HS Batting order - sabermetrics

Great thread cabbage,

When my son was in HS the coach used the traditional approach in putting the best hitter in the 3 hole, power 4 hole, etc. I have the stats from the three years my son played and one year we had a bunch of injuries so the stats don't make sense, but in the other two the number of AB's for the top 4 hitters were:

Yr 1

102

97

96

96

Yr 2

108

106

98

96

In year one they were pretty close (within 5 AB's)  and  in Yr 2  the top hitters got 10 more AB's which would statistically make a difference I would think. 

I saw teams put their best hitter in the number 1 hole a a few times when the player had good running skills.

I think it would be dependent on whether you had a high OB% player with speed on the team, but if it was me I would put him in the 1 hole and my best hitter in the 2 hole (if he had decent speed and base running skills) and go from there. For example I saw Christian Yelich bat in the 1 hole in HS, who was clearly the best hitter (for power and OB%) on his team but he could also run. 

http://www.baseballmusings.com...in/LineupAnalysis.py

It's not perfect, and it's based on MLB assumptions about run envrionment/distribution of offensive events, but it's probably a decent place to start.

If I had to guess, I'd expect the range of OBPs in a HS lineup to be wider than in MLB DH leagues, and I'd expect the average OBP (or effective OBP if you want to count ROE) to be higher in HS as well.  I'd also expect SLG to be lower in both range and on average.

The higher the run environment, the more important it is just not to make outs, and that will be exacerbated by a lower SLG/ISO. Given that, just going in reverse order of OBP might be as effective as anything else. Knowing what a kid's actual numbers are is going to be the actual problem.  Sample sizes in HS are generally miniscule, so you're often effectively guessing at who the actual best hitters are, and by how much. On the plus side, lineup order doesn't matter that much over a 162 game season, so it's probably not going to make much difference at all over 20-30 games, unless you're actively trying to pick a poor lineup.

Great thread.  At the risk if digging up any old issues, I'd start by repeating what many have said:  that unfortunately the sample size of data for a particular team will limit what conclusions you can draw and act upon, so in the end the best tools a HS coach has for putting together his lineup will probably remain his two eyes, and what they tell him in practice and in games. 

But even if you accept that 100% it's interesting to ask what about data that is wider than a given team?  What about statistics that encompass not just a team but perhaps a league, section, state or even the country? If you could show that teams where the OPS for the player in the 2 hole was higher than teams where the number 3 hitter had the higher OPS, would that mean that your best hitter should bat 2?

Maybe so, and Stat's most recent thread has got me thinking that those kind of wider statistics can be interesting, especially in terms of looking at success rates  for bunting, hit and run plays, squeeze plays, etc., or the importance of walks, or strikeouts, etc.   And they're certainly good for examining the nature of the game as it's played in HS -- IE, are the balls and bats so dead in the current era that HR's are barely a part if the game anymore?

Speaking of the nature of the game, if he's batting #2 your best hitter better also be your best bunter, or right up there.  Like it or not, as we all know bunting is HUGE in HS baseball. Around here every single team does it, and a #2 hitter, if the #1 hitter is any good at all, will be seeing a bunt sign in the first inning at least a third of the time. 

The number 2 hitter on my kid's team last fall was in many ways the ideal #2 hitter in both old school and new school terms.  He had twice as many HBP as the next closest player, and lead the team in sac bunts and ROE.  But he also lead the team in OPS, SLG, and OBP, with the #3 hitter close behind in all three.  I'm sure that the coach built his lineups on what he saw -- in fact he told me that he doesn't even calculate stats until the season is over.  That particular kid validated his decision nicely.

It was my experience that the best hitter was nearly always the best bunter, if it matters, however as JH pointed out that bunting is bad strategy IF errors rates are not taken into account. I am not sure if this holds true with a higher error rate seen on some HS teams. I still remember this game where my son's team lost 5-0 and the opposing team never hit the ball out of the infield. (they were speedy and bunted a lot and we through the ball around)

Anyway back to the thread I remember seeing Yelich in the 1 hole and thought Hmmm good idea. (after the game of course)

No one should ever lead their team in SAC (for any meaningful number of SACs) and OPS/OBP/SLG. A big part of the reason to SAC, when it's even a good idea, is that your hitter isn't good enough that foregoing the opportunity cost of taking the bat out of his hands is that big a deal.

