The All-Stars Versus The Outsiders
When Major League Baseball announced the 143 — or however many they ended up taking — players that had been picked to play in this summer’s All-Star Game last week, I noticed a few names that were missing. Not names that necessarily deserved to be there based on how they had performed so far this year, but names that you’d have expected to be easy picks before the season started. There are some pretty terrific players who weren’t selected to be part of this summer’s festivities.
The collection of names got me wondering whether it would be possible to build a roster of guys who didn’t get an invite to the All-Star Game that could compete with the AL and NL squads head to head. In other words, if I assembled a line-up from the leftovers after MLB was done selecting its All-Stars, could I field a team that was just as good?
So, I shot Dan Szymborski a note and asked if I assembled a line-up of players, would he use his rest-of-season ZIPS projections to simulate match-ups between my squad of outsiders and both All-Star squads. He said he would, so now I just had to figure out if there were enough pieces left to build a roster that could compete after MLB skimmed off the top.
In order to make the simulation easier, we skipped out on the full roster and settled on a starting nine and four pitchers, with the starter for each side going six innings, and then the three relievers each going one inning apiece — the relievers were actually Dan’s additions, so give him credit/blame for the bullpen guys. Here’s the team I ended up assembling, and yes, this the batting order I chose.
1. Carlos Santana, C
2. Evan Longoria, 3B
3. Giancarlo Stanton, RF
4. Ryan Braun, LF
5. Adrian Gonzalez, 1B
6. Adrian Beltre, DH
7. Chase Utley, 2B
8. Hanley Ramirez, SS
9. Jacoby Ellsbury, CF
SP: Stephen Strasburg
RP: David Robertson, Koji Uehara, Kenley Jansen
The line-up skews a little bit to the right-hand side, but I’m not passing on guys like Braun, Stanton, Longoria, or Beltre just to get another left-handed bat in the middle of the order. It’s kind of amazing that this is a line-up of guys who didn’t get selected to an All-Star Game, as these are some pretty big names. But, could they stack up against the starters from the AL and NL? For reference, here are the line-ups that each of those teams was given for our little experiment.
American League
1. Mike Trout, LF
2. Robinson Cano, 2B
3. Miguel Cabrera, 3B
4. Chris Davis, 1B
5. Jose Bautista, RF
6. David Ortiz, DH
7. Adam Jones, CF
8. Joe Mauer, C
9. J.J. Hardy, SS
SP: Max Scherzer
RP: Greg Holland, Joe Nathan, Mariano Rivera
National League
1. Brandon Phillips, 2B
2. Carlos Beltran, RF
3. Joey Votto, 1B
4. David Wright, 3B
5. Carlos Gonzalez, LF
6. Yadier Molina, C
7. Troy Tulowitzki, SS
8. Michael Cuddyer, DH
9. Bryce Harper, CF
SP: Matt Harvey
RP: Jason Grilli, Sergio Romo, Craig Kimbrel
All the teams are obviously quite good, but how would they do in head to head match-ups? That’s where ZIPS comes in. Using the rest-of-season projections, Dan first matched up The Outsiders against the American League squad.
The result? Basically a draw, as The Outsiders came out with a .502 winning percentage in the simulation. That’s a rounding error, basically, and shows that the leftover guys at least are in the ballpark of being competitive with a group of All-Star starters. It’s hard to celebrate a 50/50 split with any enthusiasm, though, so how would they do against the National League team?
Here, we have The Outsiders once again coming out on top, but this time, with a slightly more impressive .517 winning percentage. Braun, Longoria, Stanton, Strasburg and crew were able to win a small but decisive victory over the NL squad.
Dan also simulated a mock season in which the three teams each played 162 games against each other. The final standings:
The Outsiders: 83-79
American League: 82-80
National League: 78-84
During tonight’s broadcast, you’re likely to hear that the “best players in the world” are gathered in New York. Based on these results, that’s only partially true. A handful of the best players in the world are enjoying a four day vacation, and if they all got together and challenged either of these teams to a friendly contest, we might have ourselves a pretty interesting ballgame.
Dave is the Managing Editor of FanGraphs.
David Robertson was not an All-Star.
And Sean Marshall hasn’t pitched in 2 months.
First comment goes to the reader who missed the point of the article most completely. 😉
Santana has a .382 OBP this year, and his .466 SLG is one of the lowest marks on the squad. It’s weird to see a catcher in the leadoff spot, but on this team he fits better than anyone.
Man Trevor Rosenthal can’t even get on the leftover squad. Were the relievers taken purely on ZiPS ROS forecasts?
David Roberts, the new Johan Santa.
So you choose players based on future projections to compete against fan selected players. The fan selections skew toward favorite team bias and first half aberrations (breakouts, slumps, and injuries).
Yes, the fans got first pick of the talent, but your ‘outsiders’ are from both leagues (larger pool). Your choices are focused to the comparison that you want to make. The all-star choices have a split goal: win a single game and reward good first half performance.
It is an interesting concept, but not quite an even comparison.
But wasn’t that kind of the point?
Geez, Puig couldn’t even make the All-Star castoffs team 😛
He needs rest, even in hypothetical scenarios.
