Yan Gomes expected to be activated Friday by Indians
September 14, 2016Low-cut Cowboy: Bode Plots
September 15, 2016The 2016 Major League Baseball season is down to its final two-and-a-half weeks. With that in mind, it’s time to talk year-end awards. And for the fifth straight year (i.e. every full single-season he’s had in the bigs), that means we have to wrestle with what to do with Mike Trout (h/t Sports Illustrated’s Tom Verducci).
The Los Angeles Angels of Anaheim franchise has fallen on hard times. From 2002-15, the team averaged 90 regular season wins and made the playoffs seven times. Yes, their World Series win was way back in 2002, and they were subsequently swept in the ALDS in three of those next six appearances. But the Angels were a consistently good team with great talent for a long time.
In 2016, that certainly is not the case. The Angels are 63-82, sitting 23 games back in the AL West and even 16 games behind the second wild card spot. Their starting rotation and bullpen are both among baseball’s worst. They have a collection of below-average performers all over the lineup. This hasn’t been a good team at all.
Yet the even worse prognosis comes into the future. Very little of their existing MLB talent is under 25. And prior to the season, every reputable outlet ranked the Angels as having the league’s worst farm system. The analyses at Baseball America, Baseball Prospectus, MiLB.com and Minor League Ball are all brutally ruthless:
“There are times when you just kind of run out of ways to describe something, and the state of the Angels’ system defies much in the way of analytic vigor these days.” “The best thing to say about the Angels system, at this point, is that these rankings stories are over.” “UGHHHH: The Angels get a special category of badness all their own. … There’s no way to spin it: this may be the worst system in recent memory.”
Baseball America’s J.J. Cooper also wrote an extended piece on the Angels’ consistent win-now approach, trading away younger prospects for older players year after year. It’s ugly over there in Anaheim. And the one consistent bright spot happens to be the greatest young player in our lifetime.
Mike Trout is batting .317/.434/.558 in 142 games this season. He has 30 doubles, 28 home runs, 89 RBI and 25 steals. For his career, Mike Trout is batting .306/.403/.559. Per 142 games in his career, he has 31 doubles, 30 home runs, 87 RBI and 25 steals. He’s been remarkably consistent and excellent throughout his five-plus seasons in the majors. His on-base percentage has notably improved in 2016.
Pure offensive stats alone don’t tell the full story, even though Trout’s .962 OPS at this stage in his career places him among legendary peers. Guys like Ted Williams, Babe Ruth, Jimmie Foxx, Lou Gehrig and more. You can easily get trapped in a Baseball-Reference rabbit hole on the topic of Trout’s greatness alone.
But let’s take Albert Pujols for example. In his first five seasons, Trout’s current teammate hit .332/.416/.621 in 790 games. The Machine was one year older than Trout, relative to their MLB experience. But Pujols had three main strikes against him, in comparison: 1) He played first base, a less valuable defensive position; 2) He provided relatively little value with his base-running; and 3) He played in a more offensive-prone era.
According to FanGraphs, Pujols had a 166 Weight Runs Created Plus (wRC+) in 790 games from 2001-05. That statistic, a frequently cited offensive metrics in the sabermetrics community, helps account for “park effects and the current run environment.” But despite Pujols’ much larger slugging percentage, Trout actually has a 167 wRC+ in his career so far.
During those first five seasons, FanGraphs states that Pujols provided -20.5 Runs Above Average from baserunning, defense and fielding position adjustments. Trout, on the other hand, has provided 55.5 such Runs Above Average. Considering the conventional wisdom of 10 Runs per Win that gives Trout a WAR advantage of about 7.5, or about 1.5 per season. That’s a fairly substantial difference given the relatively equal weighted offensive production.
There are simply very, very few modern peers who Trout can be fairly compared to in terms of his all-around production. Any position player of the ’90s has to be compared to the overall slugging numbers of that era. Young prodigies like Alex Rodriguez, Vladimir Guerrero, and Pujols are a clear step below. Ken Griffey might be the best such comparison, but even he pales in comparison to Trout’s league-adjusted value at this stage in their careers.1
Mike Trout leads all of baseball with 9.2 rWAR (via Baseball-Reference) and 8.3 fWAR (via FanGraphs) this season. Nobody else is particularly close. And yet, because of the archaic ways in which Most Valuable Player voting works, Trout likely has no shot at the award. He only has one such honor to his name, back in 2014. He was the runner-up in 2012, 2013 and 2015. He’ll likely finish third or worse this season.
