The Rockies have lost Matt Holliday, Brian Fuentes is gone, and Todd Helton Has Seen More Glory In His Rear View Mirror Than He Can Expect On the Road Ahead. Will 2009 Be A Better Year? What Does "Marcel the Monkey" Have to Say? ...

by tpaine2009 on Dec 27, 2008 1:37 AM MST Comment 26 comments

If I am reckoning correctly, the Monkey thinks things will be looking up in '09 and the Rockies will do better than .500 . "Marcel the Monkey" tells me that the Rockies will go 83-79, with a .512 winning percentage.

Here is how I got there:

I started with the 2009 player projections published by TangoTiger, using The Marcel the Monkey Forecasting System, available at http://www.tangotiger.net/marcel/.
I then selected the hitting projection stats for what I thought would be the likely roster of position players for the Rockies for 2009. To get there, I excluded Holliday and substitued in Carlos Gonzalez, the one position player that I saw the Rockies get in the trade with the A's. I also knocked out Taveras, Herrera, and Bellorin. (I excluded hitting stats for all pitchers, an easy choice since Tango Tiger's Marcel 2009 projections excludes pitcher hitting stats.)
I then selected the pitcher projection stats for what I thought would be the likely roster of starters, middle relievers and closers for 2009. With Fuentes effectively gone, here is the list I selected: Starters -- Francis, Cook, Jimenez, Reynolds, DeLaRosa, Greg Smith; Relievers - Embree, Buchholz, Corpas, Huston Street, Grilli, Vizcaino. I also included the predicted stats for Hirsch, Morales, Register, and Speier.
This basically got me to a roster of 37.
I then used these stats to calculate an overall projected James' Pythagorean Winning Percentage for the 2009 season, using the following formula: RS^1.82/(RS^1.82 + RA^1.82). Where "RS" is the team projected Runs Scored for the season, and "RA" is the projected Runs Allowed for the season.
How did I get to the projected 2009 RS and RA for the team? This may be overly simple -- and comments are welcome here -- but I simply aggregated the predicted runs column (R) for each of the position players and the predicted runs column (both earned and unearned) for the likely pitching roster, as published by TangoTiger for each of the players.
Using the Pythagorean Winning Percentage, I then calculated projected wins using the 162-game season.
A similar approach has been used elsewhere by others -- http://sabermetrati.blogspot.com/2008/08/welcome-to-sabermetrati.html -- to compare predicted and actual performance as a season progresses.

So here is the bottom line that I got and how it compares to the Rockies' overall team 2008 performance:

	W	L	WPct	RS	RA	PythWPct	Pyth W	Pyth L
2008 Actual	74	88	45.68%	747	822	45.66%	74	88
2009 Marcel	-	-	-	841	820	51.15%	83	79

Is this the best projection for 2009? Is it the best way to calculate predicted RS and RA for 2009?

If so, what accounts for the projected improvement over 2008? Holliday is gone but Helton, according to his Marcel projections, should contribute more -- assuming he goes a whole season? The predicted improvement is to the offensives side, if you compare the projected RS and RA for 2009 with actual runs scored and allowed for 2008? How can that be? Can a full season of contributions by Helton be the sole explanation? Not likely, but what else?

Let me know what you think. Comments and Critiques are welcome.

Repost from: http://tpaine2009.wordpress.com

Eat. Drink. Be Merry. But the above FanPost does not necessarily reflect the attitudes, opinions, or views of Purple Row's staff (unless, of course, it's written by the staff [and even then, it still might not]).

0 recs | Comment 26 comments

Story-email Email Printer Print

Comments

This is a bit simplified... but pretty close to what I have

The big negative for the Rockies on the run prevention front that you can’t figure in with Marcel is defense, and with Hawpe, Spilborghs and Atkins, the team will likely have three liabilities built in from the beginning, a couple of big questions on that front is how much of a defensive rebound can we expect from Tulo, and how long will Atkins stay at third?

The positives that Marcel isn’t seeing for RA is that Jorge De La Rosa could be a different pitcher now than he was for the bulk of his career, and that some of our other young pitchers should be ready to take large steps forward this year. Depth of young pitching is going to help the Rockies and Dodgers a lot in 2009, and hurt the D-backs. The Giants don’t have quite the quality of depth of the Rox or Dodgers (for now, in 2010, their quality will be better) but at least they have some. Marcel regresses these pitchers toward average, but with above average abilities, any of them could take steps far beyond what the monkey sees.

As for run scoring, I don’t think we’re going to be quite as rosy, I have us around 820 runs right now, but I see scenarios where we could get up to 840 or 850. Still, since I have the pitching and defense only allowing 790, our spread is about the same, and I came out with an 84-78 record.

Fowler may be a bit more ready to contribute in 2009 than Gonzalez, especially given CarGon’s recent work in the VWL.

by Rox Girl on Dec 27, 2008 8:28 AM MST reply actions 0 recs

Playing time.....

I think you need to adjust playing time for what you think is realistic for each player, and then recalculate the number of RS/RA based on your estimates of playing time.

The other factor to consider is managerial tendencies. In the Clint Hurdle years, he has underperformend the Pythagorean winning percentage by three to four games, despite last season’s on-target performance. So if, say, the Rockies project to 83 wins, I would cut that prediction by 3 wins or more, to 79 or 80.