A simple approach is calculate how many base runners your team averages per game. If it's ten there will be thirty-one plate appearances minus double plays your team averages per game and how many base runners you average getting thrown out per game. So if your team averages ten base runners, a double play and a runner thrown out per game I suggest batting your best hitter second to make sure his spot in the order comes around that extra time. This calculation is on the fly. There may be other factors. But it sets the example.

Jacjac - I hear you, I really do, and I am not going to say you're wrong in any way, but rightly or wrongly that's just the way the game is played.  Everybody in our league bunts, and everybody in our section bunts. My kid's team's coach will bunt anybody in the the line up. I do think he does a good job of mixing it up. He'll have the kid show bunt and slash with no strikes, then don't show but do bunt with one strike etc.  And it works for him. He won 22 games in a row this year, including 5 or 6 by one run.  Also, in HS, BOF's point is a good one -- bunts put a lot of pressure on the defense and they can turn into ROEs.  Like I said, the kid I mentioned led his team in both SAC and ROE.  I don't think that's a coincidence. 

cabbagedad,

I really don’t know what one of my “own agendas” is, so I suppose this is gonna make everyone mad just because I posted, then ruin the thread.  But, that’s your problem not mine, so I’m gonna post and let whatever happens, happen.

IMHO, you’ve begun the thread by failing to define something very important because it changes from coach to coach. Unless you define “best hitter”, anyone else trying to duplicate what you come up with is gonna be hard pressed unless they happen to use the same definition you do. I’m not talking about some long algorithm, but rather whatever you define as the criteria for best hitter.

In the class I’m taking, the “best” hitting metric is defined as the one that best predicts runs, with the order of the 5 most common metrics being this. BA worst, OBP next best, SLGP next best, OPS next best, and RC being the best. Those are all pretty simple metrics, and personally, I don’t care which is chosen, but you should start someplace in order to get everyone on the same page. But just saying “best hitter” gets things started poorly.

Again, this is probably gonna be taken me going off on one of my “own agendas”, but I have to risk it.

You said: My question is... how would you apply some of the new thoughts on batting order to the typical HS game where the games are only 7 innings and there tends to be a gap in ability between the top six or seven and bottom two or three hitters?

I don’t understand what difference it makes how many innings are played? Are you saying in MLB there’s no gap in ability from the top 6 or 7 to the bottom 2 or 3, or that the gap is much less than in HS?

cabbagedad- Here ya go  http://www.beyondtheboxscore.c...izing-your-lineup-by

Let me know if I can help in any other way. Lineup optimization isn't a hugely impactful aspect of the game, but I find it to be an interesting study nonetheless. I think you may find the stats WPA and RE24 interesting also: 

http://www.fangraphs.com/blogs/get-to-know-wpa/

http://www.fangraphs.com/blogs/get-to-know-re24/

Stats, I think the innings is important. in a 9 inning game, with perfect game every batter gets up 3 times. In a 7 inning game only first 3 batters do. Simply put (cause that is how my mind works). I would rather have my best hitters up more often. In less inning even more important, I would think  

JH,

When I saw that article the last time the link was posted, I wondered how long it would be before someone started snooping around to see if it could be applied to any other levels, the way I see if other MLB metrics can be applied to other levels. When I read it the 1^st time, I saw something that intrigued me, so I looked at my data to see how we compared.

The article said: The Book says OBP is king. The lead-off hitter comes to bat only 36% of the time with a runner on base, versus 44% of the time for the next lowest spot in the lineup, so why waste homeruns?

So I ran one of my reports by BPOS to see how our numbers related, and I have to admit I was somewhat surprised. While our numbers also show the #1 BPOS as having the lowest percentage of PA with runners on, ours was 10% higher. Where the MLB numbers for the percentage for the #1 spot were 8% lower, ours were about 10% lower. Certainly not a direct match, but very much alike. I can only guess that in HS there are typically more runners that reach base, which is where the difference is.

My question is this. Is there anyone else out there who can verify their HS team/league or whatever has similar numbers. The reason I ask is, while I’d expect the numbers for about anyu level or place to be the same because the leadoff hitter always comes up at least once each game with no one on, unless the numbers are much the same as far as the percentages go, if they try to apply what the articles says about the book for the leadoff hitter, the whole paradigm would fail.

