Loading ...
Okay…I’m about to post something which probably isn’t going to be very popular on a number of different levels, but I believe it needs to be said.
I don’t think the Reds will win as many games in 2010 as they did in 2009; that is, they will win fewer than 78 games. How many, I don’t know, but somewhere between 71 and 78.
Why do I feel this way?
Well, due to the Pythagorean winning percentage method (click the link and then scroll down to get to the formula and definition).
In 2009, the Reds won 78 games and lost 84. If you’ll check a source such as baseball-reference.com you will see that the Reds’ expected won-loss record was 76-86, based on the number of runs the Reds scored and allowed. This “surplus” to the win total is usually credited to the manager, in this case, Dusty Baker, as to finding a way to help the Reds scrap to two additional victories. Some attribute the extra wins to blind luck, but whatever the case, the Reds overachieved.
However, in reading the new Baseball Prospectus book, I find the Reds overachieved even more than at first glance. Baseball Prospectus has taken the Pythagorean method to a new level, calling it Pythagenpat, and it’s used their equivalent runs methodology to get a better understanding of team talent than to have the numbers distorted by outside independent influences. Using this method, the Reds’ projected 2009 performance should’ve been 71-91, which means Dusty and the Reds came up with seven wins above expectations. That’s rather exceptional.
The problem lies in the mirage.
For two consecutive years, Dusty’s Reds’ teams have outperformed their Pythagorean expectations (two-three wins each year using the traditional Pythagorean method). That’s probably an even stronger indicator that something may go wrong, or put in proper perspective, a “correction” may take place in the balancing of luck.
Back in 1983, Bill James wrote in his 1983 Baseball Abstract about the “Johnson Effect” of the Pythagorean won-loss principal. He referenced it in his 1989 book “This Time Let’s Not Eat the Bones.” Let me quote James’s book instead of trying to paraphrase:
“The Pythagorean theory of won/loss percentage states that the ratio between a team’s wins and losses will be similar to the ratio of the square of their runs scored and the square of runs allowed…..It can also be shown that teams which deviate from this relationship in a season tend to be drawn toward it in the next season; in other words, if a team wins ninety games with runs scored and runs allowed figures which should lead to only eighty wins, the odds are strong that the team will decline in the following season.”
Reading from James’s 1983 Abstract, James referenced a Toronto newspaper writer named Bryan Johnson who had proposed this theory (thus the “Johnson Effect”) that teams will move more toward the middle year after year, especially from wins seemingly derived from ‘luck.” James conducted a study of National League teams from 1900-1980 which had actual won-loss percentages of 40 points (about six games) better than their projections. Of the 42 teams that met this criteria, 29 of those teams declined and 13 teams improved, or stated another way, 69% of the teams declined the next season.
To check his study, James lowered the threshold to 20 points (about three games difference) which happens far more often than six games of variance. James used American League teams from 1960-1980 and had 50 teams to use for the study. In this study, 15 of the 50 teams improved, and 35 of the 50 teams declined, which falls very closely to the National League study. In this case, 70% of the teams declined.
The conclusion: by historical norms, there’s about a 70% chance the Reds will win fewer than 78 games in 2010. Likewise, that means there’s about a 30% chance the team will win more than 78 games in 2010.
There’s a few things that can offset this projected decline. An acquistion of a true impact player, such as a superstar, without losing much talent from the existing team, can improve a team by several games. Also, an across the board improvement of a young team can make a difference, too. However, before we get on the “we’re a young team” bandwagon, it should be noted that the Reds are not exceptionally young. In 2009, the National League had a league average age of 28.7 when it came to positional players. The Reds’ average positional player age was 27.9, the fifth youngest in the NL. The National League’s average pitching staff was 28.6. The Reds’ pitching staff averaged 29.0, which was the sixth oldest staff in the National League. The age issue seems to be a wash to me, and we haven’t added any real game changers in the offseason; and we already have Joey Votto, who was one of the best hitters in the league in 2009.
As for the farm system, Baseball Prospectus pretty much says there’s not much help for us there at present time. The highest ranking Reds prospect on the BP prospect list is pitcher Mike Leake, who ranks 58th, and is probably a couple of years away. Next is infielder, Todd Frazier, who ranks 67th, and is predicted to replace what will probably be a traded Brandon Phillips, whom BP states is our best overall player. Since that’s all we have listed in BP’s top 100, there’s not anyone in our farm system that will add to our team this year. Also, keep in mind that most any prospect over the age of 24, is quickly losing prospect status. With prime peak baseball age being 26-29, prospects are younger, not older. Young players over 24 can still contribute, but are far less likely to be impact players; more than likely, they will be role players, which every team needs, too, but they aren’t likely to be gamechangers.
