Weighted On Base Average: The New Batting Average
by Tigerdog1 on Dec 13, 2011 12:00 PM EST
in Detroit Tigers Analysis
Email
PrintwOBA is not the ultimate "one size fits all" statistic. There will never be a single best answer for that search. However, one of the best stats we have for that purpose is called wOBA, and we're going to make a concerted effort around BYB to use that in our daily conversations and posts.
Fangraphs explained wOBA like this:
Weighted On-Base Average (wOBA) is based on a simple concept: not all hits are created equal. Batting average would have you believe they are, but think about it: what’s more valuable, a single or a homerun? Batting average doesn’t account for this difference and slugging percentage doesn’t do so accurately (is a double worth twice as much as a single? In short, no). OPS does a good job of combining all the different aspects of hitting (hitting for average, hitting for power, having plate discipline) into one metric, but it weighs slugging percentage the same as on-base percentage, while on-base percentage is more valuable than slugging. Weighted On-Base Average combines all the different aspects of hitting into one metric, weighting each of them in proportion to their actual run value.
How do you calculate weighted on base average? The formula and rationalization of it is here, but simplified a bit by tangotiger, on-base percentage is going to be worth about 1.7 times more than slugging average. Okay, now I’m happy! Sit down right here, go no further, and see if you’re comfortable. I am. OBP is worth 1.7 times as much as SLG in the equation. Got it! I suppose that I take comfort in the knowledge that there are some really smart people, like engineers and rocket scientists and Philadelphia lawyers that have spent countless hours, even years, working on this stuff to boil it all down to the right equation. I trust their judgment. More to the point, I don’t have the time or the patience to challenge them. So, you can just take their word for it, use these numbers in conversation, and look really smart throwing them around. Actually, if you can get to the point of understanding the general concept of wOBA, you are well ahead of the curve, and apparently ahead of a couple of folks that I’ve mentioned in Part 1.
To make it easier to figure out what would be a good, average or bad wOBA among regular players, I've included a chart below showing select players from the 2011 season. There were 145 qualified players.
| Rank | Name | wOBA |
| 1 | Jose Bautista (Blue Jays) | .441 |
| 2. | Miguel Cabrera (Tigers) | .436 |
| 21 | Alex Avila (Tigers) | .383 |
| 25 | Hunter Pence (2 teams) | .378 |
| 34 | Victor Martinez (Tigers) | .368 |
| 48 (top one-third ends) | Nelson Cruz (Rangers) | .352 |
| 72 and 73 (middle point) | Emilio Bonafacio / Gaby Sanchez (Marlins) | .3415 |
| 75 | Carlos Lee (Astros) | .339 |
| 97 (Bottom one-third begins) | Danny Espinoza | .325 |
| 114 | Austin Jackson (Tigers) | .309 |
| 125 | John Buck (Marlins) | .301 |
| 145 (last) | Alex Rios (White Sox) | .266 |
Source: Fangraphs
The chart below shows a few of the Tigers' regulars finished 2011:
| Name | wOBA |
| Miguel Cabrera | .436 |
| Alex Avila | .383 |
| Victor Martinez | .368 |
| Jhonny Peralta | .353 |
| Brennan Boesch | .346 |
| Delmon Young (time in Detroit only) | .325 |
| Ryan Raburn | ,314 |
| Austin Jackson | .309 |
| Ramon Santiago | .305 |
| Don Kelly | .296 |
| Brandon Inge | .247 |
Source: Fangraphs' Tigers page
So you can see the top five on that list were above average compared to qualified major leaguers. Meanwhile Raburn, Jackson Santiago, Kelly and Inge would have ranked among the worst 33% in baseball had they had enough at bats.
Now, if anyone thinks that they’ve found the perfect statistic, think again. I hate to rain on your parade, but it is my opinion that the search for the one size fits all statistic is an exercise in futility. I say this for one simple reason. Baseball presents a different situation with each at bat. No two at bats in the same game are the same. So you’re not going to find the perfect hitter for every situation in one statistic. Show me the player that has the highest OBP, the highest SLG, and the highest OPS in the league and I’ll show you the best player for every situation. But there is, at most, one of them on the planet, and there are nine guys in your lineup. So you’re going to have to mix n match nine guys into nine lineup slots. If you’re looking for a lead off hitter, you shouldn’t be looking for the same traits that you find in a good No. 7 hitter.
