By In Stuff

Another Tango Question

Tom Tango has another fascinating question going over at his blog. The idea is to try and figure out what replacement level in baseball means — not on a mathematical level but on a gut level.

It seems like many people simply cannot get their arms around the replacement level concept — and I will admit to having some issues there myself. The way I usually think about it is as the sort of player you can easily acquire or call up should someone on your team get hurt.

Look at the Yankees:

Mark Teixera is hurt, replaced by Lyle Overbay. His WAR this year: 0.2.

Curtis Granderson is hurt, replaced by Vernon Wells. His WAR this year: 0.4.

Derek Jeter is hurt, replaced by Jayson Nix. His WAR this year: 0.4.

Alex Rodriguez is hurt, replaced by Kevin Youkilis (minus-0.2 WAR) who is replaced by David Adams (minus 0.3 WAR).

The Yankees tend to get bigger replacement level names, but the concept is pretty much the same whether it’s a once good player like Wells or a lifelong utility guy like Joaquin Arias or a Class AAA guy called up to stabilize things for a good prospect not quite ready for the big leagues. Replacement level talent is what the words suggest – the level of a player you could find as a replacement.

But, I readily admit that the concept can be a bit wonky. There are players in baseball who get a lot of at bats who are BELOW replacement level … so how does that figure? Sometimes, a replacement player is actually pretty good, sometimes he is much worse than replacement level, it’s can be a complicated and hard-to-pin down concept.

And it should not be: Replacement level should be something that you feel naturally, something that just sort of makes sense without any real thought. This is something that Tango has spent quite a bit of time considering, and he offers a cool way to think about tit.

Consider the following four pitchers:

Bob Gibson

Roy Halladay

Pedro Martinez

Nolan Ryan

Now, here’s the thing: You can have any of those pitchers for their entire career. But you get the career exactly as it happened — there are no exchanges or refunds. You get their strengths and their weakness. You get their greatness and their limitations.

So you can have:

Pedro Martinez, ERA- is 67 (Gave up 33% fewer runs than league average), pitched 2,827 innings.

Roy Halladay, ERA- is 76, pitched 2,721 innings.

Bob Gibson, ERA- is 78, pitched 3,884 innings.

Nolan Ryan, ERA- is is 90, pitched 5,386 innings

Obviously, this is oversimplifying things — but you can see your challenge. Pedro Martinez is the most dominant of the pitchers — not just for his career but he had the most dominant individual seasons as well — but he threw about half the innings of Nolan Ryan. You will have to fill those innings with someone else. But who? Replacement pitchers.

What about Gibson against Ryan — considerably better run prevention, more dominant seasons, but there are still 1,500 innings to make up. Again, who will fill the gap? Obviously the answer is: Replacement pitchers.

So, for fun, do this little exercise — rank the pitchers you would want in the order you would want their careers. Put your ranking in the comments section.

And what you say say something about replacement level. What does it say? We’ll ask Tango.

Print Friendly

76 Responses to Another Tango Question

  1. P Brady says:


    I’d rather have the far more dominant pitchers for shorter than Ryan for longer.

  2. Pedro. Peak is more fun and more helpful in trying to win a championship.

  3. Brian says:

    Pedro, Bob, Roy, Nolan

    I assume I can find a replacement level player (ERA- 100) to fill whatever innings gap they have versus Nolan. When you do that, you get these results, over 5386 innings:

    Pedro + Replacement (82.67)
    Bob + Replacement (84.13)
    Roy + Replacement (87.87)
    Nolan (all by himself) (90)


    • Scott says:

      ERA- 100 is “average,” not “replacement level,” I believe.

    • tGP says:

      A replacement level pitcher wouldn’t give you a 100 ERA-. That’s league average, and by definition, half the league is below average. Average pitchers don’t grow on trees.

    • Brian says:

      Dumb question — wouldn’t then both those be the same?

    • Brian says:

      So, if you want me to assume that I can’f find an average pitcher, then the math obviously changes. As someone that once tried to convince himself that Andy Hawkins was a lot better than the media gave him credit for, and thought that Bob Wickman’s rookie season was sending him on his way to Cooperstown, I am WELL aware that average pitchers don’t grow on trees.

    • I think what Brian says above is the point. You should be able to replace a star pitcher (through free agency) with at least a league average pitcher.

      If you really replace these pitchers with replacement level pitchers, well, things might change.

    • Jimbo says:

      Using Brian’s technique but adjusting the value of the replacement pitcher.
      If a replacement pitcher gives you ~107 or less then Pedro is best.
      If a replacement pitcher is between ~107 and ~121 then Gibson is best.
      If a replacement pitcher is over 121 then you want Nolan.

    • Tangotiger says:


      And replacement-level is equivalent to the neatly costless solution: free agent pitcher signed to a minor league contract.

    • invitro says:

      I would use this technique with whatever the actual ERA- of a replacement pitcher is.

