105 + league average for 57 games = 133.3 runs. So depending on how much you want to punish someone for requiring an extra roster spot or two, I would imagine the answer would have to be either after 140 or after 130.
What does “created 105 runs” mean? Right now, Bautista leads in Runs Above Replacement with 68.5, in 112 games. If someone created 105 runs in 105 games, that’d be historic.
What am I missing? Or have saberists gone so far up their own asses that they’ve lost track of what their own statistics mean?
Comment by The Kid Calling Out The Naked Emperor — August 19, 2011 @ 10:49 am
So the question is how you handle 57 games of production from somebody else. Do you ignore it completely and just judge the guy on his 105? Do you assume replacement level production (about 20 runs)? Average (about 30 runs)?
it is a question not based on any fact from what I can tell, just wondering how you handle a player that won’t play the entire season (PED suspension, injury, etc.) when comparing them to people that played the entire year
yeah, big money teams want elite production whenever they can get it, because even with unlimited resources, elite production is hard to find. You’d rather have a smaller amount of truly elite if you’re the Yankees than a lot of slightly above average. Whereas a team like the Braves would probably prefer a lot of above average to a small amount of spectacular.
yeah, just picked up on that. Though that seems kind of backwards to me. As an econ person, value is based on marginal value (value-replacement level), not the reverse.
It’d be better just to straight up ask what repl level should be. To me it doesn’t make sense to twice separate what’s essentially a guess from what you’re trying to get at. Unless you want to show people are inconsistent with their estimations of repl level.
No, it’s more of a thought experiment, and not to be taken literally or at face value. It’s essentially asking, how much below average do you think the replacement level is? In an indirect way. It is a bit poorly worded, but you don’t have to go off the deep end about it.
When considering the MVP award, why are you giving the 105 gm player credit for the runs created by his replacement, whether that replacement is replacement level or league average? My presumption is that the MVP award considers strictly the performance of the player in question.
you’re not giving the player value for the runs created by his replacement, you’re figuring out how much you’re losing with him not being in the lineup, and then comparing that to the players that played more games.
Total value to the team (That’s what MVP is, isn’t it?) is not a rate stat. Someone who contributes at higher rate for fewer games may not be as valuable in total as someone else who contributes at a not quite as good rate for more games. Galt’s right. Given the parameters of the question, and depending a bit on who plays the other 57 games, our 105 guy falls somewhere between the 130 and 140.
I think it’s just an organic way of making us give our gut estimate at what the replacement level should be without thinking about it (though if you break it down, of course you understand that’s the point of the question).
Surprised voting for Player E is so low. I would have thought that more people would think just in terms of number of runs created, and ignore replacement level.
Comment by strongbad56 — August 19, 2011 @ 11:18 am
That is the same thing as adding his value, you would essentially be figuring out how much you lose by the replacement level production being added to the 105 player’s production as well as the other players, you could do that or simply subtract the baseline replacement level production.Either way, it’s an attempt to measure total production you get at X position and determine an MVP based on that. I still would agree with Stephen’s point that replacement level production is largely (though not completely) irrelevant vs what player 105 actually produced as compared with other candidates for an individual award.
It seems to me the real question has nothing to do with replacement players, it’s instead how much we value durability in a player. This isn’t about what’s best for the team, or how much total value is added. It’s simply, “What’s better, a monster, but injury shortened, season or a super good, if not quite monster, season?”
I recognize that part of the argument, from a teams sake, is how much production they get from the back up player. But that has next to nothing to do with our options. The question is which player contributed the most, not which team received the most value from its options. It’s not even total player value, it’s situational value. Those 105 runs in 105 games would have, all else being equal, provide more wins than an equal amount of production spaced out.
But the fact is, the guy who produces 105 runs in 105 games IS more valuable than the guy who produces 105 runs in 162 games.
You put the first guy on the 60 day DL and put someone else in his slot. That he DIDN’T use a roster slot for 57 games is value he adds.
Claiming that the guy worth 105 runs in 105 games is worth the same as the guy worth 105 runs in 162 games is equivalent to claiming that someone who produces 0 net runs while starting in 57 games is not harmful. This isn’t true, someone who produces 0 runs while starting in 57 games is costing you millions of dollars above and beyond his salary.
There’s a reasonable question how much more valuable 105 runs in 105 games is than 105 runs in 162 games, but the first player is clearly more valuable, and to say he’s more valuable by what you’d have reasonably expected to get from a more or less free replacement player in those extra 57 games is a fair approximation. That’s not crediting him with the replacement players production, it’s crediting him with NOT using a precious starting slot for 57 games to produce nothing much.
Comment by Doug Lampert — August 19, 2011 @ 11:21 am
you’re figuring out the value of games played, essentially. We all agree that a player who created a half run per game more than average is more valuable if he plays a full 162 games compared to 105.
