THE BOOK--Playing The Percentages In Baseball

On my site I looked at average speed off bat when the hitter hits the ball anywhere from 6 to 16 degrees.  You won’t have enough sample at 11 degrees, and at this point even +- 5 points gives you less than 10 hits to average for most players.

Maybe I should be clearer.

Every batted ball has a SOB, which can be broken up into hSOB and vSOB, or SOB(0 degrees), SOB(90 degrees).

Why not, instead, do SOB(x degrees), where “x” is set for each player individually , so that SOB(x) is maximized?

So, Ryan Howard might have “x” set to 25, and your average player has it set to 11, etc, etc.

So Tango your suggestion is we measure ‘how solidly a ball is hit’ based on how close the angle of the ball coming off the bat is to the angle of the swing of the bat, and infer that second angle as the average vertical angle of all balls in play for a given hitter?  That sounds reasonable, but as Rally says since we only have tens of hits per batter not possible at this point.  A by product of such a definition is that players with the least variation of ball in play vertical angle would be said to make the most solid contact.

Either way you are right that the speed of the ball in the 11° plane does predict run value of a ball in play even better than the speed of the ball in the 0° plane.  I get an r^2 value of 0.151 for that relationship.  That is probably the speed I should have used for my last graph.

Since Ryan Howard is so topical, what speed off the bat do you get based on different angles?  That is, for every batted ball, what is the 5degree speed off the bat?  10degree?  15? 20? 25? 30? 35?

Given what we know about Howard and HR hitters, then we’d expect to see his best showing at around the 25 degree mark, if this idea for getting the hardness hit is valid.

Using:

SOB = Speed Off Bat
HLA = Horizontal Launch Angle
VLA = Vertical Launch Angle

unless/until the consensus changes the terms.

Now, my point:

I can’t see where hSOB is headed. 

Tell me a player’s SOB is increasing, that means he’s hitting the ball harder, which makes homers more likely, etc.  Maybe he’s stronger now, or swinging better, whatever.

Tell me a player’s VLA is changing, that means he’s hitting the ball higher or more on a line.  Maybe he’s changed his hitting approach to get more loft, or less loft, etc.

Tell me his HLA spread is changing, that means he’s (intentionally or not) hitting the ball to a different part of the field.  Maybe he’s trying to take advantage of a short porch.  Maybe his bat is slowing down.

All of the above subject to SSS caveats, of course.  Not suggesting we interpret noise, only signal.

Here’s what I don’t get:  tell me a player’s hSOB is changing, I don’t know what that means.  If hSOB is increasing, is he hitting more line drives, or hitting his fly balls harder, or reducing the number of popups, or what?  The only way to know is to go right back to the original three parameters.

So, why do I need hSOB? 

I’m willing to be convinced…

A batter’s swing plane changes as the bat is moving in its arc through the hitting zone during a swing.  So even if he hit the ball squarely (eith zero bat-ball offset) each time, his vertical angle would vary with where in the swing path he contacted the ball.  The best measure of what the swing plane angle is at the precise moment of contact is, as Tango mentions above, the angle that has the highest average SOB for a particular batter.  But that won’t equal the optimum vertical angle (the angle that yields the highest linear weight).  That’s because the batter can also alter the vertical angle of the hit ball by adjusting the amount of undercut (bat center below ball center at contact).  So if Ryan Howard wants to hit a ball for a maximum chance of a home run, his swing angle will be around 15 to 18 degrees on a pulled ball with enough undercut to cause the hit ball to have a vertical angle of 25 degrees.

The same is true for non HR hitters.  They have a lower (by about 5 MPH) maximum SOB so they have swing planes that are flatter than Howard’s to maximize the chance that a BIP will be a hit.  Their swing plane angle on a pulled ball will be around the 11 degree angle that optimizes linear weight for BIP.  But they also have another peak at a lower swing angle of around 0 degrees for balls that they hit to the opposite field.