      I’m curious about the statement “You should be able to replace a star pitcher (through free agency) with at least a league average pitcher.” Is this really true? Is it what actually happens? I don’t know whether to believe it without some evidence.

      I think this question is basically getting at a better way of comparing careers of different length and different peak level. I think the best answer probably depends on how players are actually replaced. Certainly it is not true that players are always replaced with a replacement-level player when they leave via retirement or free agency.

      Now to read Tango’s blog.

    • Mark says:

      If you could easily replace a star pitcher with an average pitcher — which is one of those things I believe sounds ‘logical’ on its face but reveals itself as absurd when you look at it — then there really wouldn’t be such a thing as replacement level. Or average, for that matter — if it’s so easy to find an average starter, why are there so many just-barely-major-league starters in the game? Ask the Royals of the past 10 years or so how easy it is to find an average SP.

    • Mark says:

      If you could easily replace a star pitcher with an average pitcher — which is one of those things I believe sounds ‘logical’ on its face but reveals itself as absurd when you look at it — then there really wouldn’t be such a thing as replacement level. Or average, for that matter — if it’s so easy to find an average starter, why are there so many just-barely-major-league starters in the game? Ask the Royals of the past 10 years or so how easy it is to find an average SP.

  4. The real debate, IMHO, is between Pedro and Gibson. Halladay is last, as he offers (at this point) about the same number of innings as Pedro but with less dominance. That could change if he comes back strong. I would rank Ryan next to last, as I can live with the replacement for 1500 innings to get the better rate from Gibson (although that’s a ridiculous number of innings). The question then is whether I want the shorter career and better rates that Pedro gives me.

    I am going to say yes, and go for it (ie championships) in a shorter timeframe and rank them this way:

    1) Pedro
    2) Gibson
    3) Ryan
    4) Halladay

    I could easily flip the top two and hope that the several extra seasons that Gibson gives me help me field a better team in one of them. I would not take either Ryan or Halladay over the other two, at least not as the challenge is framed.

  5. rob says:

    I think I would go Gibson/Pedro/Ryan/Halladay. Gibson seemed to have the most balance on dominance and innings pitched. I agree with Pedro due to the sheer apex of his dominance. Also, I always thought Ryan was overrated, but to have that many innings eaten and done well can really let you focus elsewhere on you’re roster.

  6. Free agency is the flaw in this argument. Gibson is the only one you can actually have for the entirety of his career. Hence I’d have to pick him.

    I don’t see why people have trouble “feeling replacement level naturally.” What team do you root for? What key players have missed half a season recently? Who replaced them? That should explain it.

    • Frank says:

      I agree with you on the free agency point. Different era, of course. Who’s to say that Gibby would not have flown the Cardinal coop had he had the chance?

      I disagree with the way you suggest getting the feel for replacement level. Just look at the Nationals. Bryce Harper goes down and is replaced, for the most part, by Tyler Moore. (Blech) Danny Espinosa gets sent down and is replaced by Anthony Rendon. (Why didn’t this happen sooner?)

      “Replacement level player” is being used as a universal constant. He is a fictional player. If Replacement Man were real, teams (at least pennant contenders) would immediately take out a player with a negative WAR and put in Replacement Man. Then they would clone him and put him in for everyone else with a negative WAR.

    • Skirra says:

      “If Replacement Man were real, teams (at least pennant contenders) would immediately take out a player with a negative WAR and put in Replacement Man. Then they would clone him and put him in for everyone else with a negative WAR.”

      That would be such a Royals thing to do if cloning became possible. They would be the last to adopt and then to save money they would clone Replacement Man (with perhaps an extra touch of grit) instead of cloning All-Star caliber players.

    • invitro says:

      “If Replacement Man were real, teams (at least pennant contenders) would immediately take out a player with a negative WAR and put in Replacement Man.”

      This is false, as easily seen from evidence. I don’t think many teams use WAR to value players; see Betancourt. Also, even if a team believed in and used WAR, there is a gap between a player’s actual WAR, and their true value, or what they can be expected to produce in the future.

      Cloning is unnecessary, as the definition of replacement-level requires an unlimited supply of replacement-level players. Well, I think it does… how many replacement-level players does the model assume exist? 100? 1000? 10,000?

  7. Well, if you’re looking at this as career value, then Halladay is last, and it’s close between Martinez, Gibson and Ryan.

    But if you look at which pitchers career would give you the best chance at winning the most championships, then you really need who had the most seasons far above replacement level. In other words, for winning championships, 5 MVP caliber seasons mean a lot more than 15 seasons at a 1/3 of those season’s WAR or (100 – ERA-) values.

    Pedro had 3 all-time seasons, and 7 other seasons at an excellent level.

    Gibson had 3 all-time seasons, and 6 other seasons at an excellent level.

    Ryan had 0 all-time seasons, and 7-8 other seasons at an excellent level.

