Just How Lucky Were The 2009 San Francisco Giants?
by giantsrainman on Jan 21, 2010 3:51 PM PST
Email
PrintI did a little research on Fangraphs focusing on wRAA for hitters, UZR for fielders, and RAR for the pitching staff. The Giants hitters (including pitchers) combined for a very bad -118 wRAA vs a NL average of -27 (AL was +31). The Giants fielders combined for a very good +51 UZR vs a NL (AL and MLB) average of Zero. The Giants pitchers combined for a very good +201 RAR vs a NL average of +128 (AL was +172).
In total relative to the average NL team the 2009 Giants were 33 runs above NL average (-91 Hitting, +51 Fielding, and +73 Pitching). An average NL team won 80 games (AL average was 82). In the NL run environment these 33 runs were worth 4 wins so the Giants should have won 84 games and thus were lucky to win an additional 4 to get to 88. Some might argue that the above pitching and fielding contained some luck too. However I think this can be disputed especially for pitchers since RAR is based on FIP (not ERA) and thus not likely impacted much by luck. I would also argue that the defense was real as the leaders of the pack where Randy Winn and his +16.5 UZR which is very much in line with his 4+ year track record as a Giant and Travis Ishikawa and his +10.4 UZR which is very much in line with the consensus scouting opinion that he is one of the top defensive 1Bs in all of baseball if not the very best.
The Giants then are starting 2010 from an 84 win base and not an 88 win base. This begs the question, are we any better? We have lost Randy Winn's defense and benched Travis Ishikawa's. Have the players added not only made up for this lost defense with their gloves and/or bats but have they increased the Giants relative runs compared to the NL average? I am not going to address this now as I still have work to do on my "Rainman Projections For The Franchise And Friends" which I need to compete before I can do this. But, by all means share what you are thinking if you have already reached a conclusion on this.
One other item I would like to open up to discussion is the very large (44 runs!) advantage in RAR that the average AL pitching staff has over the average NL pitching staff. I understand the 58 run advantage the average AL team has over the average NL team in wRAA (DH vs pitchers hitting) but I just can not get my mind around this advantage the average AL team seems to have over the average NL team in pitching. What say the McCoven?
This FanPost is reader-generated, and it does not necessarily reflect the views of McCovey Chronicles. If the author uses filler to achieve the minimum word requirement, a moderator may edit the FanPost for his or her own amusement.
116 comments
|
0 recs |
Do you like this story?
Comments
Without crunching any numbers....
I would suspect that the improvement in the offense is going to be more than offset by the decline in defense. Expect that defense to negatively impact the pitching stats also.
If we win 88 games again in 201, we will be lucky…….and will also probably finish 3rd again.
Why isn't Sabean held accountable for leading the Giants into many years of mediocrity???
The RAR For The Pitching Staff Is Based On FIP And Thus Should Not Be Impacted By Defense.
by giantsrainman on Jan 21, 2010 4:29 PM PST up reply actions
Can't comment becuz....
I don’t know what that means since it wasn’t in the English language.
Why isn't Sabean held accountable for leading the Giants into many years of mediocrity???
Does This Help?
RAR is Runs Above Replacement calculated using FIP.
FIP is Fielding Independent Pitching calculated using the three outcomes a pitcher can control (strikeouts, walks, and HRs).
by giantsrainman on Jan 21, 2010 4:49 PM PST up reply actions
In that case, then FIP doesn’t adequately address how defense affects pitching. Fielders lack of ability to get to balls that fall for hits and lead to runs doesn’t seem to be measured. The Giants pitching will undoubtedly see more balls not being fileded that perhaps were fielded last year.
Why isn't Sabean held accountable for leading the Giants into many years of mediocrity???
What Are Your Thoughts On Why The Average AL Team Pitching RAR Is So Much Higher Then The NL's?
by giantsrainman on Jan 21, 2010 4:30 PM PST up reply actions
It’s possible that the AL has just had slightly better pitchers over the past few years. Even likely.
If you take 30 teams and randomly divide them into two groups of 15, the two groups are very likely to be similar in talent. But they are extremely unlikely to be identical in talent. Odds are that one group would be slightly better than the other. Stats suggest that, over the past 5 years, that group hes been the AL. It’s supported by Interleague records, as well as advanced studies. Maybe the AL has had slightly better GM’s. Or maybe it’s a payroll difference (I’m pretty sure the AL is slightly richer).
I was promised lasagna.
If you take out the Yankees and Red Sox, what’s the AL average? Because I saw an article a while back that said the payrolls of the two leagues would be the same if you did that (or maybe it was only the Yankees, I don’t recall for sure). If one league just spent more on pitchers than the other, that would account for any difference not accounted for by random variation.
If you want to take the Yankees and Red Sox out, you also have to ignore the two richest NL teams, otherwise the comparison is invalid.
I’m not sure that the Red Sox are the second richest team, though. Do they have a higher payroll than the Angles?
I was promised lasagna.
The point of taking the Yankees and (optionally) the Red Sox out is to compare two groups of teams with equal payrolls. The AL with the Yankees included has a higher payroll than the NL. Ex ante we would assume that would lead to the AL having better players, obviously including pitchers.