The caveat to this suggestion is that we have already graduated three players who are “prospect age” and are contributing on the major league team–Jay Bruce, Johnny Cueto, and Homer Bailey. Bruce and Bailey still have high ceilings; at this point, Bailey seems to be making bigger strides than Bruce, keeping the “Verducci Effect” (Baker’s overuse of Bailey in 2009) in mind. Neither Cueto or Bruce made as many strides in 2009, but both are still young and have time to improve.
As for quality by position…BP uses their “PECOTA” system to rate potential league leaders and top ten players by position. BP predicts Joey Votto to tie for 14th in the majors in home runs with 29; Jay Bruce is predicted to be 9th in Isolated Power (total bases minus singles divided by at bats); and Francisco Cordero is projected to be 8th in saves. As for quality by position (top ten listed in the majors), Joey Votto is ranked the 7th best first baseman in all of baseball and Brandon Phillips is listed as the 7th best second baseman in baseball. The only other mention we have is that potential fifth starter Matt Maloney is projected to have the 5th best value over replacement pitcher for rookies. We essentially don’t have enough star power either.
My hope? I believe if Homer Bailey can pitch anywhere near where he pitched at the end of last year, we can at least overcome at least half of the seven game deficit (78 wins over the expected 71), may be even more. I think Bailey has the biggest chance of making a positive impact on the Reds team this year. If Jay Bruce can increase his line drive rate (decrease the flyballs) and somewhat manage the strike zone, that will most definitely help, too. Bruce’s 2009 line drive rate was 13%, which was the lowest of any player in baseball with more than 300 plate appearances. Bruce has the power, but we need more hits, too. If Bailey and Bruce don’t take huge steps forward, we won’t win 78, and we’ll be lucky to win 75.
Thus, my rationale. I suppose some will say I’m a naysayer; others will say I’m taking the unusual stance (for me) of supporting Dusty; others will say I just downplayed Dusty’s role; others will disdain what I’ve written as unnecessary; others will remind me that there are exceptions. I truly hope it doesn’t work this way, because games are a lot more fun when we win…but, in my opinion, we still need a whole lot more talent than what we currently have on our roster or in our farm system. I realize that most projections I’ve seen for our team are for around 80-82 wins, but the historical record does not support that proposition.
This is good analysis, but I think you’re missing a few things. The difference, for example, in the performance the Reds had at Center, Short, and Third last year and what they can expect this year is, even conservatively, equivalent to adding an impact player. Those three positions were total disasters for well over half of last season. This year, the Reds figure to be at least league average at all three (and this is a rather conservative estimate). I won’t be surprised if the Reds tank again, but I think your underplaying just how much this team has improved since last year.
Also, and this is pretty random, I seem to recall reading something when he was with the Cubs about how Baker’s teams have consistently out-performed their pythagorean record. I don’t know if there’s anything to that, but I thought I would bring it up.
Nice post Steve. I file this under “interesting” but it’s nothing that gets me too worked up. Projecting win totals on projected runs scored is just information to me. It sort of feels like predicting a due date for a new baby… “based on the data and the process we’ve come to understand, your child will be born on such and such a date.” Well, we all know how that goes. A birth date range seems to make more sense. Of course, there are factors that can inluence a premature birth, and those should be understood and avoided, much like factors that lead to scoring few runs (e.g. starting guys with low OBP, or allowing atrocious defense to continue).
Also, and maybe the math accounts for this, at the end of the day you can be outscored 4-12 in a series and still have a winning record. I understand it goes against the odds and the math, but it’s still the only thing that needs to be true in a season.
As for crediting the manager alone for differences in expected versus actual… I’m not sold. I don’t know how you parcel out the manager from the players there. Sure decisions made by the manager play into different outcomes, but sometimes someone like Joey Votto just has a great game and pushes the runs across, despite what the manager did. Seems like a combination of effects.
Several things…
How does our 22-1 steamrolling factor in to this equation? If the theory suggests that our runs scored (squared) and runs allowed (squared) dictate our overall record, can this 21 run disparity affect our overall projected record by more than one game? I’ve always had an issue with compiling multiple games’ worth of stats to estimate single games x 162…I view them as individual games, and this one certainly skews our overall run differential. Remove that game, and how do our numbers project?
Also, the nature of our injury situation last year certainly impacts our overall results. How does playing an injured, 30-some catcher at 1st base affect the differential? With Votto out, and immediately following Hernandez’s trip to the DL, the Reds went 1-15 over their next 16 games. The lack of 1B depth (with Alonso both injured and not yet ready) is a remedy/insurance we can afford in 2010 that we could not in 2009, though it certainly is not factored in to many 2010 projections. Throw in the Rolen acquisition, concussion, and close to the season, and we fielded “several” different teams last year.