    Halladay’s best seasons are debatable whether they are all-time or just excellent, but there’s 8 total.

    So, HUGE debate between Gibson and Martinez on “peak” helping to win championships. Gibson deserves a historical edge based on pitching in a 4 man rotation, but in terms of what he would accomplish in similar situations, they are darn close.

    In my mind, while Ryan’s career may be more valuable over “replacement” level than Halladay, assuming replacement level is a top AAA player, if you make replacement level being a league average starter you could sign through free agency, I think it’s really close, with maybe Halladay having the edge.

    So, in the end, viewing the value of these pitchers in terms of helping your team win championships season by season, and with replacement level being viewed as who you could sign in free agency to replace these pitchers, I rank them as:

    1a. Gibson (give him the slight edge for being able to pitch more innings per start)
    1b. Martinez
    3. Halladay (gets edge because I’m assuming I can replace him in his lesser years than Ryan with league average starter)
    4. Ryan

    • I think the question assumes you replace these guys with a replacement level person. I think the whole point of Joe’s post is that even the Yankees cannot simply replace their players with a league average player through free agency. Granted, their losses are due to injury, but still.

    • zeke bob says:

      What is your qualification for an all-time season? I think Ryan’s ’73 and ’77 seasons deserve consideration – 7.8 WAR and 8.3 respectively. He also had 26 and 22 complete games, respectively.

      Pedro’s three highest WARs were 8.2, 8.4 and 10.1, so Ryan would at least compare to those bottom two.

  8. Assuming you had your job forever, the pick, of course, is intentionally very close. Ryan and Pedro and Gibson all have similar lifetime WARs. (In the real world, I’d want to keep my job and pick Pedro, like everybody else, and retire when he does).

    Brian had the right idea, but replacement level starting pitcher ERA, as pointed out by others, is more like an ERA 20 to 25% worse than league average. Just to pick one example, Bruce Chen of the Royals last year had a 81 ERA+ (which is the same concept — leage ERA/ERA x 100) over 190+ innings, with a WAR of 0.3. So, you are replacing 190 innings a year with an era of 20-25% worse than average.

    This really puts the value of Ryan’s longevity in perspective. To refigure Brian’s numbers (assuming replacement is 22% worse than league average):

    Pedro + replacement would have an ERA about 7% better than league average
    Gibson + replacement would have an ERA about 10% better than league, just like Ryan. (Halladay, while awesome, of course won’t be anyone’s rational pick).

  9. Ian R. says:

    Ryan, Gibson, Pedro, Halladay. (As an aside, isn’t it weird that we call three of them by their last names and one by his first?)

  10. Marty Winn says:

    I think the rules of this argument aren’t really fair. You’d have to pay Ryan a ton of money over ~26 seasons. Sure you’d pay Martinez more per season but it seems like you’d have money left over to pay for an average pitcher rather than a replacement pitcher to make up for the innings Pedro does not get to pitch. I think Pedro is the most valuable of the pitchers.

  11. I think the headache of constantly finding new replacement players would wear me down. Here’s my ranks:
    1 – Ryan
    2 – Gibson
    3 – Pedro
    4 – Halladay

  12. Pedro/Gibson/Ryan/Halladay is what I decided.

    BB-Ref WAR suggests Pedro/Ryan/Gibson/Halladay.

    I do think you can add complexity to the discussion by hitting on things like what Marty Winn said about salary. And also would you rather have a guy who’s “guaranteed” to give you 4 WAR every year that you can build around for the long haul or a guy who can give you 7 WAR per year for a shorter time period. I can imagine arguments in both directions.

  13. Ian says:

    So if you take Pedro that means you also get 2500 innings of Cole De Vries? Ugh.

  14. Phil says:

    Martinez/Gibson/Ryan/Halladay. I find that “replacement level” is right around ERA+ of 84, which is significantly worse that “average,” and shouldn’t be too hard to find to fill the innings: all the 2000-IP pitchers who have retired due to ineffectiveness (which excludes Joss, Cicotte, Shocker, Carlson, French, Koufax, and Kile) averaged 84 ERA+ in their swan song seasons.

  15. Steve O says:

    That’s the fly in the ointment, right? If replacement level players grow on trees, why do teams play guys with negative WAR?

    In the past, it could be assumed teams simply didn’t *know* that a player was performing below replacement level, but that’s no longer true. Every team has analysts and statisticians in the front office. So why play a guy who is hurting the team?

    It could be an age/development issue. If a young guy is putting up a negative WAR, you might want to let him play through it in the hopes that he develops. See: Gordon, Alex in the first half of 2007 and Brandon Phillips in 2003.

    It could be a money issue. A guy with a big contract who is under-performing (Chone Figgins from 2010-2012 seems like a good example) will get playing time because teams usually aren’t very fast to cut bait when there’s a lot of money involved. Plus, it makes the front office look like dupes.

