The average team picks whoever’ll sign first, or knows the manager from the minor leagues, or whatever.
Now what if the team is average by way of sucking defensively and good offensively. The return on both those guys is still +10. Is the theory that they should sign the defensive specialist for balance? How much extra value is that?
And conversely, if the team’s average by way of being terrible offensively and good defensively, does the reverse hold true? How much is that guy worth?
That’s one question: does improving something the team is bad at offer greater gains than improving elsewhere, and if so, how much?
And does it matter how bad they are? Is the return on improving defense more than 10 runs if they’re league-worst? Is there a kind of elasticity to returns, where only the average team values players based on overall contribution?
Because if that’s true, and there’s a value, then we could actually start to discuss this. Say Doug and Mike aren’t asking for the same price. Defense is so highly valued that Doug already has 4 offers on the hood of his gold Land Rover for $20m/year, while Mike is looking at $10m for the next year.
Does the defensively challenged team still want Doug at that price? Or are they better off picking up Mike?
We’ve laid out my (and I’d say Dave’s, to a different extent) view on this — I think all runs are created more or less equally, and you’re as well-advised to take them off the board as put them on, so improving on pitching, defensive prowess, and offense are all equally good. I don’t know of any evidence that if a team needs a first baseman, regardless of how they did last year and how they did it, that they shouldn’t take the player who is so undervalued.
There’s a big caveat to that, which is that (and I know I mention this over and over) in constructing a team you want to get into and through the playoffs there are some things you want to have in place and should think about paying for. But in general, for teams who aren’t budgeting for 85 wins, does valuation change?
What’s the opposing theory, and what’s the evidence for it?
Comments
82 Responses to “Open question for discussion: value of need”
Leave a Reply
You must be logged in to post a comment.
Given the structure of the game and a human being’s desire to control their own destiny (i.e., their psychology), they will value the things they perceive give them more control (i.e., hitting & pitching) over things that cannot control, i.e., defense. Why can’t they control defense? Because the defense only comes into play because of another action: hitting. This, of course, causes the defense to react.
This, I believe, is the answer to your question. And, yes, in spite of the fact that it makes no difference, i.e., a run is a run.
Well I did say equal accuracy, not 100% accuracy.
djw, yeah I noticed that. You should be a politician.
So the difference is that a left fielder might not get any chances in a game, where that same left fielder is guaranteed to get at least three at-bats?
That seems like a difference in opportunity, not a difference in action though.
Beyond which, it doesn’t address the central question, which is whether there’s a difference in value contributed, or whether that value varies by team composition. It does address the perception question.
Regarding value of need : 2009 Mariners were near the bottom in runs scored, but stayed competitive by being near the top in runs prevented ( and subsequently played many one run games ) if the 2010 Mariners acquire better players than the 2009 roster, they hopefully will have more wins. Question : Can the value of the role of specialists ( set-up reliever, closer; pinch runner, pinch hitter ) be quatified on a team with so many one run games ?
Go with Mike the Masher….the error bars on offense are smaller than the error bars on defense so projecting his value is safer….
The “first base is less important” argument doesn’t apply, though. You can play a good defensive left fielder and get value out of their glove (see: M’s) and you can play a good first baseman and save runs. In the same way you can punt defense at shortstop and put a hitter there.
And then we loop around to the original question:
Again, why? Even assuming that they’re not good offensively, why does offense become a greater priority than defense?
I haven’t thought much about this before, but I was wondering, is it even possible to accurately assess a return on investment with defense?
Say you get the best defensive shortstop in baseball. Can his value truly be directly related to his fielding ability? Whenever a fly-ball/strikeout pitcher is on the mound, his value is reduced. If he has the best defensive 3rd baseman in the league playing next to him, his value is reduced (since there is an overlap in area of coverage between the two). If the infield has tall grass, his value is reduced.(since the balls which normally only he can get to might now be reachable by say the top-10 shortstops in the league) etc.
Thus, a player’s defensive value can never reach or exceed a player’s true defensive ability.
