Are we really that good?

[witesoxfan]

Have you actually seen a good presentation/plot of how well these prediction systems do at predicting people's actual performance?

 

It's an admittedly small sample size, but this begins to paint a picture.

[Jake]

It's an admittedly small sample size, but this begins to paint a picture.

 

What I thought about while reading was - okay, so we know actual WAR is very predictive of actual wins and that projected WAR is fairly predictive of actual wins. What would be nice is a comparison to other things. Maybe compare against analysts' predictions and against some other statistics, like bWAR or WARP or non-sabermetric statistics.

[Chicago White Sox]

What I thought about while reading was - okay, so we know actual WAR is very predictive of actual wins and that projected WAR is fairly predictive of actual wins. What would be nice is a comparison to other things. Maybe compare against analysts' predictions and against some other statistics, like bWAR or WARP or non-sabermetric statistics.

I think you're reaching a bit here. I won't deny that actual WAR is fairly predictive of actual wins, but projected WAR doesn't seem to be very predictive of actual wins. I mean, we're only talking about a .43 r2 between projected and actual WAR. And again, actual WAR is fairly predictive of actual wins, but it's far from perfect (BTW, not suggesting it should be). So saying projected WAR is "fairly predictive" of actual wins doesn't seem entirely accurate to me. I would say that projected WAR is directionally right more often than not, but it will still have its fair share of misses, some of which are quite sizable.

 

I'm not arguing that these projection systems don't provide value, because they do, but I would never use them more than a frame of reference when setting my own expectations. Without more human input, the accuracy of these protection systems will always be limited.

Edited December 19, 2014 by Chicago White Sox

[RockRaines]

Avi needs to have a full yr playing in the majors. there can't be

any way to determine how he would do with out a history.

That's why he played fall ball. The guy is also motivated to succeed. Look at his rehab. I said he would be back sooner than projected because he attacked it like a man possessed. I believe in the kid.

[southsider2k5]

That's why he played fall ball. The guy is also motivated to succeed. Look at his rehab. I said he would be back sooner than projected because he attacked it like a man possessed. I believe in the kid.

 

I have a lot of optimism about Avi because of his rehab. If he takes that approach to his career, he is going to be a really good one.

[nickofypres]

Line up still seems kind of top heavy.

 

Eaton, Cabrera, Abreu, LaRoche are all locks for 100+ wRC+

 

After that it kind of tails off.

 

Ramirez ended up with a 97, but that is with his hot start. Ditto Flowers (93 wRC+)

Gillaspie (108 wRC+) lived and died with his BABIP last year.

 

What can we expect from Garcia? Sanchez?

 

[YouCanPutItOnTheBoardYES!]

Line up still seems kind of top heavy.

 

Eaton, Cabrera, Abreu, LaRoche are all locks for 100+ wRC+

 

After that it kind of tails off.

 

Ramirez ended up with a 97, but that is with his hot start. Ditto Flowers (93 wRC+)

Gillaspie (108 wRC+) lived and died with his BABIP last year.

 

What can we expect from Garcia? Sanchez?

 

Here's my projections for the lineup. (Average, Homers, RBI)

 

1. Adam Eaton (CF) .305 2 HR 35 RBI

2. Melky Cabrera (LF) .295 14 HR 60 RBI

3. Jose Abreu (1B) .309 32 HR 110 RBI

4. Adam LaRoche (DH) .269 28 HR 89 RBI

5. Avisail Garcia (RF) .265 15 HR 70 RBI

6. Conor Gillaspie (3B) .265 9 HR 53 RBI

7. Alexei Ramirez (SS) .275 13 HR 61 RBI

8. Tyler Flowers © .242 16 HR 57 RBI

9. Micah Johnson (2B) .250 5 HR 34 RBI

Edited December 20, 2014 by Mike F.

[Dam8610]

I sure didn't like it then and I wouldn't like it now. It was a bad value trade.