Rambling aside, my point is that I disagree. In essence, I’m saying that the RS/RA relationship from 2009 is not the best way to view our performance, which makes it hard to use that as a benchmark for the 2010 season. Obviously, I’m an optimist about this season, and can see 84, 85 wins this year, so feel free to take my evaluation with a grain of salt.
22-1 being the Philly game, BTW.
A: .666, .636, .650
Q: What is the OPS production in 2009 for the Reds at 3B, SS, and CF?
That is the largest contributing factor to why the Reds posted their lowest runs scored total of the decade in 2009. I see a correction in those positions’ output this year without a sacrifice to the offensive production from the other positions in 2010.
@JasonL: I started typing before there were any comments here, but obviously, I agree completely with what you wrote about the team’s improvement at those 3 positions.
@Kevin Mitchell is Batman: Plus another 5 runs Janish gave up at the end of the Milwaukee loss. It was 10-3, finished 15-3.
Steve, you make my head hurt. But you may be right.
For the blowout situation…the more games that are played, the less it matters. Other teams lose like this, too, after all James went back to 1900 looking for examples. Blowouts matter in the record of individual games, but the 162 game season should be a large enough sample size to wash out outliers, especially since the 162 games is the entire sample.
Anyway, there were 11 games when our opponents scored 10 or more runs. We scored 10 or more in 8 games. However, you have to keep in mind the other side of things, too. We tossed 12 shutouts, we were shutout eight times. We allowed and we scored one run 17 times each. We scored two runs 28 times, our opponents scored two runs 16 times. Compiling those totals pretty much negates the 22 and the 15.
Also, the other teams in the study made improvements, too; they weren’t playing in a vacuum and remained steady. The studies show the teams move toward the middle; teams that need to improve do; teams that are winning are more apt to stand pat and lose ground.
As for offense…LF may produce less; Gomes hit really well last year. I also think Phillips will decline some; there’s also a chance Votto will back up some…that was a huge year for him, better than projected; a “correction” wouldn’t be a surprise. Votto also ended strong; there’s a chance his time off may have provided more strength at season’s end.
I don’t think Stubbs will hit like he did in the majors last year. He may struggle to pass .700 OPS himself. It’s hard to say about Rolen…I expected him to hit more than he did and he’s older by a year. Our catchers will hit less (both Hernandez and Hanigan). I expect (hope?) Bruce will take a big step forward. Cabrera will outhit anyone we have at shortstop, but our defense will take a step back from season’s end (Janish) but be better than the beginning (Gonzalez).
Anyway…hoping for small improvements everywhere doesn’t usually make for long term improvements; that’s hoping for career years, which are more likely to happen with young teams (1999 Reds) than veterans who are usually more consistent.
It should also be noted that we were 27-13 going down the stretch, which greatly improved our record. That could be cause for hope; may be it was just a statistical “correction” for the year.
As for Dusty’s “input” to wins…my personal belief is that managers cost teams more wins than they earn. In Chicago, Dusty was “behind” the Pythagorean method his first year, ahead the next three years. All were within four of the projection.
Oh…while the Pythagorean won-loss record may seem odd…it’s usually within 2-3 games of predictions, and as noted above…70% of the teams follow the path…we all hope we’re one of the 30%.
Homer and Jay are the keys, and Votto not losing ground. (we need to trade Cordero for a young shortstop). That is one more thing…teams don’t usually win with older shortstops…I’ll research this again, but winning teams usually have young shortstops.
I don’t know how you ignore “we are a young team.” The Reds are a young team. The “core players” i.e. the impact guys Bruce, Votto, (even Phillips), Volquez, Bailey, Cueto are all very young. The Reds are VERY old at a few key positions – 3B, Loogy, etc. There is an offset. As a statistical measure median age is more important than mean.
BP predictions on those three positions, keeping in mind that BP predicts 82 wins.
Scott Rolen: 2 wins above replacement player
Drew Stubbs: 0.7 win above replacement player
Orlando Cabrera: 1.1 wins above replacement player (computed before signing–small bump for playing in GABP–a couple of tenths of a point?)
Last year’s perfomance while with Reds:
Third base (Rolen, Rosales, Encarnacion): 0.4 win above replacement
Centerfield (Taveras, Stubbs): 3.0 wins above replacement (they rated Taveras and Stubbs really high in defense last year)
Shortstop (Gonzalez, Janish): 0.5 win above replacement
That’s a net -0.1 from last year…Dickerson’s 19 games in CF would probably erase this negative and make it even.