It’s the reverse for a batter. The top batter’s value to a team is not negatively affected by other players on the same team. In fact, his presence in the lineup can elevate the value of other players to that team. Say you’ve got a high OBP guy batting in front of a lefty pull hitter, that elevates that lefty’s value. Or you have a power hitter after one or two on-base guys. Without the power hitter, those on-base guys’ value to the team is reduced if they never get to home, but with the power hitter, the value increases.
So while there is a reduction in return on investment the more and the better defensive players you have in the vicinity of each other, you get a greater return on investment the more and the better offensive players you have.
Yes.
Yes.
In practice this happens so rarely it’s not really worth considering. You can play two astoundingly good fielders next to each other without significantly diminishing returns.
Beyond which, the same argument you use affirmatively for offensive players is true for defensive players. See: the Winn-Cameron-Ichiro outfield. When the other team is getting fewer and fewer hits because their balls in play are being turned into outs, it does proportionally even more damage to wipe out one of the few baserunners they’re getting.
Beyond which, it doesn’t address the central question, which is whether there’s a difference in value contributed, or whether that value varies by team composition. It does address the perception question.
Oh, sorry. I thought the central question was the perception question. I’ll have to re-read the post. FWIW, I don’t disagree with you on the value question. In fact, I think JackZ has more opportunity by exploiting the (pro-offense) perception. I still think pitching is the key to success, however.
I believe Leroy Stanton is just basically saying that the interaction b/w the pitcher and the hitter is the basic primal scene of analysis, and that extreme abilities one way or another, in this primal scene, may de-emphasize the value of defense.
You get your Nolan Ryan up there facing the other team’s Rob Deer, your slick-fielding Doug type at first is likely to be less valuable; just in the context of that battle, you’d be better off having a stone-footed slugging Mike the Masher that put you ahead by two runs the last inning, because chances are, the current at bat is going to end up a strikeout, or a walk, or a home run.
You get your Jarrod Washburn up there facing their Lou Brock or whomever, then your slick fielder is looking more valuable.
That was my take, but who knows what’s up with people these days.
If you don’t think you’re going to compete for a title and you still need to have able-bodied vets to at least appear competitive, I’d stock up on rangy slick-fielding types to make your pitchers look better, so you can try and increase their value (on trades and as free agents).
Leroy: the batter only comes into play because of another action: pitching. This, of course, causes the hitter to react.
Get it?
Ancient Mariner — I think that’s sort of willfully mis-reading Leroy’s point. “Reactionary” was probably not the most accurate term for what (at least I take) his point, which is that the pitcher-hitter showdown is the sort of causal transaction that everything else is dependent upon.
His point, I think, is that the value of a team’s defense isn’t equal from team to team. (This isn’t my point, by the way, but just my translation of what I took to be Leroy’s interesting point.) It isn’t likely an enormous factor, but if you have a starting rotation of guys with exaggerated strikeout and/or walk and/or homer rates, then that would influence your choice of either a defensive or offensive minded 1B, because a pitching staff that gets, say, a ton of strikeouts but also gives up a lot of walks and homers is one that has a higher number of outcomes in which it’s irrelevant whether you have Keith Hernandez or Fran Drescher playing defense at first.
fans who dont understand statistics “need” homeruns and RBI’s because they would rather win 86 games and lead the league in offense than win 89 games with a balanced team. this is more a question of psychology.
I disagree. If each estimate fits a normal distribution, the higher defensive error applies to both sides of the estimate. Hence if your estimate of defensive ability is +10 runs and the error is +/-10, you might get 20 and you might get none; if he’s +10 runs on offense and the error is +/-5 runs, you might get 15 and you might get 5. Either way, the mean value (10) is the one you should use to make your expected utility calculation.
w/r/t: w/r/t
Some years ago, someone on MARINER USENET wrote an interesting post that included the abbreviation w/r/t. I forwarded the post to my son.
He was familiar with the abbreviation; noted that it had been used before.
So I just asked him about this. Here’s his reply:
“The first writer I noticed using w/r/t (with respect to) was David Foster Wallace. He could’ve got it somewhere else, though.
A.M., jack:
What I said was reactive (vs. proactive). Yes, I get it, a pitcher pitches and the batter reacts to the pitch. The point is that they both seek to do something and the game cannot continue until they do. The defense can never initiate any action, therefore they are reactive. It’s not a bad thing, it’s just the nature of the game.