We should be talking about Mike Trout more because he is an anomaly in professional sports. No player this young has ever dominated this thoroughly in the modern times since the very start of his career. I understand the complexities of national baseball fandom, but it just seems this goes often underappreciated. Should we just start up a new award for Trout to win each year, instead of MVP?
Here is an assortment of Ohio sports-related articles I’ve enjoyed recently:
- On Urban Meyer, mental health and his journey back to coaching [Brandon Sneed/Bleacher Report]
- On Johnny Manziel and experiencing all of the bars that he likes to frequent [Michael J. Mooney/Bleacher Report]
- On Kyrie Irving and how we attempt to rate him in today’s NBA [Jason Concepcion/The Ringer]
- On Andrew Miller and the Cleveland Indians’ usage of elite relievers [Ben Lindbergh/The Ringer]
- On the Cleveland Browns and using business analytics in the sports world [Eben Novy-Williams/Bloomberg]
- On Tom Herman and 11 more things to know about Houston’s extraordinary rise in the college football landscape [Pete Thamel/Campus Rush]
- Ed. note: Which is crazy to think about given how transcendent Griffey was. [↩]
130 Comments
they count the same, but the leverage is different. When you’re in a pennant race, the at bats hold a higher level of importance because of the situation. So, while the wins themselves certainly count the same, they tenor of the games are quite different.
So when wins count the same, a 10% increase in odds of winning a game in April equals just as much as a 10% increase in odds of winning a game in September. So your win probability added should be exactly the same.
If you have a measure of game tenor, I’ll be glad to check it out.
An at bat at the end of the season has a greater effect on the potential outcomes of the rest of the year. There is less time to recover from an error, so those games are higher leverage.
The tenor of a September game in a pennant chase should be easy to explain. Maybe you’ve never been to a meaningful ballgame in September? Go look at the attendance or prices for tickets in May as opposed to September.
While I understand that every game is 1/162, there is more pressure at the end. That makes it a higher leverage situation.
What a convincing article… You cited three stats that you probably can’t even begin to explain. Sabermetrics are a bunch of crap I don’t understand how people can so strongly believe in statistics that they don’t even understand. For all you WAR lovers out there please explain to me the difference between Ken Griffey Jr.’s 1997 and 1998 seasons and why his WAR is so drastically different when his stats were essentially identical across the board and he was an all-star, gold glove winner, and silver slugger winner in both seasons.
Hey Matt I promise I will believe your argument if you can do one thing: without looking at Google go ahead and tell all of us how WAR, WPA, wRC+ are calculated, what’s included, what isn’t included, and the strengths and flaws of each. How about adjusted batting wins? Adjusted batting runs?
After that maybe explain to me how Mike Trout 2016 season was better than Ken Griffey’s 1997 and 1998 seasons by comparing their WAR. Griffey clearly was the better offensive player, and also actually won Gold Gloves, unlike Trout.
Unfair the best player of a generation is stuck on a cruddy team? How about its unfair that one of the three greatest hitters in history lost like 4 MVP awards to Bonds when he was juiced out of his mind.
And did you just suggest a stat called pennant probability added? Please tell me you are tripping balls right now.
So RBIs are a useless stat because it has to do with the players around you but you are saying you can’t vote for Trout because his team sucks? Please, explain.
There’s no difference. Wins aren’t weighted.
So what you are trying to do is measure how good someone is under pressure. How do you measure that? In any given situation a guy who goes 0/10 but hits 10 line drives right at a defender would have the same “clutch” factor as someone who went 0/10 with 10 SO. You can’t accurately measure the “nerves” of a player without being inside his head and if you are going to try to measure “nerves” you also have to measure how “lucky” the player is in clutch situations, and where does just absolute randomness play in how someone performs?…and it goes on and on and on making up stupid measurements for things that no one can actually measure.
I disagree with Jacob on this one, but he absolutely knows his stuff.
Yes. I agree.
The difference is not that wins count any more, but that once the season becomes shorter, each win becomes a bigger percentage of the result. That’s why those wins in September matter more.
Plus, it’s a perjorative to be called “Mr April.” Its historic to be “Mr October.”
Bra-vo!
First of all, in baseball a single player’s contribution really has nothing to do with the overall team success so the idea that people obsessed with sabermetrics they don’t understand are voting for a baseball player on the team with the best record is completely contradictory. Second of all, the magnitude of the effect games in September have is relevant to how the team performed in April. So if the player sucked it in April then played above average in September it pretty much evens out and also means that the player is inconsistent.