    But sometimes it is neither; sometimes a player who isn’t making a lot of money and who isn’t a prospect anymore still gets PAs even though they’re killing the team with a negative WAR. The assumption is that Replacement Man is out there, ready to step in and steady the ship with his 0.0 WAR production. I’m not saying the assumption is wrong; just saying it might not always be that easy.

    • MtheL says:

      Good analysis, but you miss two other reasons: (1) What if you don’t have a replacement level player to supplement the negative WAR guy with? Either (A) your minor league guy is not yet ready for the big leagues or (B) you simply don’t have a guy in your minors at that position who is replacement level. (2) So if you don’t have anybody available, theoretically you can go pick someone up. But does that make sense for a team like the Astros or Marlins, who aren’t going anywhere this season anyways? You can either (A) trade for someone or (B) see if you can get a free agent. But it certainly doesn’t make sense to give up a player in a trade, even a relatively not valuable one, if all you’re getting is a replacement level player for a team that isn’t going anywhere anyways. And signing someone assumes you can find at least a replacement level player at the position (not always possible) or are willing to spend some money for a decent free agent (again, why bother if you’re the Astros or Marlins, or why bother if you have a raw talent in the minors who may not yet have replacement level value but will by the next season?).

    • Steve O says:

      I agree, for the Astros and Marlins it makes no sense to start somebody’s arbitration clock early, but even hopeless teams do some roster wrangling to try and squeeze out more wins. Sometimes that goes to an absurd level (remember the hapless Pirates acquiring the aged Matt Morris and his bloated contract?), but usually it’s just standard deck-shuffling/waiver-wire playing. I think a good example is the Astros signing Trevor Crowe in the off-season. Houston knew they were going to be a last place team, but even so, they were trying to grab some “replacement” type players. Crowe’s probably a picture-perfect example of a “replacement” player. 0.1 WAR so far this season for the Astros, plays all OF positions, and any team could have picked him up in the off-season.

      But outside of the Astros and Marlins, occasionally you see a contender playing a terrible, negative WAR guy who isn’t young and/or expensive, and if replacement players really grew on trees, you’d think it would be easy to replace a negative WAR guy with a 0.0 WAR guy.

    • invitro says:

      “So why play a guy who is hurting the team?”

      The GM/manager/coach doesn’t believe he is hurting the team.

      I am curious if there is any hard evidence that teams are constructing their rosters by WAR value more so than in the past.

    • Rob Smith says:

      If WAR value was used, it would be hard to explain why BJ Upton and Dan Uggla are playing every day…. except obviously that they both make a ton of money and everyone is hoping they’ll do better. In Uggla’s case, he’s been pretty bad for more than this year, so he’s a lost cause & I’d gladly take a replacement player and plop him down at second.

    • Steve O says:

      invitro: Could be that the team doesn’t think the player is hurting them, but I don’t think everyone is that oblivious. It could be that they’ve tried to find a replacement but haven’t been able to acquire a serviceable one, and in the meantime, it’s good PR for the manager/coaches/GM to say that “Yeah, Johnny Grit is a heck of a player out there, and he does the little things, and he’s a winning ballplayer, the kind of guy you want to go to war with…” Sure, it’s a mish-mosh of boilerplate nonsense, but it sounds a lot better than saying “Yeah, Johnny Grit stinks on ice, but we can’t find anybody better. Oh, well…”

      Rob: Yeah, that’s the other possible scenario: a guy who doesn’t make a ton of money, and isn’t especially young, but who is performing well below expectations. If someone used to put up great numbers, and they’re currently terrible for seemingly no reason, it might make sense to stick with them and see if they can get back to their old level of performance.

  16. BobDD says:

    The context of the rest of the team/competition might make the decision for you. If the only way you can compete for the title is with a pitcher with 7+ WAR, then you NEED Pedro or Gibby. If your team needs only half that – 3.5+ WAR then you win more with Ryan (Gibby second, but Halladay closing). Why would you sign Ryan if he wouldn’t be enough to get you there? Or why would you sign Pedro if you knew he could only get you there half as often as Ryan? So the context can make any of Pedro, Gibby, or Ryan the best ingredient for the maximum number of championships. That’s using Joe’s parameters, not Brian/Marty/etc. changes.

    If anything, looking them all up in Baseball Reference impressed me all over again with Roy Halladay – that’s quite a sustained HOF peak there.

  17. Mark says:

    Speaking of Gibson, why did he only start 34 games in 1968? Were the Cardinals experimenting with a 5 man rotation?

    • MtheL says:

      Mark – They had a 5 man rotation that year, and 4 of the 5 pitchers (including a young Steve Carlton) produced ERA’s under 3.00. The starters: Gibson (ERA 1.12), Carlton (2.99 at age 23), Nelson Briles (2.81 at age 24), Ray Washburn (2.26) and Larry Jaster (3.51 at age 24).