Irrespective of UZR, which most of us accept to be flawed, do folks really thing Ishi is (among) the best defensive 1B in baseball?
I’m honestly curious.
Billy Hayes: His job is better than yours.
I wouldn’t be that surprised if he really is a +10 1B. His +10 from this year was in only 817 innings. I would agree that probably should be chopped in half, but if you do that you still wind up with 9.5 UZR/150.
Is there some reason to think that the best fielding first baseman should not be worth at least almost as many runs as the best fielding third baseman? I realize they don’t have to make as many of those plays charging in, but other than that they have the same amount of territory to cover. And the best third basemen are around 20 UZR/150.
Is there some reason to think that the best fielding first baseman should not be worth at least almost as many runs as the best fielding third baseman?
Well, yeah. They have fewer balls hit to them than any other position.
I was promised lasagna.
But where does the fact that they can save other teammates’ errors with their glove come in.
To me, it’s pretty big when a 1B can pick bad throws out of the dirt or leave the bag to make a catch that prevents an overthrow and keeps the runner at first.
that doesn’t figure in to UZR does it?
"We're in this thing!" My adopted Giant: "Raptor Jesus" Guzman, "Sweet Jesus" Guzman and Jesus H. Guzman.
Scoops matter a tiny bit, but not as much as most fans probably think.
When we regress these numbers in order to turn them into "true talent" scooping ability, we get slightly smaller values, depending on the number of matched throws (sample size) of course. In fact, it appears that the spread in talent between the best and worst "scoopers" at first base is on the order of 2-3 runs, plus or minus (a 4-6 run spread). So before you start opining about how your favorite first baseman is so great defensively because he "saves so many errors," consider that scooping ability is probably worth less than a ¼ of total defensive ability or value at first base. Fielding grounders is at least 75% of the package and "scooping" is the rest. But every little bit helps.
http://www.fangraphs.com/blogs/index.php/first-basemen-scoops/
#1 FanShot Champion
Does their definition of “scooping” apply to more than just throws in the dirt? It seems like minimizing the damage from all sorts of bad throws would be worth more than 1/4 of 1% of a 1B’s value.
"We're in this thing!" My adopted Giant: "Raptor Jesus" Guzman, "Sweet Jesus" Guzman and Jesus H. Guzman.
Baseball Info Solutions (BIS) tracks wide throws, throws in the dirt, etc.- essentially everything you could imagine in terms of a first baseman’s abilities to handle poor throws.
Their findings are essentially the same as MGL’s, who used only raw data, and Sean Smith’s, who did something very similar. So what you’re seeing in MGL’s article gives the same results if you include all types of throws.
How big of a difference is there in number of chances between first and third? If 40% of hitters are lefty, there shouldn’t be much difference unless you assume that righties are much more likely to pull the ball than lefties are.
In terms of total ABs, so including LH appearances by switch-hitters. I got that from this article. The split is about 55/30/15 R/L/S, and since S is about 2/3 L, there’s your 40%.
I think not including S at all would be the proper way to look at it, so it’s about 30/85 L….
On second thought, going back to the context we’re in (balls to 1B vs. 3B), I guess your way is the right way to look at it.
by Missing Barry on Jan 22, 2010 11:50 AM PST up reply actions
For some reason I do have a feeling RHH hitters pull the ball more, but I have no idea why. But even if it’s 60-40, that’s pretty significant, and there’s also the throwing thing.
I was promised lasagna.
60-40 is not very significant at all unless you think hitters pull the ball a pretty overwhelming percentage of the time. If we say that 60% of hitters are right-handed, and 60% of all balls in play are pulled, then we would have 52% of balls in play to the left side and 48% to the right.
No, because it’s not right side vs. left side, it’s down the 1Bline vs. down the 3B line (or close to them).
If I’m reading them correctly, last year there were 5,422 fielded by 1B, and 7,024 fielded by 3B. So that’s 56-44.
I was promised lasagna.
It doesn’t matter (mathematically) whether it’s close to the line or not. If I say I have 100 balls hit within x feet of a baseline and 60% of the hitters are RH, all I need to know to know how many went towards third compared with first is the respective propensity of RHHs and LHHs to hit a ball close to the line to each side. So if I say 75% of balls hit by a RHH within 20 feet of the baseline are hit towards third and the mirror image is true for LHH, then 55% of balls within 20 feet of the baseline will go towards third base.
As to the 5422 and 7024, that certainly is useful info, but it isn’t exactly the information that we need. We need to know how many balls were hit into the first baseman’s and third baseman’s zones, not how many were actually fielded by them. Otherwise, it’s biased by the fact that the average first baseman has less range than the average third baseman.
Even based on that number though, there’s only 30% more balls hit to third than first. Between this and the lack of throwing there’s gonna be less variation in 1B than 3B. But even half as much seems like a bit of a stretch, and if it were half then a 10 UZR/150 out of a 1B would be expected from the best defenders.
Well, I think your other assumption is also a bit off. We’re not talking about whether or not it’s possible for a 1B to pull of a +10, but rather whether it’s possible that it’s his true talent level. Whether he can he sustain it. How many 3B have actually sustained a +20 UZR over, say, 3-4 years? I’m not aware of any. Over the past 3 years, Zimmerman is +14, the next few are 10-12.