The key to maximizing productivity of a baseball lineup is to try and get the players that have the best SLG and OPS in position where they will see the greatest number of situations with men on base. It logically follows that you want to put the players with the highest OBP ahead of the guys with the highest OPS and SLG. In other words, you need OBP at the top of the lineup, both OBP and SLG in the heart of the order, and players with a good SLG but not necessarily a high OBP behind them. This would generally give your team the best chance of scoring runs during a game, all other things being equal
What does this all mean for the Tigers in 2012? The Tigers ranked 11th, 9th, and 13th in the league in OBP in the top three spots in the lineup in 2011. With all the sluggers that follow them in the order, there were relatively fewer chances to drive in runs. They need to do better. The Tigers have "the guys that drive em in." They have the sluggers. Four different Tigers were among the top three in the league in SLG at their respective positions in 2011. The Tigers are just missing a couple of guys at the top of the lineup that don't necessarily make the big bucks. They just get on base.
So, if you run into Dave Dombrowski, please steer him away from Coco Crisp with his OBP of .314. Point him in the direction of Maicer Izturis, Alberto Callaspo, or Chase Headley. And if you go to Tigerfest in January and you see Jim Leyland, please tell him that it’s not about "on base, on base, on base". It’s actually about slugging percentage plus 1.7 times on base percentage. Good luck with that one.
To read more about wOBA, begin here and follow all the links.
85 comments
|
2 recs |
Do you like this story?
Comments
Your formula doesn't jive with me
Miguel Cabrera, 2011
OBP = .448
SLG = .586
Formula as described above = (1.7 * OBP) + SLG
Substitute: (1.7 * .448) + .586
Evaluate: (.7616) + .586
Evaluate: 1.3476 (does not equal .436 from chart above)
Must be something missing here. Mathematically, the formula above is guaranteed to be greater than a player’s OPS since it is essentially the same as the formula for OPS, but increases the OBP before adding SLG.
The other sports are just sports. Baseball is a love. ~Bryant Gumbel, 1981
Here's the long formula
((0.72 x NIBB) + (0.75 x HBP) + (0.90 × 1B) + (0.92 x RBOE) + (1.24 × 2B) + (1.56 × 3B) + (1.95 x HR) / PA (source)
NIBB = Non-intentional BB
RBOE = Reached base on error
The other sports are just sports. Baseball is a love. ~Bryant Gumbel, 1981
so then...
if i understand this correctly…reaching on an error is weighted higher than getting a single??? i’d rather get a hit…
Reaching on an error has a higher probability of moving runners over than a single does, I would assume.
Of course, hitting for an error isn’t really something that a player can try to do.
i see what you mean
thats the only possible reason i could think of, but i thought our goal here was to evaluate the hitters, not the defenses they are hitting against
Isolating defense is still not something I've wrapped my head around.
Especially on the pitcher side, where I am not fully on board with FIP.
Maybe errors also have a chance of leading to more bases.
Throwing errors to first often lead to two or three bases. If an outfielder bungles a bounce, it might get by them and result in a “triple”. Just spitballing here.
The other sports are just sports. Baseball is a love. ~Bryant Gumbel, 1981
Exactly Right
Think for a moment about situations where a single, a walk, or an error might occur.
Singles and walks merely put a runner on first. Errors might also put a runner on first, but often it puts a runner on 2nd or even 3rd. Runs score more often from 2nd and 3rd than from first base. A higher probability of producing a run therefore means the original event is given more weight.
Tango’s explanation:
When you reach base on error, you can reach 2B, 3B, or even home plate. A single where you stretch to 2B is actually a double (d’uh). And we don’t care about who was responsible. We are just counting "what happened".
What’s more, reaching on errors is actually considered something of a skill, typically associated with speedier players who force more rushed throws.
“What’s more, reaching on errors is actually considered something of a skill, typically associated with speedier players who force more rushed throws.”
I like that. Makes perfect sense to me.
Drink, and Drink, and Drink, and Drink, and Drink, and Drink, and Fight!
by Brand New Hero on Dec 13, 2011 8:49 PM EST up reply actions
So IBB have zero value?
