The AL's top starting 5!

by roger on Jan 1, 2006 9:28 PM EST

Many of you may have seen a series I did on Seth's site this past fall. I developed a formula to rate all of the pitchers and position players in the Twins minor league system.

I recently modified that formula to compare pitchers at the major league level. I then completed the comparison of the top 5 starters on each staff (adding the totals of the 5)...ranking the success in 2005 of the starting 5 for all 14 teams as follows:

Twins 508.723
Angels 498.803
ChiSox 497.680
Toronto 491.236
Oakland 487.378
Cleveland 479.422
Boston 445.704
Yankees 444.247
Seattle 425.850
Baltimore 404.907
Detroit 404.584
Texas 381.787
Tampa Bay 383.551
Kansas City 343.612

Based on this formula, the top 25 starting pitchers in the AL were:

#1 Halladay Toronto 135.823
#2 Santana Twins 131.771
#3 Buehrle ChiSox 121.365
#4 Harden Oakland 114.502
#5 Garland ChiSox 111.476
#6 Colon Angels 110.170
#7 R.Johnson Yankees 109.239
#8 Byrd Angels 109.033
#9 Towers Toronto 108.975
#10 F.Hernandez Seattle 108.296
#11 Radke Twins 107.914
#12 Millwood Cleveland 106.818
#13 Silva Twins 106.508
#14 Haren Oakland 102.728
#15 Lackey Angels 100.832
#16 Wells Boston 99.919
#17 Sabathia Cleveland 98.172
#18 Garcia ChiSox 98.049
#19 Lee Cleveland 97.625
#20 Washburn Angels 97.521
#21 Rogers Texas 96.970
#22 Mussina Yankees 95.142
#23 Blanton Oakland 94.423
#24 Wakefield Boston 94.324
#25 Contreras ChiSox 94.046

0 recs | Comment 14 comments

Story-email Email | Print |

Comments

Formula

That thing is prety cool, although there are some people in the top 25 that I disagree with. Garland and Colon should probobly be lower and Radke souldn't probobly be in the top 25 at all. There are a few others...

Anyway, I'm curious how this is calculated, and what the theory behind it is. It looks pretty good.

by AdamOnFirst on Jan 2, 2006 1:45 AM EST 0 recs

Disagreement

This ranking is based strictly off a mathematical formula based on each pitchers performance in 2005. It is not an attempt at rating players, it just says who did best in 2005 based on the criteria chosen.

The formula is a combination of era, innings pitched, strikeouts, walks, home runs allowed, WHIP and bonus points for complete games or shutouts. (This formula was discussed at length on Seth's site last fall, however, that was in regards to ranking the minor league prospects which also had factors such as age, level of play, awards received, etc. The criteria for the major league rankings was also adjusted from the minor league formula.) Each category is weighed differently compared with a norm...with a pitcher achieving the norm in each category earning a total of 100 points. Therefore, there were only 15 pitchers on the starting rotations in the AL that were better than the "norm."

Radke clearly is one of the better pitchers in the league and certainly belongs in the top 25...11th using these criteria.

by roger on Jan 2, 2006 8:35 AM EST up 0 recs

Too lazy to

search Seth's site for the formula, so would you mind briefing it here?

Also, please define "norm".

Also, are these park adjusted numbers? Is Cleveland really going to see an improvement by replacing Millwood with Byrd?

Thanks, Coop

by cooper7d7 on Jan 3, 2006 11:28 AM EST up 0 recs

Byrd vs. Millwood

Perhaps Shapiro and Roger know more than the other GMs in the league, especially Daniels. But the one thing that jumps out at me on this list is that Byrd is ranked higher than Millwood. I would certainly take Millwood over Byrd, all things being equal. But apparently the formula weights low walks very highly and does not rate ERA as highly.

This is not the only anomolous rating, but it is the most glaring. At the minor league level, it tended to rate guys like Blackburn and Slowey more highly than guys like Garza. Part of that is that it relies on one year's worth of stats. But I do think it overvalues low walk totals and undervalues ERA. Also, there's no acounting for upside. And I don't know how you would account for it in the formula.

While I like the rating system, I think it could use some tweaking. That's to be expected considering all the variables you're trying to control for.

Joe Mauer for MVP.

by cmathewson on Jan 3, 2006 11:53 AM EST up 0 recs

Missing the basic presumption

As I have clearly stated, this formula ranks pitchers performance in 2005 based on the criteria chosen for the formula. It does not intend to state that pitcher A is a better prospect than pitcher B, or that one would be a better free agent signing than another. It merely states that based on certain factors, this is how they did last year. If I were a GM, I expect I would also rather sign Millwood than Byrd, but Byrd was surprisingly effective in 2005.