Google is paying 97$ per hour! Work for few hours and have longer with friends & family! !mj336d:
On tuesday I got a great new Land Rover Range Rover from having earned $8752 this last four weeks.. Its the most-financialy rewarding I’ve had.. It sounds unbelievable but you wont forgive yourself if you don’t check it
!mj336d:
➽➽
➽➽;➽➽ http://GoogleFinancialJobsCash336DigitalTechGetPay$97Hour… ★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★✫★★::::::!mj336d:….,…..
Maybe you’re not an athlete (or haven’t been a successful one) if you think the beginning of a game/season is the same as the end? Just curious.
Let me help:
It’s like churning butter; the first few stirs are as important as the last few, but the last few are a lot harder.
Is this a trick question? We want to isolate Trout’s value. RBIs dont do that. Measuring the talent around him doesnt either.
Im guessing that Pujols is your guy? Great hitter, not quite Bonds or the top three all time.
We want to measure how much a player contributes to winning a pennant, and not just number of wins, yes? I have no idea how feasible it is to translate WPA to some kind of PPA, but that what we want to measure.
No, I understand what you are saying, but its not an appropriate critique of WPA.
Its a measure of how many wins one has added. It doesnt measure “tenor”, a purposefully nebulous term from the so-called-scientist. It doesn’t measure the leverage of each game.
As I have said, a pennant probability added measure could be useful. As the expert on “tenor”, I’ll let you figure that one out.
So far, you havent demonstrated a working knowledge of WPA. You have demonstrated that you arent going to bother to acquire one before attempting to criticize it though. I’m all ears if you have something that shows (preferably scientifically, considering your profession) why its fatally flawed, but so far this is peanut gallery stuff.
Since you enjoy sabermetrics so much perhaps you can come up with some sort of adjustment for the fact that Bonds elevated his career (and WAR) from being hall of fame worthy to greatest of all-time worthy by taking steroids. Additionally make an adjustment for the fact that Ruth, Gehrig, Williams, Musial played in a much less competitive era due to the fact that baseball was not an international game at that point and blacks (and to an extent, hispanics) were not allowed to play. They also played against like 7 teams meaning they saw many more at-bats against crap pitching.
When you make those adjustments then yes, Pujols is (will be) easily on the Mount Rushmore of greatest hitters ever. Pujols just joined Bonds and Aaron as the only players have to have 600 2B and 575 HR (without roids it would only be Aaron and Pujols). Forgive me for using such basic statistics…Additionally without Bonds taking roids Pujols is at least a 5-time MVP, and surpasses Bonds as having the most MVP awards ever.
Actually, is it not more important that a player is able to elevate his performance when potential runs are on base? You know the whole “clutch” factor thing. Because runs win games and wins get you to the playoffs.
But sure, I understand that you want to isolate a players individual value and RBIs are generated because that players teammates got on base. What I am saying is that if you are so concerned with focusing on an individuals contribution and hence value, then why do you care what the team’s record is because the teams record has even more to do with the player’s teammates than do RBIs….
I see the issue. I’m not the one saying to not vote for Trout because his team stinks.
There is no real evidence that elevating ones performance with runners on is a repeatable skill. Thats not to say its not valuable, but we also want to isolate a players true talent from good breaks.
I have a real hard time trying to devalue players performance based on what we think should be morally correct. Its impossible to know what Bonds plays like without PEDs, or even to know that Pujols wasn’t using. Both guys failed the exact same number of PED tests. Which is also the same number as Aaron, who has admitted to using greenies, and Mays, with his red juice, and Mantle, who loaded up on meth from a Dr. Feelgood.
Pujols has had a great career, even if he retired today, 30th all time in WAR. But its impossible to make adjustments based on unknowns.
Steve, you’re saying its impossible to know what Bonds plays like without steroids and so you won’t make an adjustment based on unknowns but you are trying to make a single, all-encompassing statistic for pennant probability added of a single player? Give me a break. And don’t give me the “I never failed a test” crap, Lance. And Adderall might give you a little boost and keep you focused and relaxed but you didn’t see Miguel Tejada hitting 73 HR at age 36 (assumed age 36) now did we?
Morality has nothing to do with it. We aren’t discussing whether or not steroids SHOULD be legal. If I was in charge I would love to see players hit 70 HR every season but the fact is they are currently banned because of that very reason (morality be damned).