I find that a pretty limiting stat since some of the best baseball players in the game get an alarming rate of IBB. I can understand not weighting it as much as a NIBB, but I have to think there are times (almost all of them) where an IBB would at least slightly increase run expectancy which mean that you are being more productive than if you were out.
Agreed
This is odd. I guess it’s not a PA, so it really has no effect. I would tend to think that it increases RE, but perhaps it doesn’t. Putting Bonds on 1st in mid-90’s seemed to be a much safer play than pitching to him. I guess IBB are a result of being damn good, which is probably already measured in the rest of the equation.
The other sports are just sports. Baseball is a love. ~Bryant Gumbel, 1981
What wOBA is doing, though, is isolating what a batter can do offensively.
A batter has no control over whether the opposing manager intentionally walks him or not. He does have control over whether he takes a normal walk, however.
I disagree
Miguel Cabrera being Miguel Cabrera has a direct effect on if an opposing manager intentionally walks him. If he was not the offensive force that he was he wouldn’t be intentionally walked as often as he is. He doesn’t directly have control over it but indirectly he does, he is being intentionally walked because the threat he poses is so much larger than anyone else on the team.
I don’t think that they should be weighted as high as the other stats, but they should have some value.
You could actually argue that they should be negatively weighted in some cases. Bautista coming to bat with a runner on second probably has a higher statistical chance to drive the runner in from second than whoever is batting behind him (I haven’t checked the numbers but that is irrelevant in this example).
I don’t know how it should be weighted, but the IBB is a indirect result, in certain situations, of the abilities of the batter at the plate.
But it's a strategy.
He doesn’t control the strategy of the opposing manager. This is setting aside the fact that baserunners also control this strategy. Managers are less likely to issue the IBB if first base is blocked.
Thus, if you wanted to stretch it, you can make the case that the GM has an impact on it by having a high OBP lineup.
Good discussion and I agree with both point of view
However, as stated in the article:
wOBA is not the ultimate “one size fits all” statistic. There will never be a single best answer for that search.
Nothing is perfect. It’s just another statistical tool to assess a player’s value and performance. The IBB can distort a little what Miguel Cabrera can do with a bat, but everyone has it. IBB are available to Miguel Cabrera and Alex Rios (Last in the list of wOPA)…so it’s not like a “exclusivity” of Cabrera.
A life spent making mistakes is not only more honorable, but more useful than a life spent doing nothing.
by JofpGallagher on Dec 15, 2011 7:13 AM EST up reply actions
Nah...the source missed the ). I just copy and pasted!
The other sports are just sports. Baseball is a love. ~Bryant Gumbel, 1981
I think it's implied that just the distribution is close to wOBA.
You’d have to scale this stat down to reach the wOBA average. They use the 1.7 and 1 weights in this calculation to highlight the importance of OBP vs. SLG, not to be practical.
yes
this is stat is mapped to OBP to make it easier to convert in people’s minds. (Personally, I don’t think that necessarily does make it easier, I think most people would have done better mapping it either to OPS or average). This is similar to FIP and xFIP being mapped to ERA.
by Kurt Mensching on Dec 13, 2011 12:57 PM EST up reply actions
Mapping wOBA
I wish that wOBA was not mapped to anything. Tango converted to OBP because it was convenient for the studies he was doing in the book. I don’t think he was thinking about mainstream usage when he created it. For most purposes, I think the conversion to OBP, BA or whatever is an unnecessary step making calculations more complicated.
However, If wOBA is to mapped to anything, I think OBP is the right choice. wOBA is just like OBP except it applies proper weights to the events. BA doesn’t have walks in it, so weighted BA (as opposed to weighted OBP) doesn’t really make sense to me. I understand that more people are familiar with BA than OBP, but I think those who are only comfortable with BA aren’t going to be all that interested in an advanced stat anyway.
I wouldn’t want to see it converted to the OPS scale, because OPS is a messy stat that doesn’t really measure anything. It’s useful for comparing players, but wouldn’t want to see it become the basis of other stats.
I’d rather just not see any conversion and have people get used to the scale the way they got used to the OPS scale
Practical isn't the right word
It is ‘practical’ to assign weights that correspond to what happens in the real world. It makes it easy to understand – which is practical.