Anyway, why not just look at the actually numbers – how many 1B during UZR’s existence have put a ~+10 consistently? I’m pretty sure there haven’t been any.
I was promised lasagna.
That’s certainly a fair point, but for starters, Doug Mientkiewicz has a career 8.9 UZR/150 at 1B in about 5000 innings, which is in the ballpark. Furthermore, if it weren’t for Franklin Gutierrez we probably wouldn’t have thought a guy could sustain a CF UZR of 25. It’s not as if we have 50 years of UZR data to look at.
We’ll have to see what Ishikawa does if he ever gets another shot, and I don’t know that even if he does he will sustain his seemingly magical glovework, but it at least seems possible that a guy could be a true 10-run fielder at first base.
As for the throwing thing, I thought there was way more variation in range than throwing ability, but I dunno. Obviously that would be more true for OFs than IFs.
I don’t think the platoon split is the biggest reason for the difference in chances, though. The biggest issue is that the 1B’s range is limited by their responsibility to cover 1B. It doesn’t matter if the Ozzie Smith of first basemen could range 30 feet to his right if there was no one to throw too at first, and that range would be completely wasted when holding a runner.
VAE PVTO DEVS FIO
He's probably among the best defensively
But not +10 runs over the course of any season going forward.
Aaron King is still my homeboy... iffy mechanics and all
McFAQ for all you newcomers out there.
GET THAT VORP AND WHIP SH!T OUTTA HERE!!!
Whomever Sabean signs this off-season will make a good platoon partner with Ryan Gark-ohh... nevermind...
I do believe the Giants were fairly lucky last year; despite their performance, they also continued to project as one of the lower teams according to BtB’s metrics, though I have a very hard time believing the complete accuracy of those metrics given they actually put us below the Royals at one point. An 84 win base seems accurate.
Next year, I expect us to win more than 84 games; the rotation will likely hold up longer (Johnson’s injury really messed it up), the bullpen will be better with a second lefty in Runzler instead of Howry, though Howry was a good pitcher, the offense is improved, though the defense may suffer (I’ll take an is over a may), and the lack of improvement in LA and Colorado bodes well for us, though Arizona’s plethora of good short-term moves doesn’t (then again, that hurts Colorado and Los Angeles too). The Dodgers not starting the season 20-0 or whatever will help make the division winnable as well as the Wild Card. Postseason play is unlikely, but more realistic than last year.
Regarding the league adjustment, I don’t think that the AL pitchers are better, I think that the AL hitters are better, so they have to give a pitcher additional runs for pitching in the AL because it’s a more difficult environment. The superiority of AL hitters is likely due two things: the DH rule (removing a below-replacement bat and putting in a high-value bat before it’s positionally adjusted is very valuable to the team’s offensive performance) and the presence of the Yankees, who entice many good players to both leave the NL and stay in the AL with offers of fat piles of money. However, I do believe that a lot of metrics put the AL’s advantage far too high. There are better teams there, but a lot of it has to do with the DH and money.
The AL has likely been better for these reasons, particularly the DH. The presence of a DH, even in interleague play, helps because it offers the chance to sub out a below-average player for an above-average player in NL-only games. NL teams do not have this advantage; presumably, their best offensive players are already starting. The inclusion of the DH in NL games likely does negatively impact the defense, but probably not to the extent that their offensive contribution helps, especially if they are replacing the worst overall producer in the lineup (though this is unlikely; the least productive spots are typically C/SS/2B/3B/CF, and if a DH could play one of those he wouldn’t be a DH). Still, the opportunity that a DH gives is an advantage; either playing the DH or playing the typical starter produces more runs, so the manager has a chance to improve the lineup regardless.
by quincy0191 on Jan 21, 2010 7:12 PM PST reply actions 1 recs
the rotation will likely hold up longer (Johnson’s injury really messed it up)
I don’t know about this – Johnson was the worst pitcher on our staff. Starting pitchers get hurt, it happens, it’s why a team like the Red Sox started last season with 7 of them. We have yet to sign another starter, so right now we’re actually looking at a team that’s exactly where we were last season in terms of the rotation after Johnson was hurt. And there’s the likelihood that one of our good pitchers get hurt…..
by Missing Barry on Jan 22, 2010 7:32 AM PST up reply actions
Johnson did not perform well (due in large part to a ridiculous HR/FB of 19.2%), but with him out of the rotation, we didn’t really have another option for a fifth starter. Martinez, Miller, Sadowski, etc. were probably worse than throwing Johnson out there every fifth day. In any case, losing a starter and having your next best option be a long reliever or a guy with a 5.00 ERA in Triple-A is very bad.
Yeah, and my point is that that’s exactly the situation we’re in right now, so until/unless we sign a 5th starter, we can’t expect any sort of benefit from not having a hurt Johnson, because instead we have no johnson!
by Missing Barry on Jan 22, 2010 1:04 PM PST up reply actions
Yeah, but either way, you’re not getting laid.
by Missing Barry on Jan 22, 2010 1:56 PM PST up reply actions
Yes
Right now is equivalent to having a hurt Johnson, because a hurt Johnson sees no action (GET IT???). But the point I made was that Johnson’s injury messed up the rotation more than his absence does now, because we didn’t have the fallback option of Bumgarner then; at least now we could go long-reliever, AAA starter, OR top prospect who isn’t ready. None of them are great choices, but having three meh choices is better than two (especially if one doesn’t work out).