Also, when I said vastly I meant it, like me pitching to my 6 year old daughter. MLB players are essentially evenly matched with each other as opposed to how they’d stack up against the other 6 billion people in the world. No, I wasn’t clear on that.
Against my daughter, I am vastly superior (yeah, I’m awesome) so the outcome is pretty much certain. I don’t need no stinkin’ defense! But in MLB, since the players are evenly matched, value in defense is just as useful as value elsewhere.
I think you’re wrong. The variance in the number of wins that the 99 run/game team would experience would likely be much higher, leading to a wider range of expected outcomes. However, the chance that they would underperform their Pythag record would be just as great as the chance that they would overperform their Pythag record. The 1 run/game team would be just as bad as the 99 run/game team; they’d just be more predictable.
But when I run the numbers through Pythag, it agrees with you; that is, the low-scoring team would only be expected to win 20% of their games, whereas the high-scoring team would be approaching a 50% rate. Pythag thinks that scoring one additional run and allowing one additional run is beneficial. I’m not so sure about that.
DMZ…….
Is this a babysitting post?
The answer to this one is pretty easy, and I can’t believe it hasn’t come up yet. Two words: fantasy baseball. More words: FBB is very largely about balancing your investments efficiently and equally around many strengths. And I think I can make the assumption that most readers of a site like USSM play fantasy, and when we play GM, we’ve gotta balance our investments across the board.
Be careful with the Pythag example! It makes a difference whether you assume the team scores more runs than the opposition or less.
For a team that scores LESS runs, yes, they are better off adding runs on offense than subtracting with defense (according to Pythag).
For a team that scores MORE runs, however, they are better off subtracting runs with defense than adding on offense (according to Pythag).
A team aiming at the playoffs places themselves in the latter category. Mind you, the effect is very small when you’re talking about players who are between +/-50 runs on teams that score/allow > 600 runs a year.
I don’t remember who said “Half of the game is ninety percent mental” but I think it applies here. While a run is a run, most pitchers seem to have more ‘mental’ problems with power hitters than singles hitters. And power also effects the way they pitch — if Gutierez is hitting third with a power hitter behind him, pitchers will probably throw him more fastballs trying not to walk him than if he has a singles hitter behind him. Maybe statistics show he shouldn’t pitch that way, I don’t know, but that’s the way it’s generally done.
There’s also the effect on our pitchers. The coach always tells them to not try to be perfect on each pitch but just go out and throw. But if a team can’t score, and they know that giving up two or three runs means probably losing, it puts a lot more pressure on them and some of them don’t handle it well.
My last point is that for me baseball is entertainment, and while I enjoy good fielding, I truly love watching my team’s runs cross the plate, and I think a lot of fans are like this. Winning is important, but when I see a whole string of low-scoring pitchers’ duel with great defense type games, I start reading a good book and watching the game just when something I find exciting is happening.
What I missed in terms of the Mariners’ scoring last year wasn’t so much power as batting RISP. I don’t care how a guy on base gets in, I’m happy with a single or double to do it, but another year of stranding guys who got on second or third with no outs is going to take a lot of fun out of the game for me even if we win some.
This is untrue. In Baseball the defense holds the ball, therefore they initiate every action.
I still can’t believe people can’t understand the ‘reactive’ point.
Who cares if the defense holds the ball, and initiates the action? The pitcher would have no reason to throw the ball if there were no batter at the plate. You may as well say the batter initiates the action by stepping in and out of the box. That is pointless metaphysics. Either point gets us nowhere, in terms of strategy and analysis.
Set aside the pitcher and batter on one hand (the left), and the act of fielding on the other (the right). Broadly speaking, the right hand is always responding to what happens on the left hand, not vice versa. There are subtle shadings of the reverse (a hitter taking advantage of a shift, etc), but the defense has to react to what occurs in the pitcher-batter duel.
onetree,
A good pitcher on a bad (or average) team is going to feel added pressure to put the team in position to win that’s true. I don’t see how it matters if the team is bad at run production or run scoring.
