Effects of Intentional Walks on Non-Intentional Walks
Intentional walks (IBB) are usually given to good and/or unprotected players in a lineup. Pitchers would rather face the next, weaker hitting batter. The IBBs lead to an inflated walk rate (BB%) for hitters. By removing IBB from a player’s BB%, a true walk rate emerges. A problem I noticed was that when a player’s IBB% increases so does their non-intentional walk rate (NIBB%). Here is an attempt at putting some numbers behind the assumption.
I have seen the phenomenon of the dual increase a number of times while looking at data. Here are three examples:
| Adrian Gonzalez | PA | IBB | NIBB | IBB% | NIBB |
| 2010 | 591 | 35 | 58 | 5.9% | 9.8% |
| 2011 | 715 | 20 | 54 | 2.8% | 7.6% |
| Robinson Cano | PA | IBB | NIBB | IBB% | NIBB |
| 2009 | 674 | 2 | 28 | 0.3% | 4.2% |
| 2010 | 696 | 14 | 43 | 2.0% | 6.2% |
| Paul Goldschmidt | PA | IBB | NIBB | IBB% | NIBB |
| 2010 (A+) | 599 | 0 | 57 | 0.0% | 9.5% |
| 2011 (AA) | 457 | 12 | 70 | 2.6% | 15.3% |
Robinson Cano matured into a good hitter in 2010. Pitchers would rather not face him and just walk him. Adrian Gonzalez on the other hand moved from San Diego, where he had no protection in the lineup, to Boston where he had plenty of protection. Both his IBB and NIBB dropped. Like Cano, Paul Goldschmidt became a more feared hitter as his IBB went from 0 to 12. Goldschmidt’s BB% historically should have gotten worse as he went from the High- to Double-A instead of improving. In each of these cases, the NIBB% moved in the same direction as the IBB%.
The two values moving hand in hand logically makes sense. If a batter is feared, pitchers will not throw anything near the zone at times and hope the hitter chases a ball out of the zone. Personally, I have seen way too many non-intentional walks where the pitcher wanted nothing to do with the hitter. While the concept makes sense, there has been no value to show how much of a change should be expected. I tried several methods to determine the amount changed, but found that most players generally don’t have a bunch of IBB. There just seemed a be a bunch of noise when looking at the data. I needed to look at only hitters who actually a decent number of IBBs and a fairly long career.
I looked at established hitters (> 9 MLB seasons) since 1980 who had the highest % of their walks being intentional. I ended up looking at 20 players (Vladimir Guerrero, Barry Bonds, Albert Pujols, George Brett, Miguel Cabrera, Mo Vaughn, Ken Griffey Jr., Eddie Murray, Sammy Sosa, Carlos Delgado, Darryl Strawberry, Will Clark, Dale Murphy, Chili Davis, Bobby Bonilla, Manny Ramirez, Todd Helton, Lance Berkman, Wally Joyner, Albert Belle). For each of the players, I ranked their seasons by IBB% and then split their careers in half by PA and breaks in the IBB rate. Finally I compared the IBB% and NIBB% of the two sets of years and determined how much NIBB% moves in tandem with the IBB%.
Of the 20 players, 14 players saw their NIBB% increase when their IBB% increased. The values ranges from 1.95% increase in NIBB% points for every 1% point increase, to a -1.41% point decrease for every 1% point decrease. The average % point increase works out to +0.50% with a median value of +0.64%. For the sake of ease — and also because there is total lack of acronyms in sabermetrics — I will call the amount of increase in NIBB from IBB “IBBump.”
There are some points to be cautious about when using this data. First, it is just a small sample of data (20 players). Few data points will always be an issue when looking at this data because not that many players have a large IBB%. Generally, only a few hitters are given free passes. The second issue is that I only looked at good hitters. Hitters in the the 8th spot in the NL get quite a few IBB in order for the pitcher to face the other team’s weak hitting pitcher. Those hitters may have a different IBBump value. I wasn’t looking to find all the in and outs of IBBump, I was mainly looking to see if there was any data that backed up my hunch. I see plenty of more work that needs to be done of the subject in the future.