For comparison…when I mean impact player….to the extreme, Albert Pujols was 12.7 wins above replacement level last year.
Joey Votto was 4.8 wins above replacement level last year.
Tim Lincecum was 8.25 wins above replacement level.
Playing time and injuries make a difference.
Age is adjusted for playing time–it is what it is.
I do think it’s an offset as I said…we’re pretty much in the middle, not young, not old.
LF OPS was .721 and RF OPS was .816. I don’t see LF producing less this year.
Phillips was dead-on his Cin career average. I don’t know why we should expect him to decrease?
I suspect Stubbs will be between 700 and 750 with a better obp than he showed in his brief time last year. Something like .350obp/.375slg.
Why would you expect the catchers to hit less? They combined for a .671ops last year.
The Reds played the Pirates (after Pitts traded away their entire team) 11 or 12 times over the final 40 games and only lost 1 game to them.
@Steve Price: We’re using different sources. For example, FanGraphs Fans and Chone have Stubbs at 2.6 and 2.5 WAR, respectively. Rolen at 3.5 and 3.0 and Cabrera at 1.4 and 0.5.
I have said in multiple places that I think Cabrera is, at best, as good as Janish, but, over a whole season, I do think he’ll be better than Janish/Gonzo.
I’m not sure why there are such big differences in the projections of Stubbs and Rolen, but that’s 2.5 wins, minimum, that Fangraphs sees, but BP doesn’t. It will be interesting to see how it plays out.
Also, Zobrist rated the highest of any player on Fangraphs (Pujols was just a tick behind) at 8.6 WAR. Votto rated a 4.5 there. It seems as though BP has a much wider range. I’m not a subscriber, so I have no idea why that is, but it’s pretty interesting.
That’s a pretty pessimistic projection on Stubbs and suggests that they should have retained Taveras?
1 – I think the defensive numbers on Taveras are inflated. He was not a positive on defense last year.
2 – CHONE projection for Stubbs is 2.5 WAR, and that includes a negative 0.5 for offense. Bill James and Marcel both have Stubbs projected as positive 0.2 and 0.3, respectively.
I am not sure the Reds can be considered extremely young team anymore. They will have 3 of their 8 position players (Rolen, Cabrera, Hernandez) 35 or older. Phillips is 29 which is average at least. Gomes is 30. Arroyo 33, Harang 32, Rhoades 40, Cordero 35. On the bench you have Miles at 33. You also have Burton and Masset in the 28/29 range. If Lehr makes thes team he is 33. This is 13 of the 25 roster spots.
BP has Taveras at 0.6 projection (Stubbs at 0.7). They have Taveras rated at 1.0 for last year. They nail him on offense and give him credit for 16 runs saved on defense.
You know..that is possible…the Reds’ defense was greatly improved over the previous year…we all remember Taveras making some really foolish plays out there…but, we may have noticed it more after we were fed up with him. We may not have noticed his positives (not defending…considering human nature here).
BP consideres Stubbs “otherworldly” on defense, so that’s not the issue. I don’t think Stubbs will hit .350 OBP. I’ll be surprised, may be shocked, if he passes .325. I think he needs to hit seventh or eighth in the order…just checked: BP’s PECOTA predicts .323 OBP for Stubbs.
As for catchers and 2b…Hernandez is older and a catcher; he’ll decline. Hanigan just can’t drive the ball…he didn’t do well at season’s end. We can only hope he’ll repeat what he did.
Phillips is 29…he took a few more walks last year; it’s a big year for him, career wise with the big $11 mil coming up for next year…I think he knows he’ll be showcased with Frazier coming up to replace him. I expect him to swing for the fences more…driving down his production. That’s conjecture on my point. BP has his EQA dropping from .289 to .283, not much difference.
The great ting about a thesis like Steve’s is it’s wide open for all sorts of interpretations. And really, who is to say any of them are wrong? Or right? Almost 100 percent of it is subjective and speculative. It’s the kind of study that is a baseball blog’s dream. Awesome stuff.
Insightful post, Steve. Thanks.
I think a very common mistake to make at this stage (every year, for all fans of all teams) is to assume that the guys who played well will do the same this year, and the guys who were bad will do better.
I do strongly think that Jay Bruce, for one, will improve this year. He just seems to be so much better than last year’s numbers.
But Joey Votto exceeded our expectations last year. His minor league numbers were good but not great. I think we need to be open to the fact that Votto might backslide a little bit this year.
Hard to say overall.