    • Rob Smith says:

      Interestingly, when you look at it, Gibson pitched 304 2/3 innings in those 34 starts & had 28 complete games. Multiply 34 starts times 9 innings and you get 306 innings. So, wow, Gibson averaged nine innings per start!!!

  18. mgarbowski says:

    I think Marty Winn above is on the right track here with his comment. This is not a standard Replacement Player scenario. You’re not getting a replacement for Pedro for a month, or even a season. He pitched 7 fewer years than Ryan. That’s 7 years when Pedro’s roster spot is open and his salary is zero. You should be able to get somebody to fill that gap who is substantially better than standard replacement level. I’d take Pedro. By this logic nobody would ever want Sandy Koufax.

    • BobDD says:

      The subject of the article is about replacement level, so changing that parameter would move us away from the exact idea Joe is trying to get to.

    • Rob Smith says:

      I think the point is that Joe is creating a false choice. As noted, you have to take Pedro with the shortened career. It’s a false choice because you don’t replace Pedro with a replacement player. You could, but most likely teams roll that salary into a Free Agent, or bring up a young stud. So, though Ryan was inoordinately productive into his 40s, that doesn’t mean that Pedro’s ultimate replacement wouldn’t have been close to, or as productive as Ryan in his 40s. Bottom line, it’s a silly concept that Joe is proposing & not based on any real world scenario. Nobody purposely puts a replacement level player into their starting rotation, though sometimes a player underperforms and it works out that way.

    • Tangotiger says:

      The point is that if you replace Pedro with a league average pitcher, you’ll have to PAY that league average pitcher his 6MM$ or 10MM$ each year.

      The point of replacement level is that you’ll sign a pitcher to a minor league free agent contract (or a bunch of them), and then get the best one of the lot to pitch the season for you. That’s the (nearly) cost-free solution.

      Pedro at 25MM$ a year x 15 years or Ryan at 15MM$ a year x 25 years is the same cost. And the missing 10 years is the minor league free agent pitcher.

    • “The point of replacement level is that you’ll sign a pitcher to a minor league free agent contract (or a bunch of them), and then get the best one of the lot to pitch the season for you. That’s the (nearly) cost-free solution.”

      “And the missing 10 years is the minor league free agent pitcher.”

      Do these minor league free agent pitchers include players that you’ve drafted and are still in their pre-arbitration years and playing at league minimum?

    • Tangotiger says:

      They are free agent pitchers, so, they are not part of the arbitration or pre-arb group.

      They are signed to a minor league contract. So, think about pitchers who are 27-32 years old, and no team wants them, and the only deal they could get was a minor league one.

    • But I don’t wanna sign 27-32 year-old pitchers to minor league contracts!

      My strategy would be to retain the services of the very dominant pitcher (Pedro), hope each season I attain the maximum IP he can provide, but have a backup plan in case he can’t, with the backup plan being a strong farm system with young, pre-arb pitchers waiting to be tested.

      I already know with great confidence that the 27-32 year-old pitchers signed to a minor league contract are replacement value. That is unacceptable! If my farm system is up to par, and my scouts, instructors, and coaches are competent and reliable, then I can promote with confidence to the Bigs — when Pedro is unavailable — an above-replacement value (ARV) pitcher at minimal cost.

      But presuming I likely will not have too many B-plan ARV pitchers available in any one season, I do not want to fill my rotation with five Pedros. I think I’ll go with three Pedros and two Nolans, because then if all three Pedros are healthy come the post-season, they can dominate a seven-game series while one Nolan starts one game and the other serves the long-relief role (or they both start if one Pedro is unavailable).

      Which raises another point — given $X for team salary for any given season, it becomes more than Pedros and Nolans, and instead becomes about complete team construction, where the depth is located at the positions, and forming a team that (1) maximizes PEAK performance but (2) minimizes REPLACEMENT value use by (3) having depth and a strategy to utilize ARV when PEAK is unavailable or not present at a given position.

    • Tangotiger says:

      You have 375MM$ to spend for one pitcher slot.

      You can give 25MM$ a year x 15 years to Pedro, and then fill in the other 10 years at 0.5MM a year with a free agent pitcher signed to a minor league contract.

      You can give 15MM$ a year x 25 years to Ryan.

      Choose one.

    • I realize I’m not playing along, and I see your and Joe’s point/objective, but why form the question in that fashion? It seems unrealistic because (1) it is outside the scope of the reality of roster construction, (2) why would any logical GM sign a pitcher to a 15- or 25-year contract, and (3) we can estimate/project Pedro’s and Ryan’s future production, but this “perfect” knowledge that is involved is also unrealistic.

      Therefore, wouldn’t one’s feel for replacement level in your hypothetical not translate to reality?

      But according to my own desired strategy, and also what appears to be the majority in other responses, the boom-and-bust strategy should win more pennants, so Pedro and Crappy Pitcher is preferred to Ryan.