No Josh Donaldson here either? Not only Longoria ahead of him, as an AL 3B… but the DH selected, Beltre also an AL third basemen…
How about the concept… a team, comprised of only AL third basemen, compare v either AL or NL all star over 162 game schedule?
Miggy-Machado-Longoria-Beltre-Donaldson-Encarnacion-Seager-Callaspo… and one more, say it’s Chisenhall.
Lineup:
Manny
Longo
Miggy
Beltre
EE
Donaldson
Seager
Callaspo
Chisenhall
Pitcher- Chris Davis
I guess Moustakas would have to come in to close… that may be a stretch, but the rest of the team I am going to bet would fair pretty well against any of the other three lineups in the article.
lol Callaspo
Who would pitch? (I mean, 3rd basemen *are* known for their arms, so maybe it would work out.)
Chris Davis
Maybe “SSS”, sure… but consider 9.00K/9 and 2.59FIP. He could get the job done.
I wonder how many times you re-ran the simulation until the Outsiders came out on top.
I don’t think simulations work the way you think they work.
Well, he reported a .502 winning percentage, which means he ran *at least* 500 iterations of the simulation (to generate a winning percentage of 251-249), since 502 is only divisible by 251 (a prime) and 2, and couldn’t be generated by, say, 8 iterations, or 10, or 100, or so forth. If .502 is the result of rounding, and the actual percentage is something like .5019876, that would imply a considerably larger number of iterations.
Are you suggesting he ran 500 iterations of the simulation, and got a winning record of .495 or something, and thought, “no way, it is not worth running a column unless I can generate a winning percentage!”, and then ran another 500 iterations, and kept doing so until he got a result greater than .500? Then we would have to consider the potential variance in outcomes for several N=500 simulations to determine whether outcome differences in excess of .003 (from a theoretical .499 (i.e., losing) percentage to .502) would even be possible. A
gain, all for this potential possibility that Dave would, dishonestly and for the sake of his own egotism, repeat the process until he generated a winning percentage, when either a winner or a loser would make for an equally-worthwhile column; and in fact the results he reported suggested a more-or-less .500 outcome, which one would think would be the least-interesting outcome.
So, no, I don’t think he did that.
This is exactly what I would have said if I knew the fancy math you do, so instead I’ll just say it with words: Dave is right.
Not to be overly nit-picky (hint, this is nit-picking), but plenty of smaller denominators could produce that winning percentage:
101/201=.50249
102/203=.50246
et cetera
Still, your point stands. There’s zero likelihood that it would be interesting for Cameron, Szymborski or anyone else to keep crunching until the numbers skewed one way. Just looking at the lineups, I wasn’t at all surprised by the results.
The one thing that does occur to me is that David Ortiz got in as a DH. If the outsiders had to pick from the DH pool instead of the second-best 3B available, how much would that diminish their prowess? I like what I’ve seen from VMart lately but I don’t see him coming close to Beltre (or Donaldson) ROS.
101/201 rounds to .502
Also 102.5/204. Stupid Selig and his ASG ties.
No need to resort to that. ZiPS ROS is weighs seasons prior to this one a bit more heavily than the season to date.
And Ellsbury’s 30+ SBs 9th…. Im sure he’s got some statistic to say base clogging is good and speed is useless.
Can he not steal when hitting ninth?
With Yadier Molina catching for the NL squad, it’s likely that even Ellsbury won’t be running much.
Can we make a team made out of the all-star bench players and throw them into the mix? It seems like they are being left out in limbo. Also, Maybe expand the sim to include five pitchers when you do a fun season run? Who would the other 4 starters be for the outsiders?
Price, Shields, Kuroda, Hellickson? Samardzija maybe.
Clearly he’s never seen him run the bases.
It amazes me how much pitcher wins drive perception. Last year, Cliff Lee pitched wonderfully but no one noticed. This year, that honor rests with Stephen Strasburg. Though Strasburg’s last start was the worst of his career (probably since high school), the rest of his season has been otherwordly. But with the worst run support in baseball, no one has noticed.
I really wish the pitcher win stat would disappear forever.
This is a big topic going around these days, thank you, Brian Kenney.
When you are discussing a pitcher and if he is HOF worthy, what is the first question asked of a starter? How many wins did he have?
I believe this has had a big effect as to why Bert Blyleven took so long to get into the HOF.
Its amazing that as good as Beltre is defensively he has to DH on this squad.
Even in simulations Beltre doesn’t wear a cup. He got hit in the scrotum again in game 141 of the 16th simulation but still insisted on playing. It affected his defense so Cameron had to move him to DH for the rest of the sims.
I know I’m pointing out the obvious, but the differences in talent become larger when you include benches. If you have a 30 man outsider team compete with the AL and NL teams, factoring in rest for the starters and a normal rotation, I think the outsiders do much worse (especially since I’d think they’re much less deep in the starting rotation).
Stay gold, Ponyboy.
Sorry, that was a reply to Dave’s title, not Dylan’s post.
Harvey v. Strasburg at Citi? The Dark Knight of Gotham would not allow this band of misfit outsiders to prevail.
How would the simulation had changed if Dusty Baker hadn’t managed the NL squad? Goldy over Cuddy at DH, Brandon Phillips and his .083 OBP hitting anywhere other than leadoff…
You don’t think that Dusty still manages the Giants, do you?
And Bonds should be the NL’s starting LF.