I see
Thanks…that clears it up. I was confused; probably from not reading thoroughly and just jumping right to the math.
The other sports are just sports. Baseball is a love. ~Bryant Gumbel, 1981
hmmm....
Is 1.7 the new “42”?
I kid, but it is hard for me to wrap around the 1.7. Why not .7? From “The Book—Playing Percentenges” link above:
“wOBA simply represents WHAT happened, without asking WHY.” Maybe it is me, but I have to know WHY?!?!
Once again, I quote:
“The only thing to remember with wOBA is that it OBP, but with a bit of stretching here and there. A walk is 1 in OBP, but a bit less in wOBA (0.7). A HR is 1 in OBP, but much more in wOBA (2.0). Single is close (1 in OBP, 0.9 in wOBA). wOBA is just OBP, but with a bit more relevance to it.”
So in my mind, and from what I read, this stat is not perfect, but paints a better picture than OPS. Still this is arbitrary, and I side with that stats are not a PERFECT indicator for baseball success, but to be used as a guideline. You can “stretch here and there” till you turn blue…
Don't drive angry!
by Dan Gurney on Dec 13, 2011 12:30 PM EST reply actions 1 recs
wOBA, OPS+ and other weighted stats
Leave me kinda cold. I don’t have a Joe Morganesque disdain for them, but I think there’s a substantive difference between “this is what happened” stats that simply record outcomes such as BA, SLG, OPS, ERA, WHIP, K/BB, etc. and stats that have a weighting factor in them, where the however-educated opinion of the stat designer comes into play.
If the 1.7 multiplier were 1.6 or 1.5 or 1.4, players would be ranked differently on a wOBA list. That makes me question its real ability to tell us what happened and understand the relative value of ballplayers (which is, in the end, the whole point of “moneyball.”)
Your mileage, of course, may vary.
"Virtually all tactical ploys—the sacrifice bunt, the stolen base, the hit-and-run—operate on average to reduce run scoring." -- Eric Walker
by johnmoz on Dec 13, 2011 12:40 PM EST reply actions 1 recs
I think of it like this.
OPS assigns an arbitrary weighting factor of 1 to all of its inputs – BB, 1B, 2B, 3B, HR. wOBA simply assigns different weighting factors. It’s not like the weighting factors didn’t exist before. To me, the wOBA weights make more sense, because we all know that a HR is not the same as a walk. Is a HR exactly 2.0/0.7 times better than a walk? Maybe not, but I am positive that value is closer than 1 is.
by rcpratt on Dec 13, 2011 1:02 PM EST up reply actions 2 recs
OPS more arbitrary than wOBA
This is correct. OPS is even more arbitrary than wOBA. The wOBA weights are chosen according to the typically contribution to run scoring of BB, 1B, 2B, 3B, HR, etc. There is no practical reason for the weights in OPS.
Reading over everything
I think tangotiger just served to completely confuse the issue with that line about 1.7
by Kurt Mensching on Dec 13, 2011 12:45 PM EST reply actions
wOBA
Tigerdog, Cabrera’s wOBA would be 1.021 if wOBA = 1.7*OBP + SA.
wOBA actually equals (.72*NIBB + .75*HBP + .9*SI + .92*RBOE + 1.24*D + 1.56*T + 1.95*HR) / PA). I don’t think IBB (intenional walks) are counted as plate appeances, since the bat is taken out of the hitter’s hands.
What I would like to know is, how did they come up with the “.72”, “.75”, “.9”, “.92”, etc?
wOBA
Ha! Sorry! I figured his wOBA from 2003. It was 1.348 last year.
Weighted versus what happened
For John Moz and Dan Gurney,
The reason why you use a weighted percentage is because it’s a better predictor of outcomes — in this case, runs scored. That’s what the weighting is optimized for, predicting what a batter will do in the future, based on similar statistical patterns from the vast corpus of data that MLB represents.
Further, the traditional “what happened” stats are just as arbitrary, if not moreso. An at bat that ends in a walk is a good thing — not a zero. Add to that the fact that neither BA nor ERA is a good prediction of individual performance going forward, and that means that wOBA is a much better tool when evaluating players.
Yes and yes...
I completely understand, and respect the article and analysis by Tom Tango.