It doesn't rate walks higher than era, era is weighted at double the value of walks. Furthermore, pitchers with ridiculously low walks don't get full advantage as no category can be scored at more than 2 times the base value. For example, if walks are weighted at 15 points and the norm is 3 per 9 innings, 1.0 walks per inning would equate to 45 points, however, the maximim allowed is 30. On the other hand, pitchers such as Silva/Radke did accumulate the maximum walk value whereas no pitcher achieved the maximum value for era as no one had an era under 2.00.

Regarding the minor leagues, yes, Slowey was rated much higher than Garza...however, his era and other stats at Beloit were also better than Garza, not just walks. Does it take into consideration potential, a little with things such as being in BA's Top 20, etc...but otherwise, no. That would not be possible without having access to scouting reports, etc. which I do not have. The minor league formula also takes age and level of competition into play which accounts for potential in part when a young player such as Moses is already playing at AA.

Does this need some tweaking, probably...I did make a few minor changes on this compared with the minor leagues. It does however accomplish what I originally set out to accomplish, having something other than a guess on which to rate minor league prospects.

by roger on Jan 3, 2006 12:32 PM EST up 0 recs

Oh

I thought it was a rating system to show how valuable pitchers are relative to each other. If it is just an arbitrary rating system based on certain criteria that you find valuable, it's less relevant. So when I suggested tweaks, it was only to make it a better guage of a pitchers' worth.

And don't get e wrong, it's a good tool and I really appeciate your work. But there's always room for improvement.

Joe Mauer for MVP.

by cmathewson on Jan 3, 2006 1:13 PM EST up 0 recs

Tweaking

I also agree that it needs some tweaking. Will work on that at year end when I do this for 2006. The big question I have yet to answer with the minor leaguers is, should it include both 2005 and 2006? And if so, how should that be done? Is 2005 weighted less than 2006 which would show two years worth of performance, yet, give more weight to an upward or downward trend.

by roger on Jan 3, 2006 2:19 PM EST up 0 recs

I'm

I'm well aware that it only represents this year performance and nothing else. And I am stating that I think Radke is nowhere near the 11th best pitcher in the league, and several others are overratted.

But like I said, there is definatly some interesting stuff there.

If you really wanted to make it super legit, you could do a ton of research on all the different stats compared to runs over replacement, create a scale for each, and run it that way. But that would require, as we say, "a f*** ton of work."

I think simple adjustments I would like to see would be ERA being worth 50% of the equation, and HR/9 being worth more compareed to K/9, etc. The HR only accounting for 5 points seemed a little off, especially since that really limits the overall effect it can have (heh, duh) and giving up a ton or very few homeruns has a MASSIVE affect on total runs.

Oh, something you could incorporate could be the defeseless runs. Like consider using DipsERA instead of ERA. ERA is kind of like a benchmark you can compare pitching staffs with in itself, because it really is a near bottom line. Since we are trying to use little pieces for a better bottom line, I think we should use something more specific.

But yeah, if you really feal like having fun, let us know what changes you try.

by AdamOnFirst on Jan 7, 2006 12:43 AM EST up 0 recs

Hmm...

Colon is the sixth best pitcher in the AL and yet he won the Cy Young.

Splendid.

Nick & Nick's Twins Blog

by Nick Nelson on Jan 2, 2006 4:19 AM EST 0 recs

Cy Young

Ain't that a shame! And I suspect that no matter how you change the criteria for ranking performance...Colon doesn't get to #1 in any manner...except Baseball Writers voting.

Halladay had a phenominal year, but it was cut short. Can be an argument over who should have gotten the Cy, Halladay or Santana.

by roger on Jan 2, 2006 8:50 AM EST up 0 recs

...except

Colon was tops in Wins, the most overrated statistic out there. Wow do i hate that.

by Petey on Jan 2, 2006 1:34 PM EST up 0 recs

QS

Agreed.

I've been on Yahoo for Fantasy Baseball for the past couple years and I've been waiting for them to include quality starts as a category instead of wins. Does anyone know of a FB site that uses QS instead of wins? Other than that, I generally like yahoo.

by TheMattWilke on Jan 2, 2006 1:37 PM EST 0 recs

Fantasy

I would plug CBS Sportsline's fantasy baseball over Yahoo's. I played both this year, and while both use Wins and Saves (of course), Yahoo's uses ERA which is hard to gauge over one week, while CBS uses K's and Innings pitched as scorable stats (walks/hits take points away). It seems more logical to me. I don't know of any site that uses quality starts, but CBS' tends to favor those pitchers who get quality starts, due to what stats they score. Of course, wins are inflated, but I guess we will just always have to take that into account when we draft.

by Neil on Jan 3, 2006 10:59 AM EST up 0 recs