I’ll agree that we’re probably better overall at CF and SS, and I hope Homer will improve. I could see us finishing slightly over .500 this year, and that’s the first time I’ve thought that in a LONG time.
I don’t think we’re playoff material yet, but I at least feel good about where we’re headed. I liked this offseason’s moves a LOT better than last year’s.
The Aroldis Chapman signing was genius. As a mid-market team that does most things by-the-book, we HAVE to have a bold stroke here or there, or else we’re going to win 70 to 80 games year after year after year…
Regarding the youth debate, here are two lists:
Bailey
Cueto
Bruce
Votto
Stubbs
Dickerson
Phillips
Hanigan
Hernandez
Rolen
Cabrera
Gomes
Harang
Arroyo
Those are all the players (I don’t think I’ve forgotten anyone) who figure to get major playing time this year. Notice I don’t have and relievers. That’s because bullpens are a giant-freaking-crapshoot and its pretty much pointless to try and guess how they’ll do. The first list is players at or below their “peak” ages. These are players who, all things being equal, should get better. The second list is players at or past their peak ages. All things being equal, these guys should get worse.
It seems to me that some of us are assuming the old guys will decline, but the young guys won’t get better. I don’t know why we are making that assumption. On the first list, only BP and Dickerson are at their peak ages.
The Reds might be totally disappointing and mediocre this year. They also might be pretty good, but don’t discount the youth just because it hasn’t done everything it could yet.
I’m staying optimistic as long as I can. I’d rather be grumpy starting in June than February if it comes to that.
the real point of this post is steve’s follow up where he says he thinks Gomes won’t hit as well, Stubbs won’t hit as well, and BP won’t hit as well. That’s his gut feeling, and he’s looking for backup.
The whole pythagorian stuff isn’t really meant to be used to predict an individual team in a given year. it’s 30,000 ft view approach. There are too many moving parts for a statistical approach based only on runs to be valid.
There are many other better projections systems out there, based on a lot more than runs from the previous year. They have the reds around .500, and I’ll take those projections over this, which is really a dressed up gut feeling.
Something I keep forgetting (I think we all do) – Votto is almost certainly not going to be quite as good, rate stat wise, as he was last year. However, he’s also probably going to play more games. So, I guess, remember that 95% Votto is totally going to be at least as good as 80% Votto/20% Hernandez was last year. Maybe even a tick or two better.
I think this is a VERY, VERY important point
EE was hurt…then didn’t play well and we acquired Rolen late in the season.
Gonzo had an AWFUL season until he went to Boston (something in the water?)
and our manager couldn’t help himself and started the worst CF in baseball all the time.
But I also want to add there is one other thing I think we can hope for
Experience!
Fisher was a rookie. I expect he’ll improve this year
Herrera was only 24 with 7 games of major league experience. I expect he’ll improve.
Masset was only 26
Bailey is STILL only 23!! But now has some experience.
Cueto was 23
Hanigan was essentially a rookie
Votto was only 25 and in his 1st full season
Bruce was 22(!!!)
Stubbs was a 24 year old rookie
all these guys are a year older and have some major league experience. It’s no sure thing but there is a chance each one of these guys could improve.
The Reds lost Votto for a good chunk of time because of what he was dealing with and lost Bruce to injury. Lets hope they can both stay healthy this year.
I would agree, Jason…even lower rate stats may mean increased overall production from first base
I would adjust the list a bit…players don’t necessarily improve during peak seasons…they should be reaching a point of established perfomrance, so consistency would be the expectation…using young, peak, past peak:
young:
Bailey
Cueto
Bruce
Votto
Stubbs
peak:
Dickerson
Phillips
Hanigan
Gomes
past peak:
Hernandez
Rolen
Cabrera
Harang
Arroyo
We’re back to balanced again…
Oh…either Baseball Prospectus changed something or my memory failed me, but their website is predicting 77 wins for the Reds not 80-82 as I previously mentioned.
I agree this is a summary expectation, not a thorough statistical model. However, 70% probability based on history is a rather significant threshold…and, for all the statistical models out there, it pretty much comes down to runs scored and runs allowed….how those runs are obtained is for the individual charts
Yes…precisely…discussion time
The new trend is defense, and, to Jocketty’s credit, he’s been on top of this. We all need to think about this in the approach to the team. That’s actually been my biggest surprise about the Cabrera signing. I also have little doubt that’s why we dragged our feet on Gomes.
The team to watch is the Boston Red Sox. I almost feel cheated that Bill James now works for them. For over 15 years he has been working on some defensive metrics and he doesn’t share his biggest discoveries anymore because he is under Bosox employment. It’s his choice and perogative…but, I know something is amiss and we’re not seeing it.
