View Single Post
Old 07-21-2015, 08:51 PM   #179
a good little dog
MissileDog's Avatar
Join Date: Jan 2009
Location: Sabo's a good guy !!!1!
Posts: 5,581
Re: Flower, Cinnamon Stick, & Sabo

The All-Star Game
(Part 1 of 2)

Sabo was disappointed that the NL lost last week's All-Star Game 3-6 in Cincinnati. He was also disappointed that his Padres (3rd, 44-49, .473, 8 GB this morning) won't have the home field advantage in this year's World Series. However, he was happy that the traditional lone Padre, J.Upton, actually got to play, and went 1-1 with a stolen base.

But the All-Star game matters now !!!1!

Yeah, I did a little Sabo-Metric essay on that last year. What I didn't figure in is that fans can vote for the All-Star teams. So how much does a fan's All-Star vote matter?

Originally Posted by MissileDog View Post
...This gives 2 * 0 * .33 * .050 == 0. Of course, BP has Sabo's Padres at a Zero% chance of winning the World Series. How much did it matter to BP's favorite, the Angels? 2 * .072 * .33 * .050 == .002736... or roughly 1/3 of a regular season game.
Well, the short answer is going to be the same as voting for POTUS or something similar... it matters zero. But let's do the numbers anyways. Well we got... prob(being a swing vote) * delta(players elected) * prob(your team making WS) * prob(WS goes 7) * delta(home field advantage). I discussed the last three last year. This year, I'll discuss the other two, starting with the possible election delta.

What could tilting an election accomplish? Well it could (a) select a better non-pitcher starter, among players who would be on the roster anyways, (b) select a better non-pitcher starter, and 'bump' a worse player off the roster, (c) select a better backup among the nominated 'last five', or (d)(e)(f) do the opposite to the 'other' league's team. OK, let's look at these cases. I'd say both (a) and (d) are an illusion. Now-a- days, a significantly better non-pitcher selected as a backup will get as much, if not more, actual playing time as an "undeserving" elected player. Categories (b) and (d) I would also say are effectively nothings. The scenario of tilting a close three way election between two "undeserving" position players vs one significantly better player, who wouldn't otherwise make the team, really is approaching zero.

Sister J went to this All-Star Game !!!1!

So did HOFers & Padres R.Fingers and D.Winfield.

So we're really just left with (c) and (f), tilting the election for the 34th player on each roster. These are from a pre-nominated 'last five'. They are all going to be quite "deserving", but they also have one thing in common... the fans, players vote, and managers/commissioner have each passed them by. This means that, in general, their Wins above Replacement are going to tightly grouped. In fact, collectively they're going to be just slightly positive Wins above All-Star Game Replacement... as that is pretty much what they are.

Very generously, I'd say the largest delta WAR in a 'tip-able' election (assuming such exist) would be 0.1. That's over a full 162-game schedule. The All-Star game is a 1 game schedule, and these 34th non-starters are, again, generously going to be playing about 1/9th of it. Of course, there's two elections that could be tipped, so again very,very generously, let's double that. That gives a delta of 0.0001372. So if we take twice today's Baseball Prospectus highest chance to win the World Series, 17.7% for the Dodgers, and double it for their chance to get in, and we assume both 34th man elections are 'tilt-able'... the overall delta gained by tilting both elections is 8.01234568e-8. That's an extra 0.00001298 wins per 162 game season.

So... no, your block of 35 All-Star game votes don't matter. It gets worse, of course. The chances of 35 votes tilting an All-Star election are much,much smaller than 8.01234568e-8. I'll skip the pretend maths.

Flower in the Sabotat
MissileDog is offline   Reply With Quote