      So perhaps that simply maintains that people will accept replacement level in their team construction, as a cost needed to benefit from non-durable PEAK value that will optimize the chance to win it all.

    • Rob Smith says:

      Tangotiger is creating another false choice. Also, making the assumption that Pedro is a $25M pitcher and Ryan is a $15M pitcher based on their career length or some other unstated criteria is just random. Point is, you pick whoever you think is better. When their career ends, you replace them with your best possible option, not with some theoretical avg replacement player.

  19. As several people bring this up, I wanted to point out that replacement level players are not, I believe, “average.” They will produce negative WAR, OPS+ below 100, etc. The reason so many teams field replacement level players with negative value is that most replacement players have negative value. In fact, I recall reading an article a couple of years ago that there there are more major leaguers below the average than there are those at it or above it, because the truly great players (Pujols was the example at the time) skew the mean.

    • Mark Daniel says:

      Benjamin, a replacement level player produces a WAR of 0. You are correct that they will produce an OPS+ or ERA+ below 100. So what you would get out of each pitcher + replacement pitcher would be essentially their career WAR.
      So, as Michael Haynes pointed out above:
      B-ref WAR:
      Pedro 86.0, Ryan 83.8, Gibson 81.9, Halladay 65.5
      And from Fangraphs WAR:
      Ryan 106.5, Gibson 91.4, Pedro 86.8, Halladay 67.9

      Using these two WAR numbers, it seems the best bet would be Ryan. I don’t know why one would pick B-ref over Fangraphs, or vice versa, however.

    • Frank says:

      Benjamin – Your comment – whether correct or not – is one of the principal reasons people can’t get the heads around WAR.

      Mark – Your comment alludes to another principal reason people can’t get their heads around WAR: Why are there two of them? Why is Ryan 20% more valuable under the Fangraphs system? I know there are reasons, but when they take a day and a half to explain and you still come out confused, then the whole thing is devalued. Sometimes I hear the response of “trust us on this” and I feel like I’m listening to a government official talk about the National Security Administration.

    • Ian R. says:

      For this sort of comparison, absolutely use bWAR over fWAR.

      Fangraphs WAR uses fielding independent pitching (FIP), which makes adjustments for the quality of defense and batted-ball luck. Baseball-Reference WAR is based on the actual runs the pitcher allowed. fWAR is a better predictive tool because, as mentioned, it filters out the effects of luck and defense, but if you’re doing a historical comparison of pitchers (especially in really large samples), bWAR is a better measure.

  20. “The subject of the article is about replacement level, so changing that parameter would move us away from the exact idea Joe is trying to get to.”

    Then change the scenario, because Marty Winn and mgarbowski are presenting reasonable arguments for why this “whatif” is flawed regarding replacement level for pitchers.

    Looking at one particular season instead of careers is likely more apropos to focus on the ‘exact idea Joe is trying to get to’.

  21. Nick Vlahos says:

    Gibson averaged 7.87 innings/start.
    Martinez averaged 6.65.

    Somebody has to pitch that extra 1.3 innings (or 4 outs). Do you want it to be a guy like Mitch Atkins?

    Gibson by a mile. Then Pedro.

  22. Tux says:

    One thing that amazed me when I went to look this stuff up: Nolan Ryan was more than just a guy with a locomotive of a fastball. He was wild, but by lots of measurements his numbers are every bit as impressive as Gibson and Pedro (Halladay, eh. He’s clearly #4 on this list – a great pitcher, but not all-time elite like the other three guys here). And if you look at Ryan’s numbers late in his career – specifically with Texas – he’s just as good if not better from an adjusted ERA perspective.

    But of the three, I gotta take Pedro. Ryan threw harder, and Gibson was meaner (and of all these guys, is probably the best money pitcher of the bunch – 9 WS starts, 9 complete games, 92 Ks and an 8-1 record. I know it’s small sample size but damn was Gibby good in the postseason), but Pedro Martinez is my Sandy Koufax. I’ve never seen a man mix pitches and make hitters look so utterly helpless, and yes, a large part of this is due to the fact that I’m too young to remember Ryan and I was -13 years old when Gibson retired. But as far as pitching packages go…Pedro had three of the best pitches I’ve ever seen. And this was later in his Sox career, after he’d lost the nitrous on his fastball and had to get by on guile because he could only hit 94 two or three times a game. The man was a magician, and if that’s romantic and sappy and not good business sense I frankly don’t care, because Pedro’s my guy and I can’t be objective here, and the addition of the steroid era makes this an even harder question for me to get over.

    That said, Nolan Ryan has the best pitcher’s fight out of all of these guys. And that has to count for something, but I’m not sure what.

