Power and OBP

There are a finite number of options in 9 innings of baseball offense.

You could make a fairly simple and rough computer program that could simulate a few thousand games and spit out the average number of runs produced per game. I would think that someone has done this somewhere, anyone know of a basic attempt at this?

You would just need to enter the percentage chance of each outcome for each of the 9 players in the lineup (the outcomes being SO, BB, single, double etc.). Each outcome has a certain percentage of effecting a baserunner in a particular way. It could get tricky when dealing with the potential for sacrifices and steals, but not too bad.

So you would enter in the distribution of each players power in the lineup, run the simulation, and you’d get your average runs scored. It seems like this would be really interesting for playing with batting order and for looking at a power vs. obp debate.

…OBP makes power less necessary. However, it doesn’t make it less valuable, unless you’re really climbing up that .900 OBP side of the spectrum (even a .400 team OBP wouldn’t really be near there, of course)

You write that power maximizes a team’s efficiency in run scoring. This makes sense, and agrees with the original thought I had when I started reading this article. Given that we have A.) a relatively obvious lack of power, and B.) a large increase in team OBP this offseason… it seems that while adding MORE OBP would be far from detrimental (in fact, would be quite solid considering our starting point), adding power would allow us to realize the full utility of what we already have gained.

After all, 3 baserunners per inning who are depending heavily on BB and 1B won’t score all that often – and that requires a highly unrealisting .500 team OBP. Put a guy with a decent ISO in after every third OBP guy (or so), and perhaps you’ll have a few shorter innings… but several which are much bigger, even with fewer average baserunners.

My point? Love the article, and the point makes sense. I just doesn’t think it proves we don’t need power hitters, so much as it shows that it’s not a critical need. The power/patience juxtaposition actually highlights the fact that a true slugger would help us get the most out of all those added baserunners.

I did some playing around with Tango’s Markov simulator where you can put in a batting line and it spits out a bunch of cool results.

The stuff I looked at was trying to compare offenses that only do two things; walk and hit home runs. Should be simple enough to understand, right?

Anyway, I played around with it for a while and have a cool spreadsheet with results (I’m sure there was a better way).

If you have a team that has BB/PA = 25% and HR/PA = 0% => 0.75 R/G
If you have a team that has BB/PA = 0% and HR/PA = 1% => 0.27 R/G
If you have a team that has BB/PA = 25% and HR/PA = 1% => 1.34 R/G

Basically by hitting 1 HR per 100 PA you have doubled your run production. This is because the run value for a home run for a team with BB/PA = 25% and HR/PA = 0% is 1.55 runs/HR.

Anyway, I think I’m a little lost in numbers right now and nothing useful is popping out of my head although I feel like there is something pretty good up there right now.

Before all these new fangled offensive metrics came out OPS was one of the best we had to easily describe the level of offensive production because it accounted for OBP and SLG and actually tracked well with offensive production. Interestingly, OBP*SLG was always a better predictor of offensive production than OBP+SLG but never caught on because OPS is just so darn easy to calculate and understand using conventional slash numbers.

The thing is, we should think of offense as OBP times SLG. Its just that simple. Maybe I’m just having a personal epiphany here but I think this makes total sense to me.

If you hit a single you increase your SLG because you got a base (SLG = total bases / at bats) and also increased your OBP because you didn’t make an out (OBP = (1 – outs) / plate appearance). Basically SLG and OBP represent the two facets of offense of getting onbase and moving around the bases and the interesting part is that we can predict production by multiplying them so run production is best thought of as depending on the product of OBP and SLG and not just the linear combination of them. Now there are two ways of thinking of this.

You can think of it as the batter (you) having some personal slugging percentage and everybody else (your team) having a collective OBP. In this scheme your slugging percentage is made more valuable based on how many players are on base. If there is 1 person on base your home run is worth 2 runs but if there are 2 people on base your home run is worth 3 runs. Your team’s OBP is basically a multiplier that makes your SLG more valuable to the team. I will call this the RBI approach because its based on how many RBI’s you will get (Think Lopez).

On the other hand you can think of the batter (you again) as having a personal on base percentage and your team having a collective SLG. Your OBP tells how often you get on base and your team’s SLG tells us how often you will get driven in for a run. I’ll call this the Runs approach because its based on how many R you get (Think Ichiro).

In this way we can see we are multiplying two numbers together. Just walks or homeruns by themselves will produce relatively poor offenses. Because SLG and OBP multiply (1×10 = 10 and 5×5 = 25) we’d like to be good at both OBP and SLG if possible. Having just high OBP or SLG is good but without the other you can never be great.

OBP*SLG isn’t rigorously the most accurate way to estimate runs but it really is I think one of the easiest ways to think about why both are valuable. I’m not sure if that helped anybody but I think it helped me a little bit. I feel like there is a better way of saying what I said but I think the concept is there and pretty simple.

Also a disclaimer: OBP + SLG does a pretty decent job of predicting run production so in effect over the relevant ranges of OBP and SLG in the MLB, OPS is a pretty good predictor. OBP*SLG is only slightly better however it is definitely slightly better and probably accounts for teams which have very lopsided OBP/SLG totals. The difference is small so usually it doesn’t matter much and linear weights works fine. Only in the most extreme cases do we really need to worry and even in those cases it probably isn’t much more than 10-20 runs over a season for an entire team.

http://www.tangotiger.net/runscreated.html

This article discusses the linear weight values of various hit types as a function of team OBP (which he calls “on base average” in the article.)