What's Holding Up The Dozier Trade?

With all due respect...you don't have a point.

If you are taking out Dozier's best year, then take out his competition's best year as well.

That's not how you would do it.  If you were removing peak performances, you would then need to project the average over the remaining years in the data for that year you just removed.  You would not simply just remove the data, you would replace the data with normalized data.

So yes, I have a point....

I understand your point... and I wouldn't call it an argument because that sounds so... I like to call it a discussion...

 

If you are going to take out Brian Dozier's best year... You have to take out the other second baseman's best year as well to be fair. 

 

Cano shouldn't get to use his 39 last year. 

Gyorko shouldn't get to use his 30 last year 

Kinsler shouldn't get his 28 last year 

Tulowitski should lose his 25 in 2014

and Hanley should lose his 30 last year

 

After you inflict the same damage on everybody else that you want to inflict on Dozier... We are back to Double Digits again. 

Except that's not how you would do it.

If you wanted to do this sort of analysis, this would be the process:

1) Remove the peak year's data for each player.

2) Determine the yearly average for each player using the data left in the data set.

3) Project that average onto the year you just removed.

This would be a very circular process which would yield the same point I already raised.

You don't account for "outliers" in data by removing them. You account for outliers by assuming the data would have been "normal" instead. In other words, you replace the outlier with the average. If you simply remove the data altogether, you create another outlier! Zero is an outlier too.

http://www.statisticshowto.com/trimmed-mean/

Yep, that exists. It's still not the right way to do it.

The way I outlined is one of the many better ways to deal with outliers, but it takes more work than the elementary grade school level link that this is.

Another way to deal with outliers is to take the mean average, take the median average, and then make a new average based on those results. This method smooths things out nicely without actually removing or changing any data (it simply reduces the weight of outliers) but works best with larger data sets.

Sigh.

http://sphweb.bumc.bu.edu/otlt/MPH-Modules/BS/BS704_HypothesisTesting-ANOVA/BS704_HypothesisTesting-Anova_print.html

What's actually holding up the Dozier trade?

And the testosterone level contained therein.

Do you normalize the low?

Entry level statistics teaches that you get rid of the highest and lowest numbers. But what if either of the highest or lowest numbers are within a standard deviation of the norm? You've started to remove valid data just to deal with some other number that is skewing things. And you're not *really* dealing with that outlier data by this process, you're just finding a way to feel more equitable about removing it. But getting the right story from data isn't about feeling fair about it. (And it's also not about removing data you don't like, it's about finding some other way to deal with it).

Not looking at Dozier's competition, here are the various methods discussed for dealing with Dozier over the past four years:

- Straight up HR/yr average: 27.75 (111 total)

- Mean + Median HR/yr average: 26.63 (106 total)

- Replacing highest and lowest with the average: 24.25 (97 total)

- Replacing the outlier with the average: 23 (92 total)

- Removing the outlier altogether: 23 (69 total)

- Removing highest and lowest: 25.5 (51 total)

If I were making a recommendation, I would use the 26/yr projection as it's the system that does not ignore his great year, it simply weighs it less. If 2016 becomes his norm, then that will work itself out over time in the calculation.

With Dozier's small sample size the story is about the same regardless of the method you use -- 42 home runs is a hell of a lot more than anyone would have expected in 2016. It would thus be a gamble to expect it in 2017, but you can take ~25 home runs to the bank until he shows that 2016 is repeatable.

You could also use regression analysis to deal with outliers, but I'll only get into what that indicates for Dozier if anyone is interested. (It's 23.39/yr.)

You really don't because you're only taking it out on one person while making him compete against everyone else.

 

That said, if we changed Dozier's 42 HR to 25 HR, he's second only to Cano and still has a double digit lead over everyone else. 

No one is arguing that if you take away Dozier's 42 that he isn't worth anything. The point is he isn't as far ahead as everyone else as is being suggested (at least until Dozier proves he can do it again). He isn't in that "super human" territory as is being suggested (at least not yet). 

What's actually holding up the Dozier trade?

image.jpeg

And the testosterone level contained therein.

No. Falvey and Levine are "rowing the boat". Unfortunately they see all too well where the Twins were (and don't want to be there), but they have no clue where they are going.

 

It doesn't matter if he leads by one home run or 20. He's the best at that particular skill and, therefore, elite.

It matters a lot in this discussion.  One of those would be a future HOFer and would greatly affect what the Twins should get in a trade. 

If the "top HR hitter" is 2 home runs above the competition, could you save your prospects and trade instead trade for the #5 HR hitter, only sacrificing half a dozen HR in the process?  Yes, that would be a smart move to keep your top prospects and keep your money.

If the "top HR hitter" is 30 home runs above the competition, passing on him for the #5 HR hitter is suddenly becomes a real, actual drop in value. 

Thus, the fact that Dozier's HR lead is due to hitting 2x his career norm in one year is worth looking at for any team. Teams would be dumb not to do that.

Also, if the Dodgers are looking for power, wouldn't they be chasing SLG instead of HR? Some players with half the HR as Dozier have similar or better SLGs. How much does HR even matter here? We could be getting deep into a discussion that is simply a red herring.

But Nick wasn't attaching value to his statement. He was referring to Dozier's skill set, which includes elite power for his position.

"Elite power for his position" isn't a value statement? What is it then?

Also, the line you quoted was not what I was responding to.

As for power, Dozier's SLG last year was .546. Very good, but not elite. There have been recent years where second basemen have had higher SLG percentages than that with a mere seven home runs. Last year he was outslugged by someone who hit 25. Knowing that, if you are the Dodgers do you still focus only on Dozier? If you are the Twins, how high to you set the bar for a trade?

Well, what might've been holding up the deal is that the Dodgers just signed Cuban Jose Fernandez. He's been out of baseball for awhile and might need some minor league time. But only 28. Do they really need Dozier now?

Entry level statistics teaches that you get rid of the highest and lowest numbers. But what if either of the highest or lowest numbers are within a standard deviation of the norm? You've started to remove valid data just to deal with some other number that is skewing things. And you're not *really* dealing with that outlier data by this process, you're just finding a way to feel more equitable about removing it. But getting the right story from data isn't about feeling fair about it. (And it's also not about removing data you don't like, it's about finding some other way to deal with it).

Do not throw away data.  Ever.  Understand why outliers might be outliers

Well... the problem with entry level statistics in the context of baseball players is a little problem called small sample size and career arc, especially when sample sizes are in years and Dozier's professional career has spanned only 8 years and only 4.5 of said years have been at the professional level. It's one thing to remove an outlier when you have a bazillion data points, but you're literally removing 20-25% of his major league career in calling it normalization.

 

More than that.  If you remove Dozier's top and bottom HR totals over the past 4 years, you remove 60 of his home runs and are left with 51.  There would be no quicker way to get a wrong answer in reality, even if your teacher might mark it as right. 

I agree that Dozier is probably good for a hair above 30 HR each year for the next couple of years. 

Elite:

noun
1.  (often used with a plural verb) the choice or best of anything considered collectively, as of a group or class of persons.

So, Mike Trout, not elite?

 

Just curious where we draw the lines of definition here.

Good point. I have nothing to refute that.

I certainly don't feel strongly enough about any of it to justify more pages in this thread. There is just way too much emphasis being pinned on both Dozier, and the trade itself.