According to STATS … eight of the top 10 players in slugging since the start of the 2007 season have hit more home runs on the first pitch than on any other count. In fact, the first pitch has been the favorite home-run scenario among all major league hitters during this span. A look at the home-run counts since 2007 reveals most players are likely to connect early in an at-bat.
Most Home Runs by Count, MLB, 2007-2010
Count HR Pct
0-0 2904 17.6
1-0 1995 12.1
1-1 1918 11.6
3-2 1616 9.8
0-1 1588 9.6Most homers apparently happen on the first or second pitch of the at-bat, either 0-0, 1-0, or 0-1. Or when the pitcher absolutely has to get it in the strike zone, at 3-2. Plus, another reason is that any particular at-bat can have only one of count, except for 3-2, 2-2, and 1-2, where the number of pitches (and thus HRs) are elevated. The only exception was 1-1, which is kind of odd since technically the pitcher has a pitch to play with there, and thus could be throwing a pitch to try to get the hitter to chase it.
The least:
As for the counts in wich the fewest home runs are hit, 3-0 and 0-2 situations top the list. No surprise there, but what stands out is that 2-0 and 3-1 counts, the so-called hitter’s counts, generate so few longballs.
Fewest Home Runs by Count, MLB, 2007-2010Giants Thoughts
Count HR Pct
3-0 68 0.4
0-2 539 3.3
3-1 845 5.1
2-0 947 5.7
So perhaps that accounts for Pablo's power shortage this season, he's been trying to work deeper into counts and missing the pitches he used to blast out for homers. Meanwhile, he's getting into count situations where hitters normally are unable to do anything more with the pitch and not get the homer.
FREE PANDA NOW!
Big Hole In This Theory.
ReplyDeleteThis only works when the pitcher has a reason to throw a strike on the first pitch and with Pablo right now the pitcher has no reason to do so. He can get ahead of the count with the Panda more often then not by getting him to chase a ball and foul it off if he doesn't hit it weakly fair for an out.
My theory is that Pablo is being messed up mechanically because of all the "help" he's been getting from people this season.
ReplyDeleteLast season, "unhelped", on the first pitch, he hit .357/.364/.626/.990 with 7 HR in 115 AB.
Also, interestingly, he did horribly once there was two strikes on him. By working deeper into counts this season, that increases the odds of him reaching two strikes. I wonder if the Giants messed with the golden geese by trying to get more out of him. Maybe it will pay off eventually, but right now, it's messing with our playoff hopes.
Your analysis on my site was amazing. Thank you for commenting and sharing your knowledge. I would love to have do a future Q & A with you on the site if you would be interested.
ReplyDeleteThis is a classic case of statistics not telling you the real story. Sandoval is not hitting right now because he is being impatient at the plate. Yes, he is seeing more pitches, but that is not making him a more patient hitter. His problems is that he is not swinging at strikes, especially when men are on base. A patient hitter is one that does not go out of the strike zone or swing at pitches that he cannot handle whether it's the first or last pitch of the at-bat. A perfect example of this was Barry Bonds. He rarely got strikes, but when he did, he didn't miss them too often. Sandoval is the polar opposite of Bonds. He doesn't appear to have a plan at the plate. He is hacking away, especially at pitches that he is not going to do much with, such as the letter-high fastball and above. Bottom line is that it really doesn't have anything to so with what number pitch of the at bat that you hit.
ReplyDeleteThere is a very serious statistical flaw in your analysis. Namely that every single AB has a pitch delivered with an 0-0 count, but many ABs never get to a 2-2 or a 3-2 count because another outcome is obtained first. In order for your analysis to be meaningful, you need to adjust your numbers so that you give the number of HRs in each count weighted by the number of times each count occurs relative to the other counts. For a theoretical example, 25% of all pitches thrown might be in a 0-0 count, 20% of all pitches in a 1-0 count, 15% in an 0-1, 10% in an 2-0, 10% in a 0-2, 10% in an 3-0, 10% in a 2-2 and 10% in a 3-2. In this theretical example the 0-0 pitch is 2.5 times more likely to obtain an HR result than a 2-2 pitch simply because there are more 0-0 pitches delivered, not because it is harder to hit a HR on a 2-2 pitch. Your suggestion that 0-0 pitch is optimal for HRs may indeed be valid, but you have to re-analyze in order to find out.
ReplyDeleteHere's a link where if you scroll a bit down the page, you can see an inkling of how often a count occurs. In the sample data shown in a table, there are 3268 0-0 counts; 1413 1-0 counts; and 1388 0-1 counts. That makes sense because some of the 0-0 balls were hit; some were balls; and some were strikes.
ReplyDeleteThe link:
ReplyDeletehttp://research.sabr.org/journals/study-of-the-count-yields-fascinating-data
Thanks for the comment Anon.
ReplyDeleteUnfortunately, you mis-read my post. The analysis is not mine, actually.