I Think Bumgarner Just Might Be Ready
Is there a risk that he isn’t? Obviously yes. Do we know he isn’t? Just as obviously no.
by giantsrainman on Jan 22, 2010 9:48 PM PST up reply actions
IMO, we need a fifth starter, and we need one a lot more than we need Huff or Molina. Risk vs. reward says to keep Bumgarner in the minors next year. If he’s ready, he wastes a year in AAA when he could have been helping the team, mostly by saving money (it’s doubtful he’d be that much of an improvement over Garland or Bedard or Davis or Sheets, etc.), but possibly with good production, though I doubt that; this would be his formative year, the “figure things out” year, much like Lincecum in 2007.
But if he’s not ready, we derail his development by putting him in a situation he can’t handle, especially if the loss in velocity is real and he needs time in the minors to get stronger. Best case scenario after that, he loses most or all of 2010, spends 2011 in the minors getting his form back, and isn’t in the rotation until 2012; worst case, he loses confidence, tries to overthrow or overcompensate in some way that gets him injured, and essentially becomes a blowout. Considering that situation is definitely possible (he’s spent a year in A-ball, half a year in A+ and half a year in AA; that is not an indication he’s ready), Bumgarner should definitely be starting 2010 in AAA.
Pitchers Are Different Then Position Players
Many have only so many pitches in there arms and you do not want to spend more of them in the minors then you have to. It is time for Madbum to start his MLB career and thus for the Giants to live with whatever struggles he has as he continues to develop. I als think the Giants track record with pitchers shows that it it easier for the professionals 9theirs at least if not all) to tell when pitchers are ready then it is for them to tell when hitters are ready. I very much trust the Giants judgment here.
by giantsrainman on Jan 23, 2010 2:19 AM PST up reply actions
I agreed with you that we should leave Posey in the minors to delay his service time. If we bring up Bumgarner, I really just don’t see the point – it completely undermines the concept of delaying service time. Basically, it would mean by saving Buster’s PT we’re using up Bumgarners, and that just doesn’t make sense to me. I’m not sure how much I buy “pitchers only having so many pitchers in their arms”, at least when we’re talking about a 20 year old who we have 6 years of team control for. I’d much rather hiim spend too much time in the minors and be ready to step into a starters role and be effective immediately. Right now, I just don’t see how he could be. Even if his fastball velocity comes back (or doesn’t but the fastball is effective anyways), he doesn’t have the offspeed stuff to be anything more than a long reliever right now. He simply doesn’t have the other pitches to go through the lineup 2-3 times consistently.
You do have the point with the Giants evaluating pitchers, as much crap as we give Sabean for hitters, he has a pretty good track record of evaluating his pitchers, so I would trust their judgment here, I just don’t know if they understand how to properly value the contribution and cost/reward aspect of it, even if they can properly evaluate how effective he’ll be. So if they’re bringing him up and don’t think he’ll be an effective starter right now, I think that’s a mistake. If they think he can be an effective starter this season (and I’m talking 1.5+ WAR type starter, and hopefully +), well I guess I’ll defer to their judgment….I just hope we don’t make the decision to bring him up for any other reason.
by Missing Barry on Jan 24, 2010 3:27 PM PST up reply actions
Neither FIP Nor RAR (which is based on FIP) Make Any Adjustments League To League
So I am having a hard time seeing how AL hitters being better then NL hitters would result in AL pitchers having better FIPs and RARs then NL pitchers.
Regarding the league adjustment, I don’t think that the AL pitchers are better, I think that the AL hitters are better, so they have to give a pitcher additional runs for pitching in the AL because it’s a more difficult environment.
I just do not see how this explains the FIP and RAR for pitchers difference between the two leagues at all.
Well
I would imagine the numbers are league-adjusted somewhat to at least account for the DH rule, since that’s a fundamental change in the way the game is played. That would be reflected in the RAR.
If not, and to explain the FIP, assuming that hitters are better in the AL implies that pitchers need to be better to be successful. Therefore, pitchers need to have better K, BB, and HR rates (i.e. the ingredients of FIP) because those are integral to the success of a pitcher. Over time, the shift of good pitchers toward the AL and mediocre pitchers toward the NL would occur. Say you have four pitchers, two good and two mediocre, and one of each is pitching in each league. The good AL pitcher would likely stay in the AL since he is valuable. The bad AL pitcher can either continue to play for a bad team, get traded to another club, or accept a lesser role. If he is traded to an NL club and becomes successful because the standards are lower, he stays. Otherwise, he continues to be traded or otherwise moved until he becomes successful or is unable to find work. The opposite happens for the other two; the mediocre NL pitcher is has the same options as the mediocre AL pitcher, only assuming the AL is superior offensively, his numbers would be even worse upon switching leagues, leading either to a trade back to the NL or the end of his career. The good NL pitcher can either stay in the NL or he can go to the AL; given the increased resources in the AL, he will likely spend time there at some point, where his numbers may worsen, but stay acceptable. Again, if the AL is offensively superior, it is more difficult to find good AL pitching, thus it is likely he will be courted strongly by high-payroll AL teams over pitchers who have only had NL success, but may be good in the AL. This is both because there is a perception that AL pitchers who succeed have proven themselves in the AL, and because there is some truth to that perception (otherwise it presumably would die out).