Here is a example of how the use IBBump. In 2010, there was a bit of discussion on how Robinson Cano doubled his walk rate (4.2% BB% from 2005-09, 8.2% in 2010). The first key is to remove the IBB. From 2005 to ’09, he had 14 IBB in just over 3000 PAs. In 2010, he had 14 in almost 700 PAs. By removing the IBB, his NIBB rates drop to 3.8% (for 2005-09) and 6.2% (for 2010). By assuming a IBBump of 0.5 and using his IBB rates (0.5% and 2.0%), the NIBB would drop by 0.2% points for 2005-09 and 1.0% for 2010. The final values work out to be true walk rate of 3.6% for 2005-09 and 5.2% for 2010. These valuse are much closer and show that, in 2010, Cano did not really have that much improvement in plate discipline. In 2011, the lack of discipline showed when his “IBBump adjusted NIBB” dropped to 3.2%
When a hitter has a huge bump in his walk rate, a person needs to look to see if the batter’s IBB went up. Besides just removing the IBB from the data, it seems like even more of an adjustment needs to be made to get a true understanding of player’s plate discipline. More studies should be done on this subject in the future, but initially it looks like a hitter will see bump in their NIBB rate due to IBB increases.
So you’re saying that 1 + 1 + 2?
1 + 1 = 2 + 1 – Juan Pierre’s mustache
=4?
I think he’s arguing that simply removing Intentional Walks to get at a hitter’s true talent doesn’t work, because Intentional Walks have a relationship to the hitter’s true talent. The better the hitter the more likely they are to be Intentionally Walked.
If that is what the author is saying, I totally agree.
If a player is intentionally walked, that is a positive event resulting from his appearance at the plate.
To subtract it from his total BB’s commits the same fallacy as ignoring BB’s in batting average.
What if if it was a semi-intentional walk (the pitcher pitched carefully to him, resulting in a walk). Should half a BB be subtracted? Maybe 1/4 or .16 depending on the pitcher’s intent.
Let’s just give the pitcher a psychological exam after each PA to determine how much of the result was attributable to him and how much to the batter.
How about excluding walks that occur with runners on second and/or third, and first base open (or, more precisely, applying the batter’s average NIBB rate in all other situations to those situations as well — this should give a decent approximation of “intnetional unintentional walks.”)
Additionally, you could estimate how much the pitcher avoids the strike zone with secondary measures like wild pitches or HBP. When a pitcher really doesn’t want the batter to get a hit he may throw it so far outside that it goes past the catcher, or so far inside that it hits the batter.
This would help the economy. There would be a whole industry created for the analysis of how intentional a walk was.
I would love to see further research on this topic
That extra adjustment is pretty easy to figure – players with a high amount of IBB also have many “non-intentional” intentional walks where the pitcher is deliberately pitching around them, but the catcher isn’t standing up – so it just doesn’t get counted as an IBB. Watch me for enough years and this becomes apparent.
Not exactly earth-shattering, is it? If you hit well, you tend to get more walks, so the pitcher can face Reynoolds or Dunn.
You could also classify semi-intentional walks by the number of pitches thrown. A four-pitch walk is at least 50% intentional, unless you’ve got a wild thing on the mound. For each additional pitch thrown, the probability of intent goes down. Nobody wants to walk a guy when the count is 3-2, particularly if there have been a lot of foul balls.
Yes, just counting some % of 4 and 5 pitch walks would do the trick. I’m sure there is variation and more scientific ways (using pitch f/x), but I think this would work okay. I’d say count 2/3 of 4 pitch walks as intentional and 1/4 of 5 pitch walks.
I thought the article would be about whether issuing an IBB affects the likelihood of a BB following right after. I’ve often heard commentators saying that the pitcher tends to lose his focus after the IBB, and I wonder if that’s confirmed by data.
I’m not sure I am wild about removing IBB’s from the equation when figuring a players wOBA. I understand why it is done as it is unfair to give 8th whole hitters in the national league such an unfair bump. Yet we have to be careful. To say that IBB’s have nothing to do with a players skill level is disingenuous. wOBA is a relative staistic and players receive IBB’s often because of their relative value to other players.How many fewer IBB’s would have Barry Bonds received in the early 2000′s if Jeff Kent was still hitting behind him. I’m going to guess quite a few. From a sabremetric standpoint we realize how stupid most inteneional/unintentional walks are. Instead of punishing Bonds for getting less than one decent pitch to hit per at bat, I would like to know how Bonds would have performed had pitcher’s “gone right at him.” What if they had pitched him in the same manner they pitch Nick Punto. 100 bombs? maybe?
Is the author saying an increase in the walk rate of a “good hitter” does not necessarily imply an increase the hitter’s plate discipline because NIBB rate rises when IBB rate rises? Translation: We don’t know the the hitter is more disciplined. Maybe pitchers are more careful. Is that the summary?
This is a bad article.
Wow. A lot of people didn’t understand this.
Good article, Jeff. Very interesting to see.
Agreed, good article. But a bit difficult to follow, perhaps because the title is somewhat opposite the analysis. It took me several reads to understand the argument, which I would summarize as “A change in walk rate may reflect either a change in batter discipline or a change in pitchers’ approach; to tease out which it is helpful to look at changes in IBB rate.”