It is the analysis by STATS, and I will admit that I didn't think any deeper about what they wrote as they are a very well known organization that produces a lot of statistics. I just assumed that they got it right.
What you say makes a lot of sense, so thank you for pointing that out. As you can see in my post, I provided a link to their PDF, and that provides a couple of ways you can notify them of their error. You can either contact them via their website or you can tweet to them on their Twitter account to let them know about their error. I am sure they would appreciate it if you brought up such an egregious error to them.
OK, in any case, looking over Pablo's 2009 pitch count data, he at least swings for more HR power: http://www.baseball-reference.com/players/split.cgi?n1=sandopa01&year=2009&t=b#count
His judgement on the first pitch was pretty good when he connected. He hit .357/.364/.626/.990 when he connected on the first pitch, 7 HR in 115 AB or 16 AB/HR. His batting line for the season was .330/.387/.556/.943, 25 HR in 572 AB, or 23 AB/HR. So Pablo, in any case, swung for more HR power in swinging on the first pitch during his ABs.
The expected held true, when he was ahead in the count, he hit better much better than when he was behind. And he had more 0-1 counts than 1-0 counts, so per the comment by Anon, not all is great when he swings, because when he misses, he gets behind on the count and this is his batting line after 0-1 count: .286/.318/.464/.783, with 6 HR in 252 AB. Which is still pretty good for most 3B, but not Panda-riffic, like we have grown to expect.
Unfortunately, I cannot tell whether that first strike was swinging or taken, else I could give an opinion on his decision making regarding swinging. But when he connects, obviously he made the right decision.
Obsessive,
ReplyDeleteActually the analysis by STATS is correct, it is just your application of it to Pablo that is faulty. They are giving the correct raw numbers, you are taking the raw numbers and drawing conclusions without properly weighting the raw numbers in the fashion which I specified above. At least I am pretty sure I am correct here, but I am not an expert on stats.
And just to clarify, your secondary analysis in post #8 is not problematic as far as I know, just your analysis in the main post. Thanks.
ReplyDeletePer Boof's point, again, I think we are probably arguing a similar point from different points of view.
ReplyDeleteIn 2009, he swung at 118 first pitches and hit a ball into play or a HR. 248 times he took a ball 1 and 267 times he had a strike 1, either swinging or taking. 18.6%, 39.2%, and 42.2%. On first pitch, he hit .357/.364/.626/.990.
In 2010, he has swung at 50 first pitches and hit a ball into play or a HR. 96 times he took a ball 1 and 125 times he took a strike 1. 18.5%, 35.4%, and 46.1%. On first pitch, he's hitting .396/.400/.688/1.088.
So his judgement on swinging still appears to be good, at least in terms of when he connects.
The problem is that now he is ending up down in the count now when he does not connect, he is in more 0-1 counts, and since there are a lot more of those than no pitch counts, his overall numbers are pretty bad relative to last season.
Now that he's struggling, people are coming out and saying that he's impatient, don't take pitches. Well, he's been like this in the majors since August 2008, and this implies that it took the MLB one and a half seasons to figure Pablo out.
Does that make sense? Not to me. I think the struggles he has been going through have been self manufactured, and that once he gets over that and all the advice that has been coming his way, he'll be back to his Panda-self. But we will see.
Well, I wouldn't call myself an expert in stats but I consider myself pretty good.
ReplyDeleteI will have to disagree with your latest comment. STATS did exactly what you said. I took their numbers and calculated the total to derive each percentage using the homers in the list and it comes out to roughly 16,500 homers for each. If they had analyzed it per what you said, each count should have a different result. So maybe what they say is the right statement, but the data they provide don't show that.
As far as the application to Sandoval, I think your statement applies to my statement because I relied on STATS data, where you are correct, they did not account for the frequency for each count, at least not with the stats they provided. My faulty logic was relying on STATS statement without checking out the numbers they presented.
I think Boof has it right - Panda is in a catch-22, the point of seeing more pitches is that you have a better change of getting one to hit. The flip side is if you don't swing at that one, you get in a hole - at which point Panda goes to hacking. Seeing more pitches is a good thing, but if the pitcher knows you're taking, what do you think you're going to get?
ReplyDeleteI saw a study somewhere recently that showed that pitches taken is not necessarily related to performance - plenty of caveats (you can't walk on 3 pitches, you can't strike out on 2, it's probably skewed by bad hitters) - but on the surface, there is barely a correlation. It reminds me of BP's analysis of clutch-hitting of a few years ago - while overthought in their usual manner, it ultimately showed that if such a thing exists, it exists in intelligent players who... "have a plan at the plate".
I'm not suggesting that in the long run Panda should be either/or, but in the cat-and-mouse game of hitting vs pitching, a hitter should be just as likely to swing at the first pitch as the last - if not, he's giving a lot away to the pitcher.