I’ve read where Baseball Prospectus and some other statistical organizations sell their most cutting edge info to some of the major league teams…and defense is the current hot topic. OBP matters on offense, but defense is being added to the mix. Just a few years ago, SLP mattered more than OBP, but with offense now in decline, OBP is going to matter more, so defense will matter on the flipside.
I predict the speed game will be back into vogue in the next three years or so.
But,the Red Sox are dumping power hitters and going the defensive route. I heard the other day that they are looking to move Ortiz now.
The last few years the eyes have been on the A’s and their plan. Now, it’s the Red Sox.
Unless this Pythagrean fella can play the oufield well and be a 30/30 man, I don’t want him.
It’s interesting that anytime you introduce new analysis, you introduce more contentious variables.
Also, and this speaks to my skepticism of all stats defensive: any system, any one at all, that states our CF defense was better in 2009 than 2008 is flat wrong. Willie Taveras is not a + defender, period. He uses his speed to help cover some tremendous gaffes. He is not that good. Corey Patterson, God bless him, for all his faults was and is a much, much better all around defender. This I knew before the Reds signed the T-virus, and I don’t care what Chone, Marcel, Pythagrean, or Marty would have you believe.
There is only so far you can extrapolate data and it really be useful.
Oh, and one more thing: I hated geometry.
Pythagrean, meh.
hmm…well…this is a tough one. I think our defense in CF was better in 2009 than 2008.
The reason?
59 games in 2009 with Dickerson and Stubbs in CF is MUCH, MUCH better than 116 games in CF from Patterson, Hairston and Freel
Patterson had an off year with the Reds defensively
when looking at team defense in CF is much more than a comparison of Taveras and Patterson. Plus, while Taveras often took horrible routes to balls a good bit of time he made up for it with his speed. I don’t think he was as bad defensively as some made him out to be. Dickerson and Stubbs were AMAZING on defense in CF
I don’t see how this makes us “balanced” actually. You still have a majority of the roster expected to get better or tread water with a few guys on the downside. Also, I’m just going to be honest and say that I’m not worried about Harang and Arroyo. I don’t see either of them being “worse” than last year. Also, you dropped Bailey from the young list, which, I assume, was accidental.
Steve,
Interesting article, but when I was reading it (at least the first half) I realized that you buy into Dusty Baker’s “Law of averages”. The fact that we won more games than we “should have” statistically the last two years has no effect on us doing the same this year. Just like if we flipped a quarter 100 times and it was heads 75 times, there would still be a 50% chance that its heads again on the 101st flip. Nothing makes me scratch my head more than when Dusty talks about a player being “due” because of the law of averages when they are in a slump.
accidental…missed him on copy and paste
I do think there’s a law of averages; however, there’s also talent and ability, and most things drive toward the middle. Taveras wasn’t doing the things he had to do to improve at the plate last year; managing the account, driving the ball, or giving effort; the thought that he would magically start hitting was crazy. Bruce should get better, but last year he hit 13% line drives…he was going for the big fly, or badly fooled (or both) all the time last year. It was interesting to me when he started hitting at year’s end because I kept hearing that he had not worked during his injury and was in Dusty’s doghouse…allegedly, that’s why he wasn’t starting when he first came off the DL. Then he started pounding the ball.
By definition, rate stats are averages…however, that’s just part of the overall picture. On another thread, I’m suggesting Cueto needs to go the bullpen, may be our future closer, because the law of averages is catching up with him the other way in the second half of the seasons, and scouts say his physical stature will lead to breakdowns as a starter.
As for the last couple of years determining how we’ll do this year…well, it really doesn’t; each year does stand on it’s own ..however, we have 80 years of data that says there’s a 70% chance we won’t be lucky this year, and our record was based on team performance, not blind chance—the players and management team are the common denominators, so it’s not just averages or chance…I’m not an odds man…I don’t bet the ponies and I go to Vegas with a vacation “budget” (fulling expecting to lose).
this stuff makes me dizzy.
Opening Day, get here quick.
Maybe it’s not “luck”, but that the analysis and prediction was wrong.
See, I guess this is kind of my point. The “young” Reds are the core players. The core of this team is extremely young and will get even younger over the next few years as Francisco, Harang, Arroyo, Rolen and Cabrera are replaced by guys like Leake, Chapman, Wood, Francisco, Frazier, and Alonso. It’s not as though the Rolen, Phillips, Cabrera contingent is what will make this team a winner. Those are nice vets, but if the Reds are going places, it will be on the backs of Bruce, Votto and the young pitchers.