This is all assuming that the AL produces more offense, which I think is undeniably true. The presence of the DH creates a better offensive environment, and again the increased resources tend to have the high-impact players gravitate toward AL clubs.
So the AL contains better hitters AND better pitchers in a vicious cycle; the offense gets better, so the pitchers have to get better to compensate (this doesn’t mean the individuals improve, but that the teams import better talent from the NL), so the offense has to get better, etc. This cycle originates because of the increased resources in the AL; higher payrolls mean better players (when used correctly), so they import better players from the NL while exporting inferior players, and their competitors must compensate, and the only reason the cycle stops is because there is an upper payroll limit (eventually a team must stop spending, and there is enough high-quality talent that the NL clubs can entice players as well because teams top out before talent does).
But then, as you say, how does this produce better FIPs, as the talent balance shouldn’t get skewed? Well, I think it’s because there is replacement level talent, and the best players benefit from that. Say each player is graded on a scale of 10, with 10 being the highest. A 10 pitcher vs. a 10 hitter should produce the same result as a 5 pitcher vs. a 5 hitter, etc. So if there are more 10 pitchers in the AL, and they face generally superior lineups, but those lineups also contains inferior hitters (which they must; payroll limits a team from having a lineup of 10 hitters), they grade out as average against the best, but excellent against the mediocre. The same is true on the hitters’ side. So the pitchers boost their FIPs by facing lower-level talent, and the hitters boost their wOBAs by facing lower-level talent, which must exist without the presence of unlimited payroll. FIP also doesn’t account for non-HR hits, which do help wOBAs and batting RAR, so it is possible to have higher FIPs and higher wOBAs in the same league. I believe that is the key source of the difference: better pitching in the AL combined with a metric that doesn’t evaluate all forms of offense leads to a better overall stat, and better hitting in the AL leads to a better overall wOBA. And that is why a metric like ERA, which accounts for more forms of offense, is closer for both leagues; the NL’s inferior hitting is balanced by its inferior pitching, and the AL’s superior pitching is balanced by its superior hitting.
I hope that makes sense and it’s a logical progression of thought. I must point out that I do not have much backup for this besides my own ideas and reason, which may be flawed and thus invalidate the entire essay (because it’s now an essay). If not, I’m not going to try to answer the question because I don’t know enough, I don’t want to take the time to learn what I’d need to know, and I’ve spent plenty of time on it already.
The Giants offense appears to be improved, although not by quite as much as before Bengie Molina signed.
The pitching appears to be another story. Jeremy Affeldt will likely double his ERA, and Matt Cain, Barry Zito. Brandon Medders and even Brian Wilson may have a hard time duplicating Further improvement by Tim Lincecum and Sergio Romo won’t be enough to fully offset the declines.
Another factor not often mentioned. The Giants need to find someone to replace the ERA’s in the low threes of Justin Miller and Bobby Howry. Hopefully they will find someone with better timing than Howry.
I know
I was about to say “Do you think Lincecum could be even better than the last two years?”. Then I saw the handle.
The Giants Way™"If anybody deserves credit for this year’s turnaround it’s these two people, Brian and Bruce," Neukom said. "The encouraging thing is we think we’re back to playing baseball the way it ought to be played."
The Giants offense appears to be improved
All I have to say to this is…PUFF Magazine!
say hey nation is the Ralph Nader of McC.-Xanthan
by say hey nation on Jan 22, 2010 11:44 AM PST up reply actions
POOF
I know you nerds know NOTHING about the real game of baseball, or any other athletic endeavor requiring teamwork under physical stress.
Mr. F! | comics | art | New Nattowear | Unofficial McImage Directory
Do You Really Expect An Increase In Team FIP?
I could indeed see an increase in team ERA as it has some luck in it. But an increase in FIP is what would have to happen to result in an increase in RAR.
by giantsrainman on Jan 22, 2010 2:33 PM PST up reply actions
I do – I don’t see much room to improve, so I think it’s a lot more likely to get worse than improve, so if I were a betting man, I’d bet we were a little bit worse (the average of all likelihoods)…
by Missing Barry on Jan 22, 2010 2:40 PM PST up reply actions
Then We Disagree - I Se No Good Reason To Expect This
by giantsrainman on Jan 22, 2010 2:43 PM PST up reply actions
Random variation. Things are more likely to move towards the mean than away from it….
by Missing Barry on Jan 22, 2010 2:56 PM PST up reply actions
I Think These Pitchers FIPs Were Already At Their Mean (True Talent Level).