It’s not at all clear that a good pitcher on the horrible defense capable offense team will feel less pressure and be more likely to realize his potential. If you’re on a bad or average team with high expectations you’re going to feel pressure and you’re going to find losing frustrating. How many times have we seen a pitcher fall apart after a key error or a double play that wasn’t turned? Bad defense can absolutely get in a pitcher’s head.
If you have great run prevention and poor offense you are likely an average or worse baseball team but you make your pitchers look even better than their true talent (at least their ERA not their record). Even a dumb jock type pitcher has a good idea if his defense is helping him or not on balance.
Conversely a pitcher on a team with a transcendent offense may look worse than he is (in terms of ERA not record) because when he gets a big lead he’s instructed to pound the strike zone and make the hitter put the ball in play. I realize we have better tools than wins and earned runs to evaluate pitching skill but they illustrate my point.
For real baseball teams who are trying to maximize their wins (not their gate receipts) it is pretty obvious that a win is a win and all the reasonable arguments to the contrary seam to apply to unrealistically good run scoring or run prevention.
I’m interested in the psychology of being a fan of a run prevention first vs. offense first team. I love watching brilliant pitching performances and athletic defensive plays. A great pitcher like Felix is more likely to have a really memorable game if his awesome defense can save his bacon once or twice in an otherwise dominant game.
I’d much rater watch a few dozen 2009 Mariners games than an equivalent number of games from the average .500 team because I’d be more likely to see that kind of performances that entertain me. I appreciate that lots of casual and fervent fans feel the other way or have no preference and appreciate 10-9 games as much as pitchers’ duels. I am sure I watched more of the Mariners last year because I thought we were seeing improvements in player evaluation and roster construction that could lead to a contender in 2010. Not many people want to watch a bad team with a crazy manager and an inept front office that has no hope of contending for years and years.
Second sentence should read “…run prevention or run scoring.”
A couple of other commenters have also alluded to this, but rlharr puts it best.
A run saved is NOT the same as a run scored due to Diminishing Marginal Returns, as shown by rlharr’s example. You can easily show this in a spreadsheet: if a team is outscoring its opponents by 10%, then runs saved are roughly 10% more valuable than runs scored (in terms of how much the team’s win percentage increases). If a team is outscoring its opponents by 20%, then runs saved are roughly 20% more valuable than runs scored. Etc. And vice-versa if the team is being outscored by 10%.
That addresses the question of runs scored vs runs saved. Derek also raises a more complex and detailed question: what about power? To answer that, one needs a model of how power translates into runs … OBP * SLG would be an example of a simple such model, but might be too simple and too imprecise (i.e. SLG% is not synonymous with power, though it is correlated with it). I haven’t worked out an example, though I suspect that power has diminishing marginal returns too (if nothing else, because pretty much EVERYTHING in life exhibits diminishing marginal returns), but I don’t have a worked out example.
For runs scored vs runs saved though, it’s very easy: just put the numbers into the Pythagorean formula. A run scored is not the same as a run saved (unless the team is a Pythag .500 team, i.e. gives up the same number of runs as it scores).
I suspect that the run scored by stronger offense vs. run saved by stronger defense is determined by who is pitching and where you are in the game.
My observation is that pitchers are very much influenced by how their teammates are batting. Speaking about an average starting pitcher on an average team, if his teammates are hitting well, he is more likely to pitch well. Conversely, if their teammates seem to be baffled by the opposing pitcher, they tend to not pitch as well. I suspect this is strongly true for starting pitchers, and less true for relievers. It is one reason real pitchers duels are so rare, and is one thing which makes an ace an ace- an ace seems less influenced and can stay strong in a game even if his team has fallen behind. The less accomplished pitcher just seems to press when he falls behind, contributing to the eventual loss rather than steadily producing at his best. That is what we admire when we see a well pitched game sometimes – a pitcher who just keeps plugging along making his best pitches even if he is behind, going against the typical psychological flow of trying to be perfect in hopes of not falling further behind. Eventually his offense might figure out something and then we admire the gutty performance on a day when he might not have had his best stuff.