Look at Pujols’ stats right now. Now imagine he could add 5 more years of stats that were even better than he was in his prime. You are looking at a 7-time MVP, all-time HR leader, all-time 2B, all-time OPS & OPS+, and top-5 WAR, top-5 hits. Then again that’s all hypothetical but so is WAR, WPA, and PPA because there are so many UNKNOWNS that you aren’t accounting for.
Steve, again it is impossible to put a single event into a vacuum and say it holds X value. The value of each event is in relation to the events surrounding that event and to the events surrounding THAT event. If this is how WAR and WPA are calculated then you aren’t isolating a single players value in a way that you can reasonably compare it to players on other teams nor can you compare it across history.
It is completely arbitrary because it is not all-inclusive (just like the MLB wasn’t “all-inclusive” until the 1950s), and probably never will be.
There’s a lot of things you’re misreading or just simply missing here, and at this point, I’m not sure if you’re just not getting it or being willfully obtuse.
I am not trying to make a pennant probability added stat. I am wondering, if something like that were to exist in addition to win probability added, would that be a better measure for what Bode wants. And all of that has nothing to do with rewriting Bonds historical record. I have no idea how you are drawing comparisons.
Not having failed a test is incredibly important. Without it, you have no proof of anything, and would just be guessing out of you-know-where. And if guys were heavily using PEDs in the testing era, then it would only show that the testing is insufficient, and a guy like Pujols can’t be assumed cleaner than anyone else. If you have evidence that adderall is less effective than another PED, I’m all ears. Let’s hear the proof.
PEDs are not banned because they make guys hit more home runs. The drugs banned in the JDA are specifically chosen as Schedule I, II, and III drugs, as written by the DEA. MLB didn’t decide what was legal or not, they just followed the DEA’s recommendation. The DEA doesn’t give a darn about hitting home runs, and they created the schedules based on accepted medical use and potential for abuse of the drug. It is entirely based on morality.
I can imagine Pujols being even better. But he didn’t do that. So none of those hypotheticals matter, like you know. WAR and WPA are not hypothetical in the same way at all. I have explained elsewhere in this thread how they are created, but you seem to not have read that yet.
Frankly, I have no idea what your point is anymore, other than you want to needle at things that you clearly aren’t fully understanding before you start poking.
You have made some ridiculous comparisons so far, but that last one really takes the cake. I have no idea what you even mean by WAR and WPA not being all-inclusive.
You’re clearly admitting here that you haven’t bothered to look up how WAR and WPA are designed and used. If you want to go ahead and do that, and have a real discussion, great, I’ll wait. If you want to continue to remain uneducated about something you want to criticize, we’re done here.
Ok, Steve here you go:
Why WAR is stupid
1.) There is no single calculation for WAR, many people calculate it differently, which means the statistic is still being developed and is incomplete. In 5 or 10 years you will probably be arguing that some brand new made up statistic is THE most important statistic.
2.) I disagree with how wOBA is calculated because: HBP is not value generated by the hitter but rather the pitcher, and IBB may very well be a result of the hitters talent rather than just being an arbitrary managerial decision; it gives specific weights and value to batting events but neglects to consider the fact that for the most part, hitters don’t choose where to hit the ball and therefore randomness, luck, and misfortune play a huge role in whether or not a player winds up with a 1B, 2B, or 3B. Rather than focusing on the result of the bat off the ball what should really be focused on is what type of hit resulted (aka a blooper, fly ball, line-drive) because this tells us more about how well the batter hit the ball while accounting for the bad luck or good luck a player might have by hitting a line-drive right at a defender or hitting a dinker that luckily split defenders, or hitting a ball awkwardly off the outfield wall and results in a 3B simply out of luck.
3.) The fielding component is completely flawed starting with the fact that the positions are compared against each other rather than comparing similar positions against each other. Multiple hitters can generate the same result because on offense everyone is the same position: hitter. On defense it is impossible for a team to field 9 catchers or 4 centerfielders so comparing the fielding value of a second baseman and catcher is impossible without using relative values for those positions, its apples and oranges.
4.) More on fielding: WAR does not account for the fact that fielding shifts have become a thing, which rather than being a result of a player’s individual fielding talent or “range factor” is just a managerial decision–when a second baseman is shifted to play shallow right field, is he even a second baseman any more? So how does one who is calculating WAR account for managerial decisions or game plans that are not obvious and differentiate that from just bad fielding?
