Tangotiger Blog

Give me a hitting projection for Kershaw, pitching projection for Trout and fielding projection at each position for both players and I could pop it in my sim.

My guess would be, as long as Trout could find the plate and put the ball in play enough times he’d win.

The key one is the pitching projection for Trout.  That’s what will make or break it.

Kershaw however is left-handed, so, he’ll have QUITE the challenge on the field.  But, who knows, the non-pitching Kershaws might have learned to throw right-handed.

For fielding, let’s say that if the average fielder turns .700 outs per BIP, then Trout would be a .725 and Kershaw would be .650.

Trout might smarten up and not go for the fences, and turn into Ichiro.

In fact, I think peak-Ichiro would easily win.

Peak Ichiro could also throw 90mph:

https://www.youtube.com/watch?v=tvp-kuAX5Bg

I can’t watch it from the office.  Is that the one where he faced Hideki Matsui in the All-Star game in Japan?  And the opposing manager removed Matsui?

Watching a “little” guy like Ichiro throwing 85-90 should give you an idea as to how incredibly athletic and talented he is/was.

MGL or Tango, do you have any empirical data from your retrosheet files on what happens when an elite hitter faces an elite pitcher? Hitters with a 140+ wRC vs pitchers with a RA- of 75 or less.

pm: It’s in The Book.  Have you read it?

Tango @5,

You nailed it in regards to what’s in that clip.  It’s fun to see baby-Ichiro.  It’s only too bad we didn’t get to see him in MLB earlier.  I wish I could’ve watched his whole career develop.

Kershaw’s career ERA- is 65, but the last over the last four seasons, it’s been better than that (and Trout’s career wRC+ is actually 165, but it’s probably better to be conservative with these things).

We also know that Kershaw is a basically a 35 wRC+ hitter over the last four years.  So doesn’t the question simply become, “Would Trout’s pitching be better than a 286?”  Because that’s the math, right? (286/100*35=100, give or take)

So, would Mike Trout give up three times as many runs as an average ML pitcher?  If no, then he wins, right?

I have no idea what the answer to this reader’s question is, by the way, but I LOVE the question!

Doesn’t count for much (and certainly isn’t nearly as good as Xei’s sim) but just for kicks I ran all-Trout vs all-Kershaw in sabersim.com (using 2015 Steamer projections). I arbitrarily used 2014 Mike Pelfrey (8.00 ERA / 7.50 FIP) for Trout’s pitching, and there’s no fielding incorporated, plus non-closer bullpens are just league average, AND there’s a home field advantage for Kershaw. So lots and lots of flaws but here are the results regardless:

http://www.sabersim.com/gameSim?simid=627482815777

Just about exactly 2/3 won by Team Trout. Add in defense and it goes further to Trout, but add in bullpen and it goes towards Kersh. I’ll take an educated guess and say it ends up well in Trout’s favor in the end, but there’s certainly a large error bar there.

Right, you really need to nail the pitching for Trout.

There’s been plenty of nonpitchers-as-pitchers.  You can easily use that as your guideline.

I’d guess such a group would have about 2x as many walks as K, and a BABIP around .330 or so.  Easy enough to see what that gives us:

http://tangotiger.net/markov.html

I put in this:
80 AB
35 H
10 2B
3 3B
10 HR
20 BB
10 SO

That’s 22 runs per game, or about 5 times as much as MLB average.

Is the above reasonable?  I have no idea.  The tools are there.  Now we just need some effort to put in reasonable numbers.

Someone want to try?

Oops… Hits should equal to 20, if I make HR = 10 and SO = 10.

Anyway, I already closed my window, so someone else can take it from here.

The Kershaws would be playing lefties at short, third, and catcher. Trouties would run wild on the basepaths. 

The Trouties could fake those positions way better. Trout even has a catcher’s body (at least now!).

I think the Trouts would score a ton of runs. I think they’d win.

Edit to #13: Also a lefty at second.

I’m guessing the Trouts are athletic enough to actually turn a DP with a slow runner, Kershaws far less likely, especially given the speed of the Trouts.

I looked up non-pitchers pitching and Baseball-Reference has a page with just that information:

http://www.baseball-reference.com/friv/fieldPitch.shtml

I then eliminated anyone who started a game and used just players who appeared as pitchers since 1901. Here are the numbers:

252 different players
567.7 IP
696 H
477 R
417 ER
65 HR
359 BB
178 K
7.56 RA/9
6.61 ERA
1.86 WHIP
6.6% K%
13.3% BB%

Some of these guys pitched 10 or more innings, so I further reduced it down to only those who pitched 5 or fewer innings:

231 players
366 IP
462 H
317 R
296 ER
53 HR
248 BB
127 K
7.80 RA/9
7.28 ERA
1.94 WHIP
7.2% K%
14.0% BB%

Some of the non-pitchers pitching names were interesting:

Ty Cobb-5 innings, 2 R
Dave Kingman—4 innings, 4 R
Wade Boggs—2 2/3 innings, 1 R
Ted Williams—2 innings, 1 R
Razor Shines—1 inning, 0 R

Modifying my numbers to something more realistic:

80 AB
27 H
5 2B
3 3B
5 HR
20 BB
10 SO

That’s 12.4 RPG

I kept the BB/SO ratio at 2:1, which is consistent with above.  But I also figured that in today’s game, BB+SO will be quite high.