So keeping a run off the board and keeping the starting pitcher from worrying about pitching from behind is likely more important while the starters are in, but the run that is scored later in the game which psychologically might represent the momentum shifting (a purely psychological happening…baseball games do not actually have any real inertia) means in later innings the offense run is more valuable because it either puts another nail in the coffin for the team in the lead, or it appears to be a momentum shift for the team which is behind. Late inning, a great defensive play which saves a run just keeps things mostly at status quo, not really contributing as much to the psychological story of the game, except it might add retrospectively to the seeming inevitability of the outcome the game is marching towards.
So, early in the game, defense saving a run more valuable since it helps the psychology of the starter. In later innings, I suspect offense runs are more valuable since it helps signal the psychological momentum of the game for both the team in the lead and the team behind, and which I believe then influences performance on the field.
This is not true.
Really. There is no evidence that pitchers are significantly affected by the score of the game. Truly great studies have been done on this, and it’s just not there.
Can you cite these studies? I am not aware of them and would love to read them, as it is quite counter intuitive based upon years of watching the game.
How many times I have seen a pitcher try to pitch too fine after he falls behind, especially when the opposing starter has the reputation of being the better pitcher? Hundreds.
How many times have I seen truly great starters just relax into the game when they have a comfortable lead – challenging the opposing fastball hitters with fastballs, also not worried about throwing a breaking pitch with a 3 ball count when they would be much tighter with it in a tight game? Probably hundreds again.
A study which disspells the notion that a pitcher pitches differently with a lead would be one I would love to read.
You are, if you think about it, arguing that pitchers pursue pitching strategies that aren’t as effective when they need them most.
I think there was one pitcher who turned up as a “pitch to score” guy and it might have been Roger Clemens. Tough to remember. I’ll see if I can dig it up.
But yup: doesn’t happen. Pitchers do their best all the time, more or less.
Ah, here’s a good study on this by Greg Spira.
Thanks for the link – I will digest this study.
And I am not proposing that pitchers pursue non-optimal strategies, only that situational pitching is real, otherwise there would not be any IBBs. And score certainly is part of the mix of the situation considered – “Do we pitch careful to this guy and not worry if we walk him?” certainly would never be said if you have a 6 run lead.
Situational pitching could include trying to work the power hitters differently if it is a one run lead.
So optimal strategy, given any particular pitcher/hitter combination, could and seemingly is influenced to some extent by the score. It is not too much of a stretch to posit that there are many decisions made as a pitcher goes through a lineup which take into account the score of the game at that particular point. And therefore, why not expect, that if it is early in the game, particular runs influence those choices more profoundly?
This is a bit different than the “pitch to the score” tone of the link you sent, but I need to digest it before shooting my mouth off too much (which might be too late to prevent anyway!!)
Thank you for pointing out this article to me.
I am not sure that this study really addresses whether a pitcher does or does not change how he pitches based on the score. It rather addresses whether or not a pitcher’s performance is influenced by the opposing pitcher’s performance (in terms of run support). I am suspicious that the results are just typical of most simple statistical studies of very complex independent events – that is, a return to the mean is to be expected.
We know there is situational pitching. Certainly that is not in dispute, is it? Otherwise why are there ever IBBs? Why pitchouts? Why relief pitchers – the ultimate situational pitching decisions made.
We know situational pitching is influenced by the score. So does situational pitching not have any influence on outcome whatsoever? I think this study does not show that because it doesn’t really capture data that would show if this is or is not something that happened.
Even if you could think of a way to study how score influences how a pitcher pitches, you have the difficulty of teasing out those times when a strategy was executed successfully, and those times in which it was not. Each of these subtle changes are usually a series, but are only easily studied as independent events. It makes my head hurt to think about how you could tease this info out of baseball stats.
I certainly don’t have the time nor data to conduct a proper study, but it would seem to me that if one had the time and data one could look at whether pitch selection was influenced by score, and what effect pitch selection had on pitcher effectiveness. Or perhaps you could study whether BB vs. balls-put-into-play ratios change based on score – I suspect it would go down if the pitcher has a significant lead as he would be more likely to challenge hitters. But how you could tease this data out, I just don’t know.