I see no reason to think they were not.
by giantsrainman on Jan 22, 2010 3:14 PM PST up reply actions
I think Lincecum and Cain will stay pretty much the same, Sanchez will improve, Zito will worsen, and the fifth starter will be better than Johnson. So I expect a little improvement from that and because I think Sanchez will go more positive than Zito goes negative. And if Zito can hold form, that’s a great rotation.
Even if they’re at their mean, though, we have to expect some amount of random variation, and especially with a guy like Lincecum, there just isn’t much room to go up – it’s probably a skewed distribution where he’s more likely to underperform than overperform. Especially once you factor in the chances of injury/arm fatigue and other issues that might negatively affect performance….
by Missing Barry on Jan 24, 2010 3:29 PM PST up reply actions
I Don't Buy At All The Therory Of Regression To League Mean
Players regress to their true talent levels not to league mean.
by giantsrainman on Jan 24, 2010 4:20 PM PST up reply actions
Players regress to their true talent levels
Yes, this is right. What you’re not getting is we’re estimating true talent levels and it’s subject to the sample sample error every other sample observation is. We don’t know their true talent level, and there’s a need to regress “towards” the league mean because as you go away from the mean, it becomes more likely there is sample error pushing you that direction. Not every player should be regressed to the league mean – I am 100% sure Tim Lincecum’s talent level is not the league average, it is undoubtably better than that. That’s not the issue. The issue is our sample – it’s more likely his true talent is closer as opposed to farther from the league mean than the sample we have. In addition to that – it’s probaly much easier/more likely for a players “true talent level” (these are always changing, remember) and/or performance (also highly variable) to go from a FIP of 2 to 3, rather than 2 to 1. Given our pitching was at the high end of the spectrum….well, it’s simply more likely to get worse than get better.
Also, injuries matter – they will happen. No idea who it’s going to happen to, but you have to account for that possibility and include it in some form in your estimates. Like I said, if Lincecum gets arm fatigue or pulls his groin or some other minor injury he can pitch through – his performance will be worse. If he gets hurt more than that, we lose some of his production. So…yeah, the possibility of that happening has to drag your expectations down in some way.
by Missing Barry on Jan 24, 2010 8:43 PM PST up reply actions
I think you’re pretty accurate on your assessment of our talent level, GRM. My guess is our injury situation last year was fairly average, so we should expect a similar amount of injuries this season….though we’re screwed if Sandoval or Lincecum goes down for a long time, so let’s hope that doesn’t happen. I do think we’ve gotten a little better this offseason, but only marginally so. We’re maybe an 86 win team now? Not good enough for me, because that’s not a playoff team. And who knows what Bochy is going to do to screw up our lineup.
I think it all comes down to Renteria
With the exception of injuries to key players, I think Renteria is the key in 2010. If his defense and offense really were affected by injuries and we get who we thought we were getting, the team had a very good chance of improving on last year’s 88 wins.
A healthy and productive Renteria will have a ripple effect in that it would allow Uribe to be a super sub and give Bochy more opportunity to use the roster coreectly with rest days and matchups.
If Renteria sucks again and Uribe has to play SS full-time, it won’t be pretty since most of expect regression from Uribe and the negative ripple effect will compound things.
"We're in this thing!" My adopted Giant: "Raptor Jesus" Guzman, "Sweet Jesus" Guzman and Jesus H. Guzman.
Even if Renteria is better, isn’t the point of this thread that we’re really building on 84 wins, not 88?
by Missing Barry on Jan 22, 2010 10:33 AM PST up reply actions
I think the team that’s currently assembled is about 4 wins better than what we should have expected from last year’s team.
"We're in this thing!" My adopted Giant: "Raptor Jesus" Guzman, "Sweet Jesus" Guzman and Jesus H. Guzman.
How so? I think the pitching will still be very good but it should regress a little. Freddy Sanchez looks like he’s going to miss some time to start the year which is never a good thing for a 2B in his early-30’s. Huff isn’t any better than Ishikawa and DeRosa isn’t any better in LF than Lewis.
The team looks remarkably the same to me.
#1 FanShot Champion
The pitching looks certainly worse to me at this point. Instead of the twilight of Randy Johnson, there’s….??? MadBum? JOEY? Pucetas? Probably not that much worse than Johnson, actually, but a little bit. And of the four remaining starters, I only really have confidence in Lincecum being as good. Cain outperformed his peripherals, Zito is still Zito, and Sanchez, who the hell knows?
The bullpen was crazy good last year. However, the guys who are back (Affeldt, Romo, Wilson, Medders) are various degrees of unlikely to be as good, and more important, Howry and Miller are gone. Is it reasonable to expect Joaquin and Runzler to replace their production?
As to the lineup, I totally agree. With Sanchez out, they don’t really seem significantly better anywhere. I suppose we can expect more out of Uribe than they got out of the Manny Burriss Experience. And maybe a smidge more from Derosa than Lewis. Maybe. All told, I’d guess two wins more from the lineup, at most.
The people who think this team is better than last year’s are confusing to me, because I don’t know where the improvement is supposed to come from.