5.) There is no current adjustment for the time period a player played in. If the structure of the league changes, and different people are allowed to play in different periods then the talent pool is distorted. There are many more humans playing baseball in the world today than in 1920 and the league is almost twice as big and what this means is that the average player today is much better than the average player in the early 1900s. If the average player of the early 1900s is worse, then the “replacement level” player is worse and this results in the WAR of the best players of that era being distorted because there is a bigger gap between the best and the average in 2016 than in 1920.
6.) If you can come up with a specific value that a batting event generates in any situation, I think you can probably come up with a specific adjustment for the added value of steroids.
7.) So, if we are comparing Pujols and Trout and I say “there is no way Trout, through age 25, generated more value than Pujols even though Trout has a higher WAR,” most WARshippers (trademark pending) will reply that “WAR includes fielding and base running too, so Trout is clearly the better overall player.” So here is something I threw together:
Through their first 5 seasons, if you multiply avg. PA*(HR%+LD%+BB%-SO%) you are left with a value that assesses how well a player can hit the ball and how good their batting eye is. The averages leave you with Pujols=203.056 and Trout=162.624.
On to base running: if you add (SB+BT)-(CS+PO+OOB) and take the average over the number of years you get a value that compares positive base running events against negative base running events and if you compare Pujols against Trout you get Pujols: 14.6 and Trout: 33.6
Fielding: If you compare Trout and Pujols’ Total Zone Total Fielding Runs Above Average, BIS Defensive Runs Saved Above Average, and BIS Plus/Minus Fielding Runs Above Average, if you add up their totals then divide by career years you get Pujols: 22.875 and Trout: 13.667.
Add up hitting, base running, and defense and you get Pujols= 240.531 vs. Trout= 209.891. Well would you take a look at that…
“5 or 10 years you will probably be arguing that some brand new made up statistic is THE most important statistic.”
Yes. And that you think this is a problem is laughable. We are constantly evaluating and improving our measures of analysis. This is a good thing. WAR is our best measure now, and that smart people are trying to find ways to improve it is a lot better than how things used to be where people just said “average, home runs, and RBI are all you need”.
“Rather than focusing on the result of the bat off the ball what should really be focused on is what type of hit resulted”
We already have people that do this. The issue is that guys aren’t all that great at repeating these, so it can be tough to tell how much hitting that line drive or that dinker was good luck or bad luck as well. I don’t disagree that wOBA might miss some loud outs and dribblers that sneak through, but it still does a better job than any other measure we have currently. Maybe this is a great opportunity for you to get to work.
“rather than comparing similar positions against each other”
This is the first step of determining defensive value. How many outs you made compared to the other guys at your position.
“WAR does not account for the fact that fielding shifts have become a thing”
Another great opportunity for you to do some homework. How do shifts affect the number of outs a fielder makes?
“there is a bigger gap between the best and the average in 2016 than in 1920.”
You have not demonstrated this. If a team of replacement level players won ~30% of their games back in 1920 and in 2016, and an average team wins, obviously, 50%, then the gap doesn’t increase. If you have evidence that a replacement level team didn’t win the same percentage of games today as they did 100 years ago, you’re on to something. Right now, you’re just posturing and making leaps in assumptions.
“If you can come up with a specific value that a batting event generates in any situation, I think you can probably come up with a specific adjustment for the added value of steroids.”
This still makes no sense. We have combed through the history of the game to see how much a single, or whatever adds to a team’s run scoring ability. We can easily test how much a single adds over an out. We cannot as easily test how much better PEDs made Bonds or Mays, or maybe Pujols, because we have no control factor and no idea to what level any player used. We’re just making guesses and assumptions about things we will never know for sure.
“Through their first 5 seasons, if you multiply avg. PA*(HR%+LD%+BB%-SO%) you are left with a value that assesses how well a player can hit the ball and how good their batting eye is. ”
Let me stop you right here. The reason wOBA is good, and something like this is not, is that the former doesn’t assume what assesses a players ability. It knows beforehand. It knows how much a hr or bb or so is worth, and weights them appropriately. Your measure doesn’t. You then proceed to just add up all the numbers as if they are equal, again without ever bothering to figure out how they should be weighted. And, of course, your defensive averaging doesn’t include the fact that CF is a more difficult position to play than 1B. You can move a Carlos Santana, a guy who doesn’t fit anywhere else on the diamond to 1B and get competent enough play. You can’t do that in CF. This is a seriously flawed approach, and doesn’t deserve the time of day.
I get it, you like to needle at some of the advances made in baseball analysis. I just have no idea why. You seem to be interested in asking questions about how well our analysis works, and wanting to find ways to improve it. You’re not some old fogey who says “average, home runs, and RBI is all I need”.