  23. rdcobb says:

    1. Gibson
    2. Pedro (on peak value as a tiebreaker)
    3. Ryan
    4. Halladay (this seems clear)

  24. Frank says:

    Here’s another way to look at this question: With any of the guys on the list not named Gibson, could the Cardinals of the 1960’s won three pennants? (or, perhaps, more than three?)

    Pedro? Perhaps not durable enough. Gibson had 240+ innings 9 out of 10 years; Pedro once his entire career. Which brings up another issue about replacement players: For Pedro, you have to have a good bullpen to sustain him – 45 complete games for his career. Gibson was his own bullpen – 255 complete games.

    Ryan? Not consistent enough.
    Halladay? Not good enough over all.

  25. Aaron Reese says:

    Ryan, Martinez, Gibson, Halladay.

    Every single time I listen to smart baseball people, they say that innings pitched is the single most underrated stat for a pitcher. I wouldn’t get more production from a single player than from Ryan. And he would stabilize my rotation for 27 years. 27! No replacement needed. Every fifth day, there’s a good chance to win and my team’s bullpen could take a nap, so I wouldn’t need to worry about wearing out their arms and replacing them. I wouldn’t have to worry about any replacement-level arms already in the bullpen. Even in his worst seasons, he was pretty much average. He also got rocked, and I mean truly beat up, very few times. He didn’t have an ERA over 4.00 until he was forty-six years old. Studly.

    It’s hard to pass up probably the most dominant pitcher ever (213 ERA+ over 7 years. Whoa). If I’m aiming at pennants, he gives the best shot at them. After I lose him to age and fragility, I’d have to scuffle through those painful rebuilding years, with all the very sad and bitter 30-year-old AAA players, who come the majors and hopefully keep their ERAs under 6.00, but hey, flags fly forever.

    Bob Gibson once pitched 10 complete games in a row and allowed two runs. #@!$%-ing TWO. I don’t care if it was a pitcher’s year, that’s mind-boggling. He had a blend of longevity and peak. Again, the bullpen gets most nights because he just won’t break down. Also players that faced him said he was scary (like Bagwell on the mound, but throwing at you). I don’t know how much that is worth, but it’s awesome.

    I’m not gonna roll Halladay out of bed, you know what I’m saying? His peak might be better than Gibson’s, but he didn’t throw 200 innings until his fourth year and probably won’t ever again after 35. I don’t think I can fault him for being my fourth-preferred pitcher in this group. He just didn’t have enough time as the beastly beast that he was.

    • Rob Smith says:

      Look, I had Angels season tickets in the 70s and there was nothing better than a Nolan Ryan pitched game. Of course the Angels had nothing but Ryan in those days…. but…. anyway, people forget that in his most dominant years, he could be pretty inconsistent and walked over 200 batters. The common wisdom was that you had to get to him early though, and once he settled in and got his groove, it was game over. But, there were more games than I care to remember where Ryan didn’t get past the first couple of innings. Lots of walks, hit batters and then a bases clearing hit. Yes, it happened, though we tend to block out such memories.

    • Wilbur says:

      They (broadcasters) said the same thing about Gibson – you had to get him early or it was ballgame over.
      But they also said that about Seaver, too.

    • Wilbur says:

      They (broadcasters) said the same thing about Gibson – you had to get him early or it was ballgame over.
      But they also said that about Seaver, too.

  26. Rob Smith says:

    Well, at least with Gibson, he wasn’t walking the bases full on a regular basis…. one of those nasty Ryan traits we tend to want to pretend didn’t exist.

  27. Dinky says:

    ERA- is based on league average. Replacement level is worse than that. Consider: 100% of the pitchers who can pitch at league average (and then some) are already in the majors, with the limited exception of young pitchers being held back for reasons such as keeping an extra season before free agency. When the Dodgers traded Pedro, he was expected to be a reliever for the Dodgers, but he was in the majors. But some players who can hit league average cannot field well enough and are stuck behind a better hitter. I think replacement level is much worse than league average. That said, I think Nolan Ryan is the most valuable. Year after year you can count on him for lots of innings not just above replacement level (which is probably worse than ERA+10) but better than league average.

    If we’re running a real team, then as many people pointed out, you’re better off with Pedro. The extra years could probably buy nearly a league average pitcher. But if you charted games by attendance, you are probably better off with Ryan (Koufax, Fernando) since you’ll fill the stadium when the superstars pitch. I’m not sure Pedro ever was that popular.

    • Rob Smith says:

      Well, hmmm, Pedro had those three Cy Youngs (and two seconds in the voting) and pitched in that non popular town of Boston…. the one that’s never on TV playing the Yankees. I’m sure that Pedro never pitched in front of full houses at Fenway, Yankee Stadium or anywhere else. Ha, what a comment! BTW: as good as Ryan was with the Angels, they couldn’t draw flies in the 70s…. although things improved somewhat when Ryan pitched…. but certainly nowhere near a full house. So much for the attendance angle.