Well, if you value players by WAR, then Cain performed how we expect and given his consistency in the FIP department, he should maintain that production. With the lineup, no Velez = good. Anything is better than Winn’s offense last year. DeRosa is alright, Huff sucks but is at least as good as we got out of 1B last year. Renteria might be as bad, but he basically can’t be worse. Our 2B production last year was pretty god awful. I expect a slight improvement. Not much. Not a $23M+ improvement, mind you (yeah, we are spending that much on our FA lineup acquisitions, blows the mind, huh), but a small improvement nevertheless.
by Missing Barry on Jan 22, 2010 2:02 PM PST up reply actions
The RAR I Used Is Soley Based On Peripherals (FIP) So Why Should It Get Worse?
by giantsrainman on Jan 22, 2010 2:38 PM PST up reply actions
I was referring to actual runs allowed, not how well the starters pitched, which was an error if we’re talking about FIP. I expect Lincecum and Cain to pitch about as well (barring injury, which scares me to even say), Zito to be a bit worse, and I have no idea what to think about Sanchez. And at this point I expect the fifth starter to be a tad worse. But the main reason I expect this year’s pitching to be worse is the bullpen. Should we really expect Joaquin and Runzler to be as good as Howry and Miller were last year? For that matter, do we expect Medders and Affeldt to be as good as they were?
FIP And Thus RAR Wise Why Shouldn't We Expect Them To Be As Good.
Miller has a bad FIP last year as did Randy Johnson.
by giantsrainman on Jan 22, 2010 3:31 PM PST up reply actions
So then you expect 88 wins instead of improving on that, right? I think we’re more like 1-2 wins better, personally, though I haven’t done the math.
by Missing Barry on Jan 22, 2010 1:05 PM PST up reply actions
I haven’t heard anything about Renteria’s defense being affected by injuries. He had an elbow injury and his problem was range, so I wouldn’t expect any improvement there.
I was promised lasagna.
He didn’t seem to get much on his throws to me. If that was because of the elbow, it might have caused him to play in a little more than usual (since he would need to get the ball earlier if he were going to throw a guy out with a weak throw), which would result in reduced range.
Note that I never spent much time watching Renteria before this past year, so I have no idea if a) his arm strength really was reduced, or if it was, b) he really was playing in a bit more. So really it’s just a theory of how an arm injury hypothetically could affect range.
For comparison
Baseball prospectus has the giants at 83 wins, with the 5 wins coming from “a magical ability to get extra runs from thier hitting (in this case EQA)” – luck.
I don’t know how team UZR correlates year-to-year, but with no Winn and less (?) Ishikawa, I don’t see that number getting any better. Will be interesting to see if that translates into worse run prevention.
TEMPORARY SIG AWAITING FINAL SABEAN APOCALYPSE
-------
PARPG- Indy post-apocalyptic roleplaying game in development
I Would Like Some More Feedback On The Following
One other item I would like to open up to discussion is the very large (44 runs!) advantage in RAR that the average AL pitching staff has over the average NL pitching staff. I understand the 58 run advantage the average AL team has over the average NL team in wRAA (DH vs pitchers hitting) but I just can not get my mind around this advantage the average AL team seems to have over the average NL team in pitching. What say the McCoven?
I dunno, I really haven’t kept up with research into the AL/NL disparity, all I can say is the W-L record over the years is pretty bad for the NL. That sounds like a huge difference, but at the same time, there really does seem to be a pretty big difference between the leagues…
So yeah, I do know there’s a talent difference favoring the AL, I just don’t know how big I should expect it to be.
by Missing Barry on Jan 22, 2010 2:57 PM PST up reply actions
A Different Approach
Using a BsR formula for the 2009 NL (RMSE for wRC was 17.2, mine’s at 14.4) has the Giants scoring 628 runs. Using FIP scaled to RA/9 and then adjusting for UZR, we get 614 runs allowed. Run this through PythagenPat and you get a .508 win%, or 82 wins and 80 losses.
So, perhaps they were even a bit luckier than what you suggest.
BP’s adjusted standings, based on UEQR and UEQA, also have the Giants at 82-80.
I like the BsR approach
Thanks for not being lazy like me
Aaron King is still my homeboy... iffy mechanics and all
McFAQ for all you newcomers out there.
GET THAT VORP AND WHIP SH!T OUTTA HERE!!!
Whomever Sabean signs this off-season will make a good platoon partner with Ryan Gark-ohh... nevermind...
Can You Provide More Details On How This Is Calculated?
In particular I don’t know what you mean by “RMSE” or you calculated “mine’s”. I am also not sure what these wRCs represent. Are they the teams’ wRC relative to some standard or what? If so. what is that standard?
by giantsrainman on Jan 22, 2010 9:58 PM PST up reply actions
Sure! “RMSE” stands for “”http://en.wikipedia.org/wiki/Mean_squared_error" >Root Mean Square Error." It’s the most common way we identify accuracy with run estimators. Generally speaking, most estimators sit around a RMSE of 23-25. Best-fit equations, such as the method I used, yield greater accuracy because you’re using coefficients that match the run environment.
I’m a bit surprised that you’re not familiar with wRC, since you are fluent in “FanGraphsese.” (not meant as a shot at you, of course!) wRC are “weighted runs created.” While we have wRAA, which are runs above a baseline of average, RC are on an absolute runs scale. The difference in the calculation is that we don’t stop with (wOBA – LgwOBA)/Scale*PA; it takes it a step forward and sort of “cheats” to convert into absolute runs allowed by using league runs per plate appearance. This gives us an equation of:
(((wOBA – LgwOBA)/Scale) +(LgR/LgPA))*PA
This estimates how many runs the team (or player) created.