I don’t disagree with any of the future…this post was only about this year. As for the age of the team, again, I was only discussing 2010 and anticipated playing times.
Now, whether all those young guys will contribute much in the future is a whole different story and discussion….
Pythagorus can be thrown off by a few things. One is that if a team wins most of its close games, due to a good bullpen (or loses most due to a bad one).
Within a season, after a decent sample of games, the Pythagarus record has been shown to be a better predictor of the team’s end of the season record than their actual record. This doesn’t mean that it’s almost always close to correct, just that it’s a better predictor than assuming a team will maintain its current winning percentage.
Using Pythagarus across seasons is tricky, because a team can get better or worse across seasons. (It can of course get better or worse during a season too, but on average not as much.) If a team is lucky one year, then the odds are it will not be as lucky the next year, so, unless it has improved, its record will decline. This does not mean that good luck this year will mean bad luck next year – that would imply “luck” has memory, which is mathematically unsound.
PS None of my above post contradicts anything that anyone in particular has said.
Not to mention that we haven’t considered one possible explanation for the Reds recent luck: Dusty is a genius !
Or not.
In 2007, when Cueto was flying through the system and the Reds were having bullpen meltdowns, Chris Welsh thought Cueto should be called up and moved to closer. Krivsky’s staff and a lot of scouts scoffed but once I saw Cueto’s stuff I sided with Welsh.
Now, with some starting pitching depth on the way, I really, really like this idea. The problem is that Cueto’s maturity and pressure-situation makeup just isn’t where it needs to be quite yet. But I wholly agree with your premise on his future role.
Love the analysis, but the way I view this analysis is that they are a lot better at telling us what happened last year than predicting future years, especially when looking at how much a team like the Reds has changed from last year. Look at it this way.
1) Different players for the bulk of the year at SS, 3b, CF
2) More time in LF for Gomes, at 1b for Votto, and at SP for Homer
3) More time in RF and expected improvement for Bruce
These factors alone indicate a lot of flux from 2009 to 2010, which really invalidates using 2009 to predict future results, good or bad. It is very interesting, tho.
Ultimately, that’s why they play the games!
Your first problem is that you’re using historical stats almost entirely from pre-free agency. you’re looking at teams that were primarily made up of the same players from year to year. In the modern game it’s not rare to have 30%+ turnover from year to year.
You can seriously tell me that if a team changes it’s starting SS, 3B, and CF that you think that using runs from the previous year is a better projection tool than the projected stats of the players actually on the team now.
I’m not saying i disagree with using RS v. RA to predict the reds’ record, I’m saying that using last years RS v. RA to predict this year’s doesn’t make any sense.
Steve, you are a valuable asset to this blog. This is really great stuff (though it still makes my head hurt).
Wow Steve, you sure opened a can of worms! I think all the sides have valid points and are, to a great degree correct. 150+ games out of Votto have got to help at 1B. Same for Bruce. 130+ games out of Rolen. No Gonzo or PJ batting second and unable to get on base helps (and I like PJ for his amazing defense). I expect Arroyo to be similar overall to last year. Harang had bounced back prior to his appendectomy and sounds as if he and Price agree about his mechanics, so he should be way better than last year. Bailey should be better as well. Not as awesome as at the end of the year, but still much improved overall. The bullpen should be as good or nearly as good as last year. BP should be about the same as his average over the last several years. That leaves the question of CF, LF and the other two SP positions. I still like CD playing CF and leading off. Stubbs and whoever else should be in LF should be at least league average. I fully expect a much better year in the 82-86 win range.
@Steve Price:
Drew Stubbs: 0.7 win above replacement player
Last year’s perfomance while with Reds:
Centerfield (Taveras, Stubbs): 3.0 wins above replacement (they rated Taveras and Stubbs really high in defense last year)
This is just stupid, frankly. Defense doesn’t fluctuate like that. Taveras’ offense was clearly MILES below replacement. His defense didn’t stand out on any metric. Thus, Stubbs’ defense, in 2 months, was somehow worth 3 wins? Yet this year, he’s going to play worse defense AND hit significantly worse than Wily did? Because that’s what has to happen for those numbers to work.
Impossible. I was looking for nothing BUT good defense from the guy. Even adequate defense. Hell, he only had 277 chances – how could he have saved 16 runs? The ONLY thing I can say is this: BR shows him has having better throwing results than the average, and he somehow had 7 assists.
Would somebody please give these projection geeks a hemroid depository?
suppository is what i meant
The holy grail for defensive evaluation IS coming, sooner than I realized.
Think I’ll make that a post unto itself.