I did conclude by saying:
“Still this is arbitrary, and I side with that stats are not a PERFECT indicator for baseball success, but to be used as a guideline. You can "stretch here and there" till you turn blue…”
I just do not agree with the 1.7 to twist and shape the stat to be “close”. But alas, I read fangraphs to find my answer:
“Linear weights are optimised for the average baseball game, and start to fall apart when you drift too far away from that. "
Duh
Don't drive angry!
I don't think there's an actual x1.7 being applied there
it’s saying “when we add it all up with our weights, it works out to being about 1.7 x OBP …”
by Kurt Mensching on Dec 13, 2011 1:27 PM EST up reply actions
I know...
At first I was miffed about the 1.7. But I from the responses here and from reading, I see that the formula is more than that:
((0.72 x NIBB) + (0.75 x HBP) + (0.90 × 1B) + (0.92 x RBOE) + (1.24 × 2B) + (1.56 × 3B) + (1.95 x HR) / PA
I am just not a sabremetrics guy and I should start a support group…“Hi my name is Jon and I like stats, but not that much.”
Jim Leyland-“Hi, my name is Jim, and I like marlboros more than I like certain stats to evaluate performance” (*not a real quote)
I know that stats paint a picture with standard deviations and linear weights, but I think that there are too many variables. I will probably get asked to leave the room, but I want to see stats of games played in temperatures below 50 degrees at game time, between 50 and 60, 60 and 70, and 70 and above. Also for different elevations above sea level in each ball park. These are stats that also should be taken into account, but are not as “sexy” as wOBA.
Anyone up for doing the math of a wOPA x (temp/elevation) + defense – (average runs per game at ballpark)/winning percentage of team playing against?
I am trying to by funny sarcastic not a dick
Don't drive angry!
wRC+ is basically wOBA that is park and league adjusted, if you really are interested.
Which would basically account for average park temperature and elevation. Although you’re not going to be able to reverse engineer it to figure out how much a player’s wOBA changes based on temperature.
Don’t worry your not alone. My head just exploded.
by syper17 on Dec 13, 2011 2:20 PM EST via iPhone app up reply actions
OK I've edited a bit to try to clear up some of the misunderstandings from the original version
But I certainly encourage everyone to follow the links if they have questions. There’s a whole lot written about wOBA and Tigerdog just did a glance over things.
wOBA
wOBA is an attempt to attach a value to what a player did. If you want to say what a player actually did, why not ( (TB + NIBB + HBP + SB – CS) / (AB + NIBB + SF) ) ? To me, a sac fly (SF) is just a fly out and someone happens to be on 3B. Not charging an at bat on a sac fly is weighting that and applying a value to it rather than just stating what happened. It is not the same as a sac bunt, which is a purposeful out and therefore not an at bat.
different events have different values
Looking at more than 100 years of baseball stats, you can learn the run value of the events, and you find these run values remain pretty consistent on a season to season basis. So you use the value of these events to “weight” the stats.
by Kurt Mensching on Dec 13, 2011 1:14 PM EST up reply actions
more to the point, IMO
is that a batter is not charged with an AB if his fly ball scores a runner from third, but he is if his groundout scores a runner from third. Those are equivalent outcomes and should be treated as such.
Anyway, I like wOBA, somebody please show this column to Lynn Henning. You can e-mail him at: baseball-luddite@detnews.com
why yes, I'm slightly drunk at the moment
by Mark in Chicago on Dec 13, 2011 2:23 PM EST up reply actions
Sorry for the confusion- and I was trying to simplify things
I believe what Tangotiger was trying to tell us is that the 1 to 1 ratio of importance (weight) given to OBP and SLG is not an accurate measure of run creation. OBP is 1.7 times as important.
Billy Beane once arbitrarily stated that OBP is three times as important. Not quite, but it is to be given more weight than SLG in terms of run creation.
Tangotiger’s exact quote from the link above is:
1.7*OBP + SLG is going to be very close to wOBA.
"King of Minutiae"
I think that's just going to be confusing again.
Because the weights are all created such that it is scaled to an OBP level. Simply using that equation is going to give a number closer to OPS. The relative values will be the same, but the absolutes will not.
right
I really don’t like that line at all.