What they’re asking is where that line in between actually lies, where you’d be ambivalent. For that you need to figure out how much you lose by the player missing those games. If you don’t consider replacement level, you’re inherently assuming that either it’s an automatic out if he doesn’t play, or that his not playing is irrelevant. To figure out how much you’re losing, yo uhave to figure out what a theoretical replacement player is.
The point of their poll is to back door in to see what people think replacement level is.
JC Bradbury explains this much better than me in hot stove economics.
Yes absolutely, I agree 105 in 105 vs 105 in 162 there is an obvious answer for an individual award.
As soon as you start moving up the ladder in this poll it becomes highly debatable. The award is not “which team got the most value out of x position for the season” it’s an individual player award and therefore I think total production at x position is really only a tiebreaker where differences in runs (whatever measure)/game is highly different such as in your example.
Runs created is R+RBI-HR would be my guess. Looking at random players, 1 run created per game isn’t an obscene pace. For instance the Phillies have 4 players beating it and a 5th close so perhaps something else is meant or as others have suggested it is a purely intellectual question.
I would set my benchmark for a ‘full’ season at 150 games. Hardly anyone plays 162 games. Less than that it would depend on why they missed time with injury being the most valid excuse. Even there though I might knock someone if I thought it was an injury another player might have tried to play through. I have heard people suggest that Victorino is unlikely to get serious MVP consideration because he only is on pace to play 130 games. Now that doesn’t seem fair to me so perhaps I would consider how much I personally like a particular player. This kind of thing is a popularity contest, right?
also, I think using theoretical terms (runs created) and then giving what we could an assume the average player is really hurt the poll. It gets the more analytically inclined thinking in terms of replacement level, which sort of defeats the whole purpose of back dooring it like that. It just confuses the more “gut” types.
But Telo told me the obvious intent was for us to calculate replacement level from the given info. You mean he can’t read minds?
Comment by Nitram Odarp — August 19, 2011 @ 11:30 am
I understand your point, I agree with you that it matters….somewhat.
But I disagree that this poll is an attempt to find the replacement level value because the question is about an individual award. I think the question is better phrased at, “how much do you care about the exact replacement level value for an individual award?”
If we knew the replacement level value, we could easily determine where exactly the line is as far as pure value (or marginal value). It is not there because that’s not the point so much as it is, do you really care all that much for the purposes of MVP?
Halfway Between Ceiling and Floor
125 – 162x = 105 – 105x
20 = 57x
.35 = x
.35/.5 = 70%
70% of league average seems to be a decent guess at the replacement level.
Comment by Samus Aran — August 19, 2011 @ 11:36 am
That is the obvious intent, if you break the problem down logically. Your answer reflects what you think the replacement level is. The point of the poll is to not only gather the calculated answers, but to also get folks’ gut reactions (non-calculators) at what they feel like the replacement level should be. And it’s cool to see that it’s working out pretty well. People are choosing the actual replacement level.
I think Doug nailed it. Every team has only ~4400 outs to play with. Whichever team gets the most out of those wins. Every at bat, every pitch thrown carries an opportunity cost. A player’s value is always marginal — how much better he is than somebody else who can replace him.
But that’s why the WAR framework is so useful — it accounts for that. It uses that opportunity cost as the baseline. A raw counting stat like RC doesn’t do that. So I’d add replacement level production to the 105 and slot him in behind the 130 guy.
It’s interesting because we see this done implicitly in the way we handle HOF voting. Guys like Kirby Puckett who only have 2/3 of a career aren’t judged on just their accumulated counting stats. We look at both peak and career value. Judging a season should use similar logic.
So about 19.95 units is the replacement production for that time. If you purely cared about nothing but overall production and production loss that would put the 105 player right between the 120 and 130 player.
I will say that, in the interest of really seeing who was thoughtful about this, it would have been interesting to have skewed the options just a little bit more to make the best answer not one of the middle of the choices.
I find many of the responses interesting. So, keep posting, especially if you were confused, or had assumptions that comes into conflict with what you are reading. It might be worthwhile to try to get everyone caught up on some of these concepts, in a future post.
Doug – Yes. 105 in 105 is more valuable than 105 in 162. didn’t claim otherwise. The question is, rather, is 105/105 more valuable than 110, 120, 130, 140, or 150 in 162? Those missing 57 games do matter … they are 57 games that Mr. Awesome was unable to do anything to help his team, for whatever the reason. But how much?
Yea, for sure. There wasn’t anywhere else to go with the choices logically. But while it might be obvious to us what it means to put in the 160/162 choice, it’s not obvious to someone who doesn’t fully understand the question. I don’t think there’s any reason not to add a few fluff answers, just to sift out some more of your guessers, and to indirectly encourage people to fully understand the question before answering.