@Chris Garber:
agreed, those #s don’t seem right for Stubbs
CHONE has Stubbs as 2.5 WAR. That comes from huge + value from defense and a negative for offense.
While 2.5 WAR is better than .7 I still think that’s low
the Values (I don’t know how they get from the values to WAR) seem low.
Stubbs started 40 games and had a defensive value of 7.6 but the CHONE prediction for his defensive value is 10.6
IF Stubbs is the starter I think you could easily triple that #
We watched Stubbs play defense and while he made some obvious rookie mistakes his defense was great.
Add to that if his main/only sub is Dickerson the total production from CF will be even higher!
I think a big question is, will Stubbs show the power he showed in the bigs last year. He didn’t hit for power in the minors.
and if Dickerson can translate his great CF defense to LF, the OF of Dickerson/Stubbs/Bruce would be the best defensive OF in baseball
too bad if both Dickerson and Stubbs start our SS will still bat 2nd.
[...] Century Defensive Metrics by Chris Garber on February 24th, 2010 in Reds – General This discussion about the difficulty of measuring defense reminded me of an exciting article I read a few days ago [...]
See, that was my first answer, but I figured it was too simple. Like talking yourself out of an answer on a test back in school because you couldn’t believe the teacher would make it easy.:lol:
Indeed.
According to sabermetric theory, winning close games is more a matter of luck than skill (bullpen or no bullpen). That’s part of the Pythagoram thought…teams that won close games this year may not win them next year.
Studies show that turnover really hasn’t changed that much…only the dollars involved. Another interesting thing…players reach the majors later today (more training) than in the old days, which is another myth.
Also..the most effective players (peak seasons) come when players are in the team’s control, ages 26-29, during which they usually don’t have free agency, since players usually get free agent rights around age 30.
Pythagarus is a basis for prediction. Sometimes the predictions are off. I read an article that looked at cases where the Pythagoran record differs significantly from the actual record over a full season. (You don’t have to subscribe to the theory that only luck is involved in such cases.) They found a correlation in those cases with quality of bullpen, but I don’t think it was a formal study of statistical correlation.
Very late to this party, but just wanted to add to the chorus that this was an outstanding post and an outstanding series of follow-up comments. Just a wonderful example of the quality of people commenting here.
Pinson, if you dig up a link to that I would be very interested. The idea of studying common threads in cases when the Pythagorean record predictions are off is an interesting one. Living in Cleveland, one thing that kept coming up when the Indians kept underperforming their Pythagorean expectations was whether management was at fault. As critical as I have been on Dusty, I really hesitate to think the reverse could be true but I suppose it could be considered…
Man this is terrible analysis.
And attributing variances in a team’s pythag to a manager? Did you make that up? Bullpen! Every run is not created equally.
At least use Base Runs and generate some second-order run expectancy numbers rather than just some simple square root calcs.
@GRF: Bullpen, bullpen, bullpen.
It has nothing to do with managers.
And jeebus, 25 year old studies as evidence. Heck, 25 years ago Bill Gates said that 64kB</b) was enough memory for a computer.
Woops
I can’t find the article, but Wikipedia has a nice writeup of “Pythagorean expectation” (http://en.wikipedia.org/wiki/Pythagorean_expectation).
With respect to the Pythagorean formula giving an expected winning percentage, it states: “There are some systematic statistical deviations between actual winning percentage and expected winning percentage, which include bullpen quality and luck.”
The writeup talks about how the most accurate exponent for prediction isn’t exactly 2, and takes a standard deviation (of runs scored by all teams divided by the average number of runs scored) into account.
It also gives a nice explanation of some terms used by sites like Baseball Prospectus, and by Tom above. First-order wins are the number of expected wins generated by the pythag formula. To further filter out the distortions of luck, sabermetricians can also calculate a team’s expected runs scored
and introduce that into the formula. This generates second-order wins: the number of wins a team deserves based on the number of runs they should have scored and allowed given their component offensive and defensive statistics. Third-order wins are second-order wins that have been adjusted for strength of schedule.
I believe I accomplished my mission…I wasn’t writing to the sabermetric community; and, 25 years old or not, the basic theory holds true.
Also…I don’t write for Baseball Prospectus, but the info I’ve seen implies they are using third order wins and neutralized statistics…that’s why the info was significant.
As for manager’s contributing to wins…did I say that was the case? I said some folks say that, and there’s major publications that keep track of this as for management input. As for the bullpen, I do think the bullpen matters…I also firmly believe that bullpen performance is somewhat random so that plays into the luck factor naturally.
Oh…the usual number used for computing the Pythagorean record is now 1.81 not to square them; but, for conversational purposes, the square works.