If people took it a step further and divided the 1.7OPS+SLG by 2.7, they’d find that gets closer to the actual wOBA but it’s still a very rough map.
I just don’t like it.
by Kurt Mensching on Dec 13, 2011 1:16 PM EST up reply actions
It's an over simplification, but I think it is meant to be just that
I wouldn’t be too worried about whether the more sophisticated fans that are up on sabermetrics can get their heads around the concept of wOBA. In fact, taking it back to a dramatically simplified equation is going to have exactly the result that we see here. They’re going to blow holes in an over simplified formula.
I think that the 1.7 factor is still useful for those that have no desire to delve deep into sabermetrics, for whatever reason (like the majority of the population that hates math), but would still like to know WTH we’re talking about with wOBA, since more and more fans and writers are holding that particular statistic in high regard nowadays. If there are other ways to simplify exactly what wOBA is, I would love to read about it.
I also believe that we are correct in viewing OPS as a stepping stone along the way to understanding wOBA. I would hesitate to take this discussion further- into all the factors involved in run creation, RAR, WAR, etc, without first laying out a comfortable resting place for the more casual fan’s understanding of wOBA.
"King of Minutiae"
until we get a defensive metric that is both accurate and agreed upon
I’m never going to be fully on board with WAR
by Kurt Mensching on Dec 13, 2011 1:35 PM EST up reply actions 2 recs
Agreed on WAR
I think WAR should be used very carefully. It works OK over a long career where some of the problems with defensive stats get evened out (regressed to the mean in a way). Thus, it’s good for Hall of Fame discussions for example.
In general though, I think offensive WAR should be listed next to total WAR as on Baseball-Reference. It’s important that the reader know how much of the total WAR is due to offense and how much is due to defense. That way he or she can decide whether the defensive portion makes sense. One could then adjust it upward or downward if it doesn’t seem to make sense for a particular player.
by LPanas on Dec 13, 2011 1:46 PM EST up reply actions 2 recs
As an illustration, sample a dozen MLB players across various defensive positions
Then compare their Fangraphs dWAR against their Baseball-Reference dWAR numbers. There will almost certainly be wild fluctuations.
The objective here was to take folks that are generally familiar with OBP, SLG, and OPS
and get them to a general understanding of wOBA. Fundamentally it is a weighted measure, so I like the generalization coming from the side of simplicity. Coming at it from an already advanced understanding isn’t going to make much sense.
But hell, I’d be happy just to get someone to explain the importance of on base percentage to our manager.
"King of Minutiae"
It correlates to run scoring as well or it differs from wOBA by a constant multiple are accurate statements related to this.
I’m not helping?
I am not a fan of Delmon Young.
Why not one step further and use WRC+?
I do like WAR stats for players ,not pitchers, because to some degree it does show what liability there is in the field or not.
It needs to be easy to understand
‘Weighted’, ‘Adjusted’ or ‘Scaled’ statistics can be (and are) superior evaluators, but a certain (large) percentage of fans don’t want to think about these sort of things.
OPS is easy and builds upon familar ground. You can’t QUITE explain it to an 8 year old without blowing their mind but you can probably explain it to a casual fan (say your significant other.)
For a baseball blog – it’s great, but OPS is better because a) it gets you 80% of the way towards player evaluation and requires 20% of the explanation.
Except that giving equal weight to OBP and SLG is fundamentally flawed methodology
Using a standard factor of 1.7 isn’t perfect by any means, but it’s another step in the right direction, IMO, and it’s one that even a fan who hates complex formulas can take. We’re just trying to advance the masses, and our manager, here.
"King of Minutiae"
For me OPS is good enough
Team OPS explains what- 89% of scoring? RC and other formulas can get you only another couple percent. I won’t knock those who want to play with more data and try to eek out a bit better predictor. I haven’t felt much need to join them either though.
I sometimes joke that it is probably the Golden Ratio that should be used: SLG + phi*OBP. Phi comes up pretty often and is about 1.7, so . . .
I am not a fan of Delmon Young.
sure, you say that only because you don't want to chagne your name
HighwOBA would look weird.
by Kurt Mensching on Dec 13, 2011 4:59 PM EST up reply actions
I literally laughed out loud. Touche.