I set HR at half the K rate, a bit higher than Bobby’s numbers, but reasonable enough I think.

I set BABIP at .340, which is again higher than Bobby’s numbers, but I don’t see how I can go any lower than say .325.  So, .340 seems reasonable.

Anyway, 12.4 RPG, which is close to 300% of league average.

So, when you go here:

We also know that Kershaw is a basically a 35 wRC+ hitter over the last four years.  So doesn’t the question simply become, “Would Trout’s pitching be better than a 286?”  Because that’s the math, right? (286/100*35=100, give or take)

All that’s left is the fielding.  So, I think it’s Trout that wins this one.

ACtually, make that 1 triple, and 12.1 RPG.  That sets him right around 280% of league average.

Tango, you have to take into account base running too. Trout would run all over the Kershaw’s. Kershaw is probably great at holding runners but I’m not sure if Catcher Kershaw can make the throws to 2B and 3B. He has the arm to do it but probably not the accuracy. So Trout steals 2 bases every time he gets on.

OTOH, Kershaw would suck at base running. No steals and very little advancement on hits. It would take 3-4 hits each time for Kershaw to score.

I’m guessing thats worth around 2 runs/game as an advantage for Trout.

I have given you the tool right here:

http://tangotiger.net/markov.html

There’s a section for baserunning.  Instead of telling my what I should do (of which I am 100% aware of what I need to do), why don’t YOU do it.

The fish is there, there’s a boat, there’s a fishing rod.

I ran the numbers through my sim and Trout trounces Kershaw.  The base running difference as mentioned above is huge.  Basically, any time Trout gets on first base it is pretty easy to turn that into a triple.

Trout won 79.3% of the games.
Neutralized the defense and Trout won 75.6%
Neutralized the defense and base running and Trout won 63.5% of the time.

Kershaw had HFA.

Obvious large error bars with something like this though.

#15, I took that list from BBref, removed the Ankiels, Ruths, etc. who were at one point real pitchers.  Then removed anyone who debuted before 1980 - trying to capture the modern game.

What I was left with were just 2 pitchers with over 5 innings - Vance Law and Jose Oquendo.  I left them in as they seemed legit to be considered emergency pitchers, although Vance Law has good pitching bloodlines.

That left me with these totals:

n=132
IP 192
R 162
ER 159
H 226
HR 30
BB 152
iBB 6!
SO 73
HBP 13

ERA = 7.45
RA = 7.59
FIP = 7.05

BABIP = .290 !

Jose Oquendo and Joe Mather had 2 IBB each.  Must have been in long extra innings as opposed to blowout games.  I remember the Mather game.

Whether Trout is better or worse than the average hitter at pitching, we can only guess.

Xeifrank, how does that compare to the pitching projection you used for Trout?  That’s the real wild card.  We can estimate pretty well how a school of Trout bats against Kershaw, and we know how good a hitter Kershaw is against a sampling of MLB pitchers.  Trout’s pitching is tough, since he’ll never be allowed to do it.  And even if he is allowed, the sample will be too small to be useful.

If Trout could pitch like Chris Davis a few years back against the Red Sox (90-92 MPH, good looking change) that would almost certainly be enough.  If he pitches like me (a wild, 44 year old topping out around 65-70 MPH) then Kershaw dominates.

Setting it as 2:1 for BB to K is good.

Setting it as 2:1 for K to HR is good.

But, there’s no way to accept a BABIP that is that good.  We’d have to set it in Glenden Rusch territory, some 30-40 points worse than league average.

So, in Tango/16+17, the only other thing is to figure out what rate of PAs will be TTO.

The BABIP for Rally’s data set has a small sample size. I don’t know how to measure the error rate, but I imagine that the standard deviation for that sample is around at least 15-20 points of BABIP.His data set has a sample size equivalent to a full season for a SP. 1 season of BABIP for a starter has high variance in terms of BABIP.

I used for Trout’s pitching.

IP-180
HR-45
H-250
SO-60
BB-120
GB/FB: 1.00

The BABIP that Kershaw ended up with as a hitter against Trout and Trout’s defense was 0.279

Anyways, that is what I used.

I took a bit larger sample, the list from BBref linked above, then removed anyone whose career ended before 1920, and anyone who ever started a game.  That leaves 230 pitchers and 450 innings.

I’m using BFP-W-HBP-HR-K as balls in play, as the list doesn’t include sacrifices.  So my total is a bit less than commonly reported BABIP, but even making that adjustment you’d at best end up with an average figure, not a Glendon Rusch type number.

Anyway, the larger sample shows BABIP of .303.  The league average changed a lot over the 100 years that sample covers.  So what I did was figure expected hits based on league average.  I just used the first and last year for the player, so if his career started in 1941 and ended in 1963, I’d average a .275 BABIP for 1941 and .269 for 1963.  You can certainly try to do this with more precision, looking at the actual BABIP of the years pitched, or the teams they were playing for.

The shortcut yields an expected figure of .284 - so they were about 20 points worse than average.