“The Kid Calling Out the Naked Emperor” huh? Jesus, you don’t think too highly of yourself. People have trouble knowing how to offer constructive criticism without coming off as a trite, pompous douchenozzle.
The point I was making is that this site employs advanced statistics which many of the commentariat don’t really understand. Yet, when pressed, they’ll rush to defend the author’s use of them. This was clearly a terribly structured hypothetical, and anyone who knows what “created runs” means could tell you that. Still though, people are defending it.
Comment by The Kid Calling Out The Naked Emperor — August 19, 2011 @ 12:40 pm
Not gonna lie, read the poll at work and thought the question was simply who’s the most valuable. Knock my reading comprehension if you’d like but I am sure others did thr same and will else the results.
I think it depends on whether why he missed those 57 games. If he was injured or platooned, then I’d put him somewhere near the end. If he was benched or in the minor leagues for those games (if it wasn’yt his fault he missed them), I’d put him first.
I think it’s clearly about more than just finding where replacement level is, in order to see how much you lose with a player who’s awesome, but unhealthy.
Here’s my take on this thought experiment. A guy who’s good enough that he’s essentially producing 1 run every GAME is pretty awesome. There’s something to be said about concentrated value, rather than more even distributed value over the course of a season.
So here’s how I frame this thought experiment. Assume that each of the MVP candidates is the only above average player on his team, and everyone else is of average value (Except, of course, the replacement player who fills in for our 105 game player). Replace each candidate with an average player, and you have a .500 team.
I won’t go into very hard math here, but Player F’s team is therefore (about) two wins above average, Player E is three, etc. and Player A’s team is and 88 win team.
In the 105 games played by the player in the question, an average team would win 52.5 games. This player’s team won 57.5 games . So, how many games did they win of the remaining 57? Assume an average team at all positions except his, and a replacement player at his position. What’s replacement level? About 20 runs per season below the league average at a given position. Over the 57 games, about seven fewer than league average. So 0.7 wins below average. That’s how far below average that team is WITHOUT their star player. Over 57 games, they win 27.8.
So they won 57.5 games WITH him, and 27.8 without him. That puts them at 85 wins, rounded down. Three wins above league average (plus change). This is absolutely isolating his contribution, assuming replacement level replacements and league average competency everywhere else. So it seems like he should slot in between players D and E.
it’s a simple math question. a player would need to be better than replacement (50% runs created in games) over the 57 games player X missed. that means the “correct” answer is after B, because player C creates less than player X + replacement level, whereas player B creates more runs than player X + replacement level.
Why is replacement level necessarily at the 50% level, and not, say, at the 75% level?
Comment by tangotiger — August 20, 2011 @ 10:59 am
Essentially, is it fair to consider a guy mor MVP that played less games but produced more per game? How many games played until it really doesnt matter anymore. The question is essentially asking which is better: 50 games with 60 runs created, 80 games with 90 runs created, 100 games with 105 runs created, 130 games with 125 runs created? How much should missing games or not playing in games discount ones contribution to a team? That is what I understand the question to be about.
Comment by PadresFuture — August 20, 2011 @ 11:02 pm
Player F is Josh Hamilton, right? That’s what I’m assuming here and I think the snag I face when considering the question is that yes, he did provide a ton of support and gave the team a huge boost while he was available last year. Did that momentum allow them more flexibility when he went down with the rib injury? His total impact is supposedly accounted for in either WAR or RC, so I would say the absolute value of whatever metric you employ would be your consideration. But, the MVP doesn’t work like that, does it? There have always been debates over player impact on his team’s winning versus player’s production as basis for MVP votes. That’s what this philosophical exercise has brought up for me, not as much the replacement-level question.
Comment by Tsunamijesus — August 20, 2011 @ 11:55 pm
I only browsed the first 15 or so comments so my contribution might be echoing someone else.
It makes a significant difference if the player in question is a platoon player with radical splits. It also makes a difference — though it probably shouldn’t — if the player did this down the stretch because he missed the first part of the year due to an injury or, hell, maybe he had been called up.
I voted for the just after 150 RC guy in 162 games.
Comment by LowcountryJoe — August 21, 2011 @ 12:54 am
I guess I wasn’t thoughtful about it at all than. I took the poll literally and chose Player F because I don’t believe a starter who misses 1/3 of the season is all that valuable. With missing 60 games, his team is then forced to find a replacement player to fill his position, who may be good or bad, but whether he is or not is beside the point.
Although another way to look at it is, this player in question IS the replacement player, promoted to the starting lineup in June and putting up astonishing numbers the rest of the way.
The former guy just isn’t MVP material. The latter may be.
Comment by Sultan of Schwwingg — August 21, 2011 @ 9:02 am
Tango structured the question to avoid that response. The poll is bogus.