Drink, and Drink, and Drink, and Drink, and Drink, and Drink, and Fight!
by Brand New Hero on Dec 13, 2011 9:32 PM EST up reply actions
I think this is proof that waiting periods to join the blog suck
"You, on the other hand, make Eeyore look like Rainbow Brite." -johnmoz
"I think of you more as the blue book style essay of sports journalism."-Kurt Mensching
Contributor, Bless You Boys
by David Tokarz on Dec 13, 2011 9:51 PM EST up reply actions
Given some of the things I post in the comment threads
I’m not sure I want to think about 8 year olds among BYB’s audience.
Better they should learn about these things in a BYB thread . . .
Than out on the streets. Or, in mlive.
I am not a fan of Delmon Young.
by HighOPS on Dec 13, 2011 4:27 PM EST up reply actions 1 recs
OPS is not easy. It's just been around longer.
Try selling OPS this way:
((H+HBP+BB)/(AB+BB+HBP+SF) + ((1B+2B*2+3B*3+HR*4)/AB))
That’s OBP + SLG. It’s not easier and it’s not a correct valuation of the various outcomes of a batters PA.
I think wOBA is fine
It’s a decent compromise between OBP and SLG. It’s not perfect by any means, but it certainly beats the heck out of OBP. Walks gets over-valued for some players, but just not nearly as much.
I would be thrill if they added SB/CS to the package
If you wanted more than a hitter view. Then add good defensive.stats to get one number. For pitcher do WHIP plus inverse of wOBA or slugging Pct.
You can add SB/CS to the package.
The general weights aren’t normally calculated on the seasonal level, but are usually +0.25 per SB and -0.50 for each CS.
The 1.7 thing is confusing. I'm not a fan of the shortcuts.
The other shortcut out there is (OBP*2)+SLG/3
So Cabrera would be (.448*2)+.586/3 = .494
His actual wOBA was .436. That shortcut has problems as well: 1) not entirely easy to do for most people on the fly. 2) it’s not that correct, either. Kincaid has a nice breakdown of it here.
My advice is to not worry about shortcuts at all.
Secondly, we need to keep in mind that the general formula linked to the in article by Tigerdog01 is not the formula that is used every year by Fangraphs.
The value of a homer changes depending on how high or low the run scoring environment is. The more precious the runs are (i.e. the lower the average wOBA in a league is), the more value a home run has. And vice-versa. So the weights are fluid and are ever-changing depending on what the league run environment is doing.
Matt Klaassen (who you may know from Fangraphs) posted about a year ago the single-season wOBA weights for all available years – 1871-2010. He gives the league average wOBA for each season, too.
In 2010, the average wOBA in the MLB was .320 and the value of the HR was 2.07. However, in 2009 the average wOBA was .328 and the value of a HR was 2.02.
Now, I don’t want to steal Tigerdog’s thunder as I assume he’ll be touching on this in the future, but wRC+ is like OPS+ only using linear weights. This is something that I think is the best offensive measure to look at. It’s adjusted for the run environment and the park. But I don’t want to get too far ahead of the game if it is in the works for the BYB authors here (I should’ve hung around here to get the audience for my wOBA primer on here a year ago, haha. Mine was poorly written, though. Everyone should read Lee Panas’ too! It’s wonderful.)
Accuracy was never the point of this article
The point was to highlight some of the fundamental flaws in the most commonly used statistics, and to introduce a better statistic for those that are not familiar with wOBA. The 1.7 factor is a generalized, over simplified number to show the relative weight that should be given to each of the more familiar statistics in terms of their value in creating runs. In between Billy Beane’s proclamation that OBP is three times as important, and Leyland saying that he prefers SLG over OBP, there is a number that is closer to reality.
Mike, I don’t have a patent on the thunder around here, and I certainly don’t claim to be the best person to explain some of these advanced metrics, so go right ahead and put your concepts and explanations out there. I have no great plan to blaze the path all the way to the pot-0-WAR at the end of the rainbow. Yourself and Lee are some of the best in the business at explaining these things. I just wanted to get more people interested in some of the advanced stats that are available. That, and I really would like to know whether Tiger management gets the basic concept of having guys with a higher OBP in the top three spots in the lineup. Judging by the players that DD signs, and by JL’s lineups, I think it’s fair to wonder about such things.