Using the article, if I were to try to let the computer compute who should be the leadoff hitter, I’d compute OPB, SlgP and for lack of a better measure of speed, SBPct for all the players, pick some minimum number of PAs to use, choose the top 3 players using OBP, then choose the lowest SlgP among them being careful to not choose a poor SBPct.

Sound correct?

I coach in the largest class in Missouri.  We have been successful, largely due to our pitching.  About 3 years ago, I switched to doing this with our lineup:

We basically hit our hitters from best-to-worst down through 6 hole, and then bat the next best guy 9th.  The two worst hitters hit 7/8.  If my best hitter is LH, I usually hit him 2nd, so that we can stay out of DP's.  Otherwise, that has worked well for us.  Here is why:

You get 21 outs, so there are only a few guys guaranteed 3 at-bats.  I want to make sure in a 1-0 or 2-1 type game (which most "big" games are) that our best hitters get the most opportunities.  If the game is a 5-4 type game, the top gets 4 at-bats.

I also like that we have a chance to get to a great pitcher in the first inning and can get pressure on the pitcher right off the bat.  It also helps you in those later innings when you have the 7-8-9 guys up, and you know you just have to get 1 guy on and you have a chance for a 2 run HR.

I did it the other way for a few years, but when we have been very successful, this is how I have done my lineup (our success is more because we have had great players).

Thanks JH, that's the one!  After re-reading this and taking a look at the WPA and re24 info, I find it interesting that both the article and yourself state that lineup optimization is not hugely impactful.  In fact, the article states... "Believe it or not, the difference between an optimized lineup and a typical, mildly foolish one you'll see MLB teams use is only about one win over 162 games."  That just blows me away.  The WPA example is Ellsbury leading off with a double and impacting the win probability.  Logic would tell me a better hitter will be more likely to succeed at this and thus give a team a better chance at winning if that player hits more often and in the right situation.  I would expect the potential difference in games won to be far more than 1 in 162.

As it relates to HS, I think the most glaring difference is the lesser number of outs/innings.  So, as some have noted here, getting the better hitters at the front of the lineup becomes more important due to the difference in accumulated number of AB's both per game and per season.

I think there are some things to be learned from the article that could cross over to HS but there is one point I question.  Regarding the #3 Spot, it says...

"The Book says the #3 hitter comes to the plate with, on average, fewer runners on base than the #4 or #5 hitters. So why focus on putting a guy who can knock in runs in the #3 spot, when the two spots after him can benefit from it more? Surprisingly, because he comes to bat so often with two outs and no runners on base, the #3 hitter isn't nearly as important as we think."

Yet is says your #1 guy should be your highest OBP guy and #2 should be one of your best three hitters, if not the best.  So, how can that lead to #3 coming to bat so often with two outs and no one on?  Perhaps this historical data does not properly represent what would be the case if the recommended changes to #1 and #2 were made and therefore inaccurate?

Realteamcoach hit the nail on the head. I do have a question however.

What do you do with a player who can swing it but is a slug? 

Yes, that is certainly the way I figure it.  For example, in HS, the top half of the lineup is likely to get 3-4 AB's per game and the bottom will get 2-3.  Whereas in MLB, the top half is likely to get 4-5 and the bottom 3-4.  So, the difference is higher from a % increase in AB's when comparing the top of the order and bottom.  Pitching and defense are usually decent in our league so I don't believe the number of baserunners per inning is that much higher to negate the 22% increase in innings per game.  I think in the small school leagues this gets thrown out the window.

Regarding the gap in ability, it crossed my mind as I typed that there is a gap for both but I didn't want to go overboard with a wordy OP and I do think there is a more significant gap for the average HS.

If your top 2 have OBPs of .400, the 3 hole hitter will bat with bases empty and 2 out in the first 36% of the time.  If they have OBPs of .500 it drops to 25%, but at a more realistic .350, it goes up to 42.25%.

i believe The Book suggests a relatively low OBP guy with some HR pop for the 3 hole, all things being equal. 2013 Will Venable, for instance. When he's up with the bases empty, he can still do damage, but it's not a huge deal when he makes the third out and your cleanup hitter leads off the second.

The objective is to win the game. If we are playing the Japanese, Korea, Australia National Teams or in our Northern California American Legion in the 80's.