  28. rpmcsweeney says:

    It seems like using historical pitchers is distracting from the essence of the question. Along the lines Tango did in the above comment, this can be simplified more:

    Pitcher A is more dominant, Pitcher B more durable. Your intuitive preference reveals your estimation of the value of a replacement pitcher. If you prefer Pitcher A, then you think replacement level is sufficiently close to league average that a combination of the two trumps a combination of Pitcher B and replacement level. If you think replacement is significantly worse than league average, you prefer Pitcher B.

    There are some complicating factors. In any given season, holding all other things constant, Pitcher A gives you a better chance of success.You might idiosyncratically (but entirely rationally) prefer maximizing your success in the short term over long-term stability.

  29. Asking which player you would rather have for the entirety of their career is pretty much the only question to ask when deciding which player was better. I have Gibson as one of the 10 best pitchers in history, with Martinez just outside. They are so close, it comes down to Gibson’s hitting and post-season. As high as Pedro’s peak was, Gibson was a legitimate ace until age 38, among one of the very best pitchers in the league more often Martinez was. Ryan is somewhere around 20 to 25th best ever; as much as people like to discredit him, Ryan was a legitimate ace for about a full 10 years, and the very best in the league for at least one year (77); just falls short of the Martinez / Gibson standard who were the very best at least three or four times each

  30. pumpkino says:

    I wrote a long post, but in writing it I realized I missed a deeper point. If one assumes that the real career values we have you would 100% know in advance, I think the question becomes one of market availability:cost per marginal WAR (adjusted by team). IF we assume that Pedro and Ryan are equal in career WAR (and if we 100% know all this in advance), and IF we assume tomtango’s salary figures of 15/25 are correct, and IF we assume they’d both be playing for the same team, then they are equal.

    If we remove any of these assumptions, then you have to look at other issues in order to make the question apples to apples. At first I was thinking, well, of course team X would replace Pedro with an “average” pitcher eventually, but, in “plucking” Pedro out of the team’s talent pool, the resource would be replacement level or below – and I’m talking long-term. Otherwise, you are talking additional resources required, which adds money into the equation. And market and what they used to call “Q” factor comes in – Albert Pujols makes additional money because he’s “Albert Pujols”. Same applies to “A-Rod” (before recently) – teams are looking at net marginal revenue, not “cost of WAR” exclusively.

    One has to be really generic here – it’s not does team x have a better-than-replacement pitcher in the minors now, it’s what would be available/required in 15 years. And that, making all the assumptions true, would be a replacement player. So Pedro and Ryan would be equal. I just don’t know if the accumulation of assumptions makes any real conclusion valid though – if not, then you take Pedro in a second.

  31. Since you ask me to take them for their careers, I’d take them based on innings pitched. If I was a Nolan Ryan fan, then I got to be a fan for more games so I was a happy camper.
    Replacement value? None of these pitchers are replaceable, so WAR is sort of irrelevant when we discuss them.
    For the Yankees this year, all of their injured guys (given current abilities) are replaceable. In their case, WAR is a worthwhile measurement. To me, “replaceable” is somewhat analogous to “fungible.” If the player you need to replace is not somewhat close to the norm, the drop off to even an average player will be so large that you aren’t comparing similar things.
    Conclusion: WAR or any discussion of being “replaceable” doesn’t work for elite players.

  32. I would take Ryan by a fair margin, but then Gibson and Martinez are so close it’s really a coin flip, and Halladay just really isn’t in the same zip code.

    I don’t know what the ERA- approximation for a replacement level pitcher would be, but my guess (and you did say it was supposed to be a gut level thing here, so I don’t want to look it up because that seems antithetical to what Tango is looking for) is that it would be around 125. That makes Ryan the favorite by a fair amount for me.

  33. Brady Childs says:

    Gibson, Pedro, Ryan, Halladay

  34. waqas farooq says:

    Lols and Gags, Lol Pictures, Lol Videos, Funny Pictures, Lol is the Laugh out of Laugh where you can Fun Unlimited and Laughing Unlimited. Visit the Best Lol Network Ever, where you can every thing is lol and Funny, Troll Images, Funny Vidoes, Prank Peoples, Funny Peoples, Prank Images, Fail Pictures, Epic Pictures, Epic Videos, Prank Videos, Fail Videos and Much More Fun and Entertainment

  35. Its Awkward to Say some one you are doing Lol…. Every thing you want in Lol and Troll, Now you Get all in one Network, LolsGag, Lol Pictures, Lol Videos, Lol Peoples, Funny Peoples, Troll Images, Awkwards Text, Funny Comics, Jokes, Funny Jokes, Funny Facts, Humour, Funny Planet, troll pictures, funny pictures, Facebook pictures, facebook funny pictures, facebook lol pictures, Funny videos and Much More only Laughing out of Laughing

Leave a Reply

Your email address will not be published. Required fields are marked *