A single aggregate average of vertical angle for most players doesn’t yield much useful information because most players are not pure HR hitters like Howard or pure non HR hitters.  Instead they used a mixed strategy depending on where the ball is pitched.  An aggregate vertical angle based on the SOB of all hit balls or all hit ball hits masks this and gives an intermediate value that doesn’t represent what the batter is trying to accomplish.

I think people might not be correctly interpreting Tango’s suggestion.  I thought he was suggesting that hSOB be redefined for each player to be the projection of the velocity vector onto an axis making angle x with the horizontal plane.  Specifically, if the launch angle is z, then hSOB=SOB*cos(z-x).  The current definition has x=0.  Do I have this right?  If so, then the issue is not the speed off the bat for balls hit at launch angle x (leading to SSS).  Instead the issue is the hSOB as I have defined it, which does not have SSS.

When I introduced the concept of vXY at last years Pf/x Summit I was not intending it to be an indicator of speed at some “ideal” launch angle.  First of all the concept of an “ideal” launch angle is flawed.  As I mentioned in Post #6, the vertical angle that leads to the highest LW depends on what kind of power the hitter has, where the pitch is thrown, and whether the batter is trying to hit the ball to the opposite field or not.  If the goal is to create a metric that measures Defense Independent Batting, and I think it should be, then we should not expect zXY (hSOB) at 0 degrees or any other angle to be a solution to that goal.  vXY was only intended to be a measure of how quickly the ball got to the fielder, therefore 0 degrees is the only measure it should have; horizontal means 0 degrees not 11 degrees.  If you want to find the combination of factors that leads to the highest expected linear weight of a hit ball in order to measure a batters DIB, we will be able to do that when we have enough Hit f/x data available to avoid small sample issues. But it will be a set of linear weights based on a ball’s SOB, some measure of its vertical angle, and whatever control over its horizontal angle that analysts determine a batter has, at the very least whether it was pulled or hit to the opposite field.  It will not be a single value like hSOB at some “ideal” launch angle.

Perhaps in the above post I should have said Luck Independent Batting, instead of Defense Independent Batting, since what we are seeking is a measure of a batter’s skill that minimizes the factor of defense, but also the factors of park and the sample size vagaries of whether the hit ball actually became a hit, and instead measures whether a hit ball SHOULD have become a hit and how many runs it SHOULD have produced in an average BaseOut situation for an average team.

So much of this discussion was covered in the Ortiz thread already, but we’re covering the same ground again in two different threads (this one and the linear weights on batted balls one).  I wish everyone would go back and read the Ortiz thread.

From Matt:

Peter:

I agree with you reasoning about the importance of HSoB, but you can’t take it too far. Say you had two hitters who had the same SoB and one averaged ASoB of 0 deg and the other 20 deg. The 20 deg hitter would be far more valuable than the 0 deg hitter even though his average HSoB would be about 6% lower (assuming I’m doing the calc correctly). He would hit for more power, and I would expect also a higher BABIP.

The reason being that once the ball hits the ground it loses a lot of velocity. Plus the low balls can be fielded by the infielder for one-hoppers or line drive outs. Line drives have the best BABIP and hit value and sorry to state the obvious. Although, straight up HSoB should be decently correlated to BABIP - so it’s certainly a reasonable approximation.

I would bet there’s a maximally efficient launch angle (~15 deg I would guess) and that you could predict BABIP based on the average SoB in that plane.

That’s where the idea came from, and then I determined that around 11 degrees was the launch angle that optimized batting average.

Who cares if 11 degrees is not the perfect plane for every batter?  It’s going to be close enough to work for every batter.  Even if it’s 15 degrees off, the cosine of 15 degrees is 0.97.  An error of 3% is trivial, and it’s better than using either vXYZ or vXY.  We don’t need batter profiles in order to determine optimum launch angles to measure SOB in that plane.  We can use SOB.11 for everyone and it will be good enough.

Btw, Peter, please don’t use hSOB to describe SOB in the 11-degree plane.  That will just lead to confusion.