We set our lineups based on several factors.

1. Does the opposing manager have any LHP? If not, I stack my left handed hitters one thru 5. If they do I alternate right and left. On base % is important for #1 and 2 and #9. A left handed pull hitter for #3. I need runner to 3b quickly. There are 11 ways to score from 3b with less than 2 outs.

2. RBI's each time at bat is important to determine #4,#5 and #6 hitters.

3. #9 hitter is switched with #1 hitter every 4th game to realize the importance on "on base" mentality.

4. Opposing managers will pitch your lineup with "breaking" balls to your #3, 4 5 hitters.

With this in mind it is good to switch the hitters in the line up.

Bob

GOODWILL SERIES INC

Originally Posted by cabbagedad:

Stats, I think the innings is important. in a 9 inning game, with perfect game every batter gets up 3 times. In a 7 inning game only first 3 batters do. Simply put (cause that is how my mind works). I would rather have my best hitters up more often. In less inning even more important, I would think  

Yes, that is certainly the way I figure it.  For example, in HS, the top half of the lineup is likely to get 3-4 AB's per game and the bottom will get 2-3.  Whereas in MLB, the top half is likely to get 4-5 and the bottom 3-4.  So, the difference is higher from a % increase in AB's when comparing the top of the order and bottom.  Pitching and defense are usually decent in our league so I don't believe the number of baserunners per inning is that much higher to negate the 22% increase in innings per game.  I think in the small school leagues this gets thrown out the window.

As I said, I sure don’t disagree that having better hitter getting more PAs is a good thing. Where I run into problems, is all the assumptions, never once looking to see if the assumptions are true, and I have no reason to believe they aren’t. Its just that my mind likes to work with facts, or proof if you will. What I’m talking about is your BELIEF that the pitching and defense in you league is any better than any other league. And again, I’m not saying they aren’t, but I’ve found out over the years that perceptions often don’t coincide with facts.

I know our team is far superior to its opponents in the games against each other as far as pitching and defense goes, but I’ve never tried to compare teams for all the other games. Another thing is, while more than half of our games every season are league games, we still play a lot of non-league games, and to try to calculate how every team’s league does relative to ours would take on heck of a lot of time. My point is, its really difficult to get good numbers for this kind of exercise because not many coaches both track the things necessary to compute the numbers, and store them in such a way that they can be readily accessed.

In our league its mandatory that all coaches use MaxPreps, so I could do some semblance of comparison, but even then, that would take a lot work to keep up with during a season. That’s a big reason I’m so against trying to use the numbers the way they do in the ML. They have the good fortune to have the good numbers and history to fall back on, where for HS, there’s a lot of guessing and perceiving. Many things would work ok, but no one really knows which wouldn’t, and that causes problems.

Regarding the gap in ability, it crossed my mind as I typed that there is a gap for both but I didn't want to go overboard with a wordy OP and I do think there is a more significant gap for the average HS.

For a few years, the guy I score for used to express things like that, so I thought I’d do him a favor and actually check it because I had the capacity to do it. I generated a bunch of different metrics by BPOS, and was really surprised at some of the thing that came out. To me it was just information, nothing else, so I sent the reports to him. He got  mad at me because he thought I was jacking the data around to try to make him look bad. It took me a lot of s’plainin’ to prove to him that all I did was run the numbers that were there. In the end, we had some “spirited” discussions about it, and he refuses to change what he does, even though he admits what he’s always thought was not true.

I think what happens a lot is, a HS coach doesn’t have the luxury of having a huge crew of assistants, great facilities, a 2 month long spring training, then 162 games over 6 months to tweak lineups based on years and years of data. Everything has to be accelerated, and every decision has immediate ramifications. It just a different world.

I had to deal with this during the season.  Our cleanup guy was a DH with a pulled hamstring.  He had our 4th highest OBP and 3rd highest slugging.  I hit him cleanup like normal.

In Missouri, you can re-enter one time for the starter.  So if I projected the game to be low-scoring, I would run for him the 1st chance that I got.  If he was the lead runner in the 1st inning, I ran for him in those situations.  I then would re-enter him for his next at-bat.  If he got on again, I did have some decisions to make, but I figured if he got on base twice in a low-scoring game, and I ran for him both times, we were probably going to win that game.