Nevertheless, I will continue to object to bad value trades even though that one, coupled with about 8 other moves, netted a world championship.

 

You consider that a bad value trade? Two OFs, one of whom became a decent player, one of whom completely flamed out, and a C who never really had more value than a backup or below average starter. That is what the White Sox traded for at worst a #2 caliber starter and arguably a fringe ace. You call that bad value? If that's the case, then what on Earth would you consider good value?

 

[GreenSox]

You consider that a bad value trade? Two OFs, one of whom became a decent player, one of whom completely flamed out, and a C who never really had more value than a backup or below average starter. That is what the White Sox traded for at worst a #2 caliber starter and arguably a fringe ace. You call that bad value? If that's the case, then what on Earth would you consider good value?

I consider that a horrible value trade if you want the truth.

Jeremy Reed was a top prospect - top 5 in some reports. (right between Grienke and Prince Fielder). He didn't pan out, but he, himself, had more value than a rent of Freddie Garcia. We could have gotten a ton more for him.

We also included our starting catcher who stayed in the league for another decade. We took a huge gamble there, with no one behind him. Fortunately, SF cut AJ.

We also traded another outfield prospect who became a solid major leaguer and is still in the league.

 

All of that for a 4 month rent of Garcia who was basically a 3 (maybe a low 2)? .

 

Yes, it worked out. Bad value works out sometimes.

Edited December 21, 2014 by GreenSox

[ptatc]

I consider that a horrible value trade if you want the truth.

Jeremy Reed was a top prospect. He didn't pan out, but he, himself, had more value than a rent of Freddie Garcia.

We also included our starting catcher who stayed in the league for another decade. We took a huge gamble there, with no one behind him. Fortunately, SF cut AJ.

We also traded another outfield prospect who became a solid major leaguer and is still in the league.

 

All of that for a 4 month rent of Garcia who was basically a 3 (maybe a low 2)? .

 

Yes, it worked out. Bad value works out sometimes.

I would consider that an outstanding trade for the sox. Anytime you can get a proven MLB mid-top rotation starter for 1 top suspect, another one who was suspended for PEDs and wasn't showing much and a below average MLB catcher, you make that deal. Top pitching is the name of every game in the MLB.

[Chicago White Sox]

This Freddy was a rental talk is absurd, the Sox knew they'd be able to extend him when they made the deal.

 

GreenSox, it seems you care more about winning trades than winning World Series.

[YouCanPutItOnTheBoardYES!]

This Freddy was a rental talk is absurd, the Sox knew they'd be able to extend him when they made the deal.

 

GreenSox, it seems you care more about winning trades than winning World Series.

 

Yep. I would do that Garcia trade over again everyday of the week and twice on Sunday. There's no point in saying "what if". The Sox won a World Series partially because of Garcia's success.

 

I'll take a World Series win over a trade "win" anyday.

[Jerksticks]

I see your point man but don't you think 2 douchebags and a top prospect should be the price for a fringey #1 workhorse?

 

Also possible that dominant pitching wasn't as common then so the demand was a bit higher than you're seeing now.

[Jake]

I think you're reaching a bit here. I won't deny that actual WAR is fairly predictive of actual wins, but projected WAR doesn't seem to be very predictive of actual wins. I mean, we're only talking about a .43 r2 between projected and actual WAR. And again, actual WAR is fairly predictive of actual wins, but it's far from perfect (BTW, not suggesting it should be). So saying projected WAR is "fairly predictive" of actual wins doesn't seem entirely accurate to me. I would say that projected WAR is directionally right more often than not, but it will still have its fair share of misses, some of which are quite sizable.

 

I'm not arguing that these projection systems don't provide value, because they do, but I would never use them more than a frame of reference when setting my own expectations. Without more human input, the accuracy of these protection systems will always be limited.

 

I'm a social scientist, so we poop our pants when we see an R2 that high