Introduction to JEST

The JEST Solution to the OPS problem
Note: Unless specified otherwise, when I refer to "walks" in this... essay (I guess it's sort of an essay), I am also referring to intentional walks, hit batsmen, etc. - anything which gets the batter to first base without being recorded as an AB.Note 2: I use the symbol 4B interchangeably with HR
Back when I was first getting into baseball, a bit over ten years ago now, and looking at the statistics, I thought about what an optimum offensive statistic would be. Based on the "a walk is just as good as a single" premise, I came up with the idea of adding walks to total bases (as one each), and then dividing this total by PA. I never gave this a proper title, but heretoforth it shall be called myOPS, really a rather bad title actually. This statistic is usually very close to OPS - BA, but not the same unless PA=AB. There are many problems with this line of thinking. The most glaring is that in many situations, a walk is NOT as good as a single - basically this is only true when no one is on base (or when the bases are loaded in a tie game in the bottom of the 9th or any extra inning) and in some cases where there is not an empty base between the batter and the lead runner. This is because singles have the ability to advance runners more than one base - a runner on first can often make it to second, a runner on second can often score. Also, if there is an empty bag between the batter and lead runner, on a walk, the lead runner does not advance at all.Furthermore, the problems with slugging percentage remain. Doubles are not twice as good as singles, triples are not thrice as good, and homers aren't four times as good.However, there is also some merit to the myOPS system. Every plate appearance is put into the same pool, which is something that everything other than OBP lacks. OBP has the inverse problem from SLUG - a walk or single is not equally good to a home run.So keeping the merits of these systems, I'm thinking that the ultimate statistic of this kind (where situationalism, lineup, ballpark, etc. are not factors) would be of the form (A(BB) + B(1B) + C(2B) + D(3B) + E(4B) + F(Outs))/PAA, B, C, D, E, and F are as-yet-undetermined coefficients, though likely E>D>C>B>A>F. For simplicity's sake, I've taken out sacrifices, hit batsmen, fielders' choices, and the like, both because they happen relatively infrequently and because they probably aren't a great measure of a batter anyway. Thus my PA is actually At Bats+BB, and my "Outs" is At Bats-Hits. If you wanted to be more specific, all of these things would be included, and you might split "Outs" up into strikeouts, groundouts, and flyouts. Conceivably, you could go to many degrees of finer sophistication, but that would go against the simplicity concept I'm going for here.I have made educated guesses as to what these values, and done some calculations with them. I wanted the scale to be topped out at 1 with the average player to be 0, which helps explain my scale; of course you could shift or dilate this if you want.A = .4, B= .5, C = .7, D = .8, E = 1, F = -.25. Rationale: Singles are only marginally better than walks, for the reasons listed above, which actually come up rather occasionally. Doubles are significantly better than singles, being able to drive in runners from first fairly often and from second nearly all the time, as well as having the batter wind up on second instead of first. Triples are only marginally better than doubles, because they likely don't drive anyone else in - occasionally a guy from first who wouldn't have scored on the double - though they leave the runner on third instead of second. And homers are more advantageous than triples significantly because the batter himself scores. Notice that the difference between an out and a walk (.65) is actually bigger than the difference between a walk and a homer (.60). This is because of the success/fail nature of the PA - a walk is a success, not creating an out and generating some offense, whereas a homer is only a better not-out. BB vs. out is like the difference between success and failure, whereas BB vs. HR is more like the difference between success and more success.I call this statistic JEST.