The reason why I opted not to use wRC or wRAA is because it is limited in the statistics used. It incorporates hit type, along with NIBB/HBP and SB/CS. Tango himself will tell you that the formulation for SB/CS weights are fudged, and I don’t think FanGraphs uses empirical lwts (which are ideal)- as far as I know, they use estimated weights based on generalities of linear weights values over history. The conversion from wOBA to runs created helps account for the missing statistics, but it doesn’t fix everything.
BaseRuns (or BsR) is a dynamic run estimator that is widely regarded as the ideal model for team run scoring. The formula is simple:
A*(B/B+C)+D
Where A = runners on base, B = advancement factor, C = Outs, and D = HR. You can generate coefficients to fit the run environment, and this increases accuracy by quite a bit. So, “my” BsR formula for the 2009 NL is:
A = H – HR + BB + HBP + .08*SH
B = .785*1B + 2.288*2B + 3.693*3B + 2.005*HR + .006*NIBB + .125*HBP – .709*IBB + .973*SB – 1.325*CS + .856*SH – .931*GIDP – .093*K + .017*(AB – H – K + SH + SF)
C = AB – H + SF + .92*SH
D = HR
Accuracy could probably be improved upon with using multi-year data to generate the coefficients, but this is good enough for now. The team that had the largest difference between actual runs scored and estimated runs scored were the Giants, actually. Take them out and the accuracy improves to an RMSE of 12.8. They scored 29 more runs than expected. The second best were the Dodgers and Brewers, both at 18. To me, this definitely indicates a bit of luck.
Wow, link fail. My apologies. Just wiki “Root Mean Square Error” and you’ll get a nice explanation of how it works.
Thanks - I Did However Know What wRC Is.
I just misunderstood 17.2 and 14.4 as being wRCs for “RMSE” and “mine’s” (which i did not not what they were) rather then being the RMSE of wRC and BsR respecitvely. I also understand Root Mean Square but I got fooled by the E for Error and didn’t put together Root Mean Square Error.
Correct me if I am wrong here but i just don’t see a RMSE of 17.2 being significantly worse then a RMSE fo 14.4. I think my estimate technique has almost as much chance to be right as yours does.
by giantsrainman on Jan 23, 2010 3:10 AM PST up reply actions
You can test for statistical significance if you’d like. Looking at some of the differences between the two run estimates, wRC was off by 27 runs on two teams and 50 on one (the Giants). BsR was off by 22 on one and 29 on the Giants. Looking at the “range” (maximum average difference minus the minimum average difference), there’s a range of 77 for wRC and 51 for BsR. wRC has an “r” of .962 and BsR has a .973.
When it comes to modeling team runs scored, BsR is going to be more accurate than wRC. When it comes to modeling individual runs scored, however, wRC is going to be much, much better than BsR.
Whoops
That should read as “maximum difference minus minimum difference.” Guess that’s what very little sleep does to you.
Isn't This Really A Different Question?
You are testing the accuracy of the estimator to real runs scored that might or might not involve luck. What I was trying to estimate was true team talent relative to NL league average. Does it really just follow that the estimator that gets closest to real runs scored is also the one that is closest to true team talent?
by giantsrainman on Jan 23, 2010 2:24 PM PST up reply actions
One More Question?
Fangraphs FIP is scaled to MLB ERA so when you adjusted to RA instead did you also adjust to NL only as opposed to MLB in total so that it was indeed comparable to the NL only RS data you used?
by giantsrainman on Jan 23, 2010 3:20 AM PST up reply actions
If Your Estimator Is Truely Better At Estimating TrueTeamTalent Then Your 82 Win TTT Seems More Likely
My estimates first compared everything to MLB Average or MLB Replacement and then adjusted from there to NL average.
One remaining question that I have is how does the fact that the AL dominates inter-league play and thus an average NL team wins fewer games (80.25 vs 81.86 for the AL) weigh in? This says to me that you can not just assume that on average an NL team scoring the same runs as it gives up will win 81 games. This is true across MLB but may actually vary league to league because of the SSS interactions between the leagues.
by giantsrainman on Jan 23, 2010 3:10 PM PST up reply actions
Comments For This Post Are Closed
Secondary Sidebar
User Tools
Related FanPosts on McCovey Chronicles
Related FanShots on McCovey Chronicles
FanPosts
Community blog posts and discussion.
Recommended FanPosts
Recent FanPosts
In Association With
Big League Stew - MLB - Yahoo! Sports
- The Juice: Nelson Cruz matches career high eight RBIs, Pujols stays hot in Seattle
- Benches clear at Fenway as Red Sox-Rays rivalry heats up
- Keeping good company: Albert Pujols becomes fourth-youngest player to hit 450 home runs
- Giancarlo Stanton misses cutoff man, and Logan Morrison is thankful he did (Video)
- Snoop Dogg delivers memorable first pitch at U.S. Cellular Field