To Lee’s point below, I think that the way that OPS is concocted is almost hilarious. But it’s like a Reese’s peanut butter cup, in that the result can be quite interesting. No, it’s not an accurate measure of offensive productivity, but it improves on SLG by watering down the impact of the extra base hits, and it’s useful for comparison purposes. Still, it’s very much a slugger’s statistic, and it’s on the path away from AVG, SLG, and toward wOBA.
"King of Minutiae"
Mike actually still has the writing privs he had since I took over the place
by Kurt Mensching on Dec 13, 2011 6:54 PM EST up reply actions
TD1
You’ve certainly created discussion and raised awareness. I’m sure that was your point , so you can consider this a job well done. I know a heck of a lot more and I contribute to discussions on an unhealthy level. Many thanks for a perfect off-season discussion.
The other sports are just sports. Baseball is a love. ~Bryant Gumbel, 1981
Hope I didn't come across as critical.
It was well written and I’m glad it got the response and questions that it got. I’m always happy to see stuff like this spread. If the entire community took up citing a players wOBA, it’d be glorious, glorious fun.
That's kinda where we'd like to go with this
We shall educate the masses and march on MLB!
"King of Minutiae"
Regarding WAR and the differences between players on Baseball Reference and Fangraphs:
There are a few things in play here (we’ll focus solely on hitters).
The biggest culprit is that Fangraphs and B-Ref use different replacement levels. Basically, Fangraphs holds replacement level players a bit higher than Baseball References (which was created by Sean “Rally” Smith of Baseball Projection, who is now employed by a Major League franchise).
This doesn’t make either of the measures wrong, it’s a matter of preference. There is no right or wrong answer when it comes to replacement level.
Secondly, the way B-Ref/Sean calculated the individual offensive contributions is way different than Fangraphs. FG just goes wOBA → Weighted Runs Above Average. B-Ref is trickier and uses the team offensive environment. He did this because great offensive teams don’t play in the same run environment as the league. He also says that you should not be using the individual batting runs part of B-Ref. Here is what I mean:
That stat “Rbat” is not meant to be taken as the players individual offense and compare to other players that aren’t on the same team.
If you want to do that, you need to click on the “More Stats” portion at the top table labeled Standard Batting. When you do that, scroll down to the Advanced Batting section. There, you’ll find “BtRns” which stands for Batting Runs. That was invented by Pete Palmer in the 1960’s or the 1980’s (can’t remember which; Lee probably knows) and THAT is the number to compare players on different teams to one another. Why B-Ref buries this, I have no idea.
Here, you’ll see that Miguel Cabrera posted 71.3 Batting Runs and that number will be close to the Weighted Runs Above Average (wRAA) on Fangraphs.
You would think Fangraphs would know that "home run" isn't one word!
That’s a minor blip in an otherwise excellent post, though. Thanks!
Where the 1.7 comes from
Tango fit wOBA (or a very similar linear weights model) using a linear combination of OBP and SLG and showed that very accurate agreement is achieved with the OBP coefficient about 1.7 times the SLG coefficient. The details are at
www.insidethebook.com/ee/index.php/site/comments/why_does_17obpslg_make_sense/
wOBA is derived to be a monotonic function of a player’s value in runs/PA. For two players with the same wOBA, the one with the higher OBP has more offensive value because he generates additional PAs for his teammates.
Thanks, GW
I had seen that site months ago and looked for it, but couldn’t find it.
Tango starts out with a perfect explanation and then goes on to break it all down.
This is a step-by-step explanation as to why you should use some form of modified OPS, and not just OPS. If someone ever talks to you about OPS and how it should be weighted, send them here.
Very much worth checking out for those that want to go past the simplified 1.7 weight given to OBP.
"King of Minutiae"
Wow
Should not have read this during exam week. My head now hurts more. Thanks, dude.
"You, on the other hand, make Eeyore look like Rainbow Brite." -johnmoz
"I think of you more as the blue book style essay of sports journalism."-Kurt Mensching
Contributor, Bless You Boys
Outstanding
More reasons why Alex Rios is probably the worst baseball player that still plays every day (inexplicably)…..
A life spent making mistakes is not only more honorable, but more useful than a life spent doing nothing.
Comments For This Post Are Closed
Secondary Sidebar
User Tools
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