It is hard to get on base 2 of your first 3 at-bats against a guy in the upper 80s and then also have another at-bat in that same game.

Where you get hurt with this type of guy is when he bats with less than 2 outs and a runner on 1B...DP candidate.  And I "had" to bat him behind 2 other guys who did not run very well.

Fortunately he was a senior that had learned to elevate the ball and we were not hurt any by his injury/lack of speed.

realteamcoach,

good insight.  I'd like to hear what your take is on the "beyond the box score - optimizing your lineup" piece that is linked in JH's post.

cabbagedad- In my opinion, where you would run into differences pertaining to lineup optimization at the HS level vs. the professional level doesn't have much to do with the talent difference between the top few players and the bottom few players. Yes, there is a difference, but I like to think of it like this...

Let's say, for hypothetical sake, that there are 8 productive hitters in the lineup for a big league team. Let's say four of the hitters produce 60% of the offense, while the other four produce 40%. The discrepancy in high school level talent may shift the difference to something like 70%-30%…different, but still =100% in totality. The difference in run scoring environment on a game-by-game basis, in my opinion, probably doesn't warrant enough of a drastic change in philosophy stricken from the research done on MLB games.

(note: this is a hypothesis. If anyone finds a way to run the numbers on run scoring environments and calculate WPA and RE24's for a high school team, I'd welcome the data with open arms)

The biggest difference I think you'll find between high school and MLB is the Isolated Power production per player. Isolated Power - shortened to "ISO" on websites like FanGraphs and Baseball Prospectus - is simply "Slugging Percentage minus Batting Average." ISO is a measure of how many of a player's hits go for extra bases, and how often. For reference, the current MLB leader in ISO is Jose Abreu, at .340. The highest single-season ISO in MLB history was Barry Bonds's .536 in 2001.

At the high school level, it is very difficult to find a collection of players capable of putting up ISO numbers in the same vicinity as MLBers, simply because the power production is not as prevalent. Due to this factor, run expectancies in given game situations are almost certainly different than those in the Major Leagues. Unfortunately, I can't say how different without research on the subject.

As for lineup optimization - as I alluded to, it's not that drastic of an impact game-by-game. However, I'd imagine that the run scoring environment in high school is not that different from the run scoring environment in MLB to the point where a team finds itself necessitating a hitter in a high level situation different from that in MLB. I'd be very confident betting that the method of lineup optimization outlined in the Beyond The Box Score piece (us nerds stylize that as BTBS, by the way ) is most likely accurate for the high school level as well.

This is the source of all the information in that article. I highly recommend it for those interested in learning more about sabermetrics: http://www.amazon.com/The-Book...seball/dp/1597971294 

Cabbage,

I think there is a lot of validity to the Beyond the Boxscore article.

One of the tough things about stats is that there is such a small amount of data for HS baseball.  I worked on DIPS for a D1 university about a decade ago, and it is just a lower level of baseball with too much difference in ability to tell us much.  Therefore, HS, College and pro baseball have to be viewed almost as 3 entirely different games.  The 2 things I have learned in studying HS stats over the past few years:

1. RBI are a much more telling indicator of a good offensive player/predicting future success at the HS level than at the pro level.

2. Bunting is a much bigger weapon in HS baseball due to errors.  Sacrifice bunting is still (largely) a waste, but bunting for hits makes a tremendous difference for offense.

I am looking at translations of JV-Varsity stats, and my main offensive hypothesis is that strikeout rate for hitters is a very telling indicator of varsity success.  So basically if a JV guy strikes out a high amount, even if his other stats are great, that guy will not project as well at the Varsity level as you would imagine.  My summer season ends in 2 weeks, and I will then be working on my data.

Realteamcoach and JH, 

I completely forgot about the re-entry rule in HS which would skew data collected at the MLB level. At least at my son's HS the top 4-5 hitters were head and shoulders above everyone else so looking back it really was not that difficult to put an effective line up together, put the top 4-5 at the top with more speed and higher OBP at the beginning and if necessary run for one of your power guys. Like Bob indicated also put a high OBP guy with some speed in the 9 hole to stack up the top of the line up to squeeze out some runs every couple of innings. It is probably more about in game management than anything else in reality if you have a couple of kids who can hit but are not too good defensively. 

Good thread. 

Re-entry changes everything.  3 years ago we had 3 2B that were good defensively but struggled to hit.  Our #2 SP was a PO, so we had to let the 2B spot "hit".  I batted that guy 8th and we pinch hit everytime the spot came to the plate.  We had 5 OF that could play that year, so we used a couple as pinch hitters and then would enter the next 2B.  Our best guy usually started the game and finished the game at 2B, but finished the season with 8 AB in 33 games.

Originally Posted by cabbagedad:

Stats agenda #1 - Coach perceptions are usually wrong as compared with data facts.  Check.

Agenda? You’re saying I have agenda saying coaches are usually wrong? Evidently, you don’t pay much attention to what I write. I’ve consistently said everyone’s perception when compared to the facts are usually wrong, including mine!

Stats agenda #2 - I can show coaches the error of their ways if they'd only listen but they don't understand stats and/or are too stubborn and old fashioned to see the light.  Check.

Holy cow! You completely missed the point I was trying to make! I thought All I did was look at some REAL numbers to see if what was generally accepted as true, really was. I said I was surprised at the results, so I’m saying right there I was definitely wrong in my perceptions, which went pretty much along with his. Then all I did was show him the results. I didn’t even include a note with it. I just sent the reports to him. He didn’t change what he’d been doing, and I didn’t make a judgment about it. There are reasons he didn’t want to change that I didn’t go into, but you made the assumption that I was somehow digging him trying to prove him some kind of fool. The reasons he refused to change had nothing to do with him being old fashioned. That is another one of YOUR assumptions.

 Let me try one more time, Stats - My question is... how would you apply some of the new thoughts on batting order to the typical HS game where the games are only 7 innings and there tends to be a gap in ability between the top six or seven and bottom two or three hitters?  If you would like to take a stab at answering, great.  If you feel I haven't qualified my question well enough, no problem.  Sit this one out.  Others don't seem to have the same issue.

I already answered the question for at least one batting position, and for the record, I can’t find any reason why if the HS game were 6 innings or 9 innings, that would change anything. I could do the same for every batting position so the computer could generate the lineup, but I suspect you’d come up with something that tried to shoot that down as well. The only reason I ask for qualifications is because I know that if something can’t be repeated, it really boils down to little but luck. I understand that when someone asks question it comes across as not believing or trusting the other person, but that’s not why I do it at all. If there’s some valid way to generate a lineup the best way, I want to know what it is.

Originally Posted by realteamcoach:

One of the tough things about stats is that there is such a small amount of data for HS baseball.  I worked on DIPS for a D1 university about a decade ago, and it is just a lower level of baseball with too much difference in ability to tell us much.  Therefore, HS, College and pro baseball have to be viewed almost as 3 entirely different games.

There are lots of similarities between the different levels, but the minutia of each is very different, to the point where thinking about them as 3 different games is likely a valid philosophy.

The 2 things I have learned in studying HS stats over the past few years:

1. RBI are a much more telling indicator of a good offensive player/predicting future success at the HS level than at the pro level.

I’m trying to get a handle on what you’re saying. It sounds as you you’re saying a JV player who gets lots of RBIs will likely have success on the V. I certainly wouldn’t mind running my numbers to see how well your philosophy plays out, but first I need criteria to be able to compute good offense. Just give me a list of how you judge offensive. It could be BA, OBP, SLGP, OPS, RC, ISO, SBPct, Run Production(R+RBI), or anything else or any combination of factors.

2. Bunting is a much bigger weapon in HS baseball due to errors.  Sacrifice bunting is still (largely) a waste, but bunting for hits makes a tremendous difference for offense.

I have to agree in great part.

I am looking at translations of JV-Varsity stats, and my main offensive hypothesis is that strikeout rate for hitters is a very telling indicator of varsity success.  So basically if a JV guy strikes out a high amount, even if his other stats are great, that guy will not project as well at the Varsity level as you would imagine.  My summer season ends in 2 weeks, and I will then be working on my data.

When you say strikeout rate, are you thinking K/PA, PA/K, K/AB, AB/K, K/BB, or something else?

I wish I could provide as much JV data as V, but I only did the JV stats for a few seasons, but you’re certainly welcome to anything I have. Give me a holler and We’ll see what I can do to get you what you’d like.