The rationale for a manager’s negative impact on PYTH winnning % is that managers with inferior in-game strategies and tactics will lose more than their fair share of close games. In the case of Clint Hurdle, I believe that case has been proven over time.

by GoRoxGo on Dec 27, 2008 11:19 AM MST reply actions 0 recs

Also, forgot to mention.....

that in the aggregate, you’ve got about 6,300 plate appearances to allocate, and about 1,440 innings.

That’s the average of how many PA/IP exist in a 162 game season (+ or – 2%). So, if your roster of 21 hitters added up to, say, 6,800 PA, you have to adjust RS downward. Or if your 16 pitchers have 1,600 IP, you have the RA too high due to the number of innings.

It’s actually too early for you to do this due to the number of unsigned free agents. Who knows. We may pick up a couple or three yet, so our roster (as well as those of other teams) is still a work in progress.

by GoRoxGo on Dec 27, 2008 11:57 AM MST up reply actions 0 recs

The last two seasons Hurdle's managed to be pretty much level with the Pythag

If you give greater weight those more recent performances, and you should, his liability to the team isn’t close to three wins. I don’t want to get into another debate about this because I always end up sounding pro-Hurdle, which I’m not, but I just don’t see the evidence that he costs the team as much as you indicate.

by Rox Girl on Dec 27, 2008 12:45 PM MST up reply actions 0 recs

I won't debate you about.....

the lack of recent empirical evidence from the last two years. But we who have followed the team throughout the years know the truth. Don’t we?

BTW, these are the actual wins vs. Pyth. wins through the Hurdle years:

2008 – 0
2007 – -1
2006 – -5
2005 – -3
2004 – -5
2003 – -4
2002 – +3

Seems Hurdle was at his best when he took over for Buddy Bell after the first 22 games of the 2002 season. I remember the surge the team had in response to the change. It didn’t last.

by GoRoxGo on Dec 27, 2008 4:58 PM MST up reply actions 0 recs

Again, I'm not pro-Hurdle

And I do feel he’ll be somewhat a liability, but you have to show the equivalent of thirty runs of liability to make him cost us three or more wins. I could see him costing us up to ten to fifteen runs after his good calls and blind squirrel nuts get factored in, which means two wins tops, but proving his responsibility for costing the team thirty runs a season during his tenure seems a little out of reach to me. Given our personnel, 2003-2006, I wouldn’t expect us to play well with the P-thag. We were a garbage team that others would use to rest their starters on getaway days after winning a couple and frequently they got burned by it in the form of blowouts.

Yeah, this is why I don’t want to get into this discussion because it sounds like I’m defending Hurdle, which I’m not, I’m just not believing that any manager could reliably be counted on to do 3 wins or more of damage.

by Rox Girl on Dec 27, 2008 5:57 PM MST up reply actions 0 recs

It's a constant baseball debate....

regarding how many wins a bad manager can cost a team, or how many a good manager can add to team wins. You are of the school of thought that a manager just simply isn’t that critical to wins given that 10 runs = 1 win, and affecting a team by 30 runs is a lot.

I beg to differ. In my view, a manager can impact a team’s RS and RA by well over 30 runs in the course of a season, and keep in mind that such run impact is also apt to be high-leverage runs, i.e. runs that are critical difference makers in close games.

Now, I have no empirical proof. I don’t even have anything to show you regarding how Hurdle impacts his team’s RS/RA differential compared to average or good managers. However, I’ll list some things that Hurdle has done over the years that have been unhelpful, and even harmful to our chances of winning:

1. Giving Willy Taveras over 530 PAs, mainly in the leadoff or #2 spot, when better hitters were kept on the bench.

2. Giving Chris Iannetta fewer PAs than Torrealba until it became too silly to continue.

3. Leading MLB in SH attempts (or in the top 2) almost every season.

4. Refusal to bring in closers in the 8th inning to get 4 or 5 outs for a save.

5. Refusal to take Helton out of the cleanup spot, even against lefties, when his SLG had dropped below .400…….until it became too silly to continue.

6. A tendency to not give relievers enough time to warm up, causing shaky outings.

7. Ignoring platoon splits too often in setting lineups against pitchers with extreme splits.

8. Or, playing platoon splits against guys like Glavine who have reverse splits.

9. Giving too much playing time to 24th and 25th guys to “keep them sharp”. Finley and Mabry, for example. Earlier, Greg Norton, Denny Hocking, Desi Relaford, etc. etc. etc.

10. A nasty tendency to make one too many moves in a key late inning situation, the last of which costs his team the game. A general lack of understanding of in-game tactics.

Managers also set the tone, and shape the team psychology. Their #1 job is to get the most out of his available talent, to get his players to play to their abilities. IMO, most of the time Clint has really fallen short in this department, 2007 being a notable exception.

Add it all up, and I think you’d be surprised by how much a manager can affect the runs differential. Personally I think he can affect RS by 25 runs in either direction, and also affect RA by the same amount. That’s a spread of 100 runs between the worst case and best case in managerial impact. 10 wins? Perhaps more, if those are high leverage runs. In short, good managers can eek out 5 more wins from a team with, say, 85 win talent, to result in 90 wins. Poor managers could result in the opposite, or 80 wins. And I think that’s being conservative.

Oh, one more thing. Hurdle has also been notorious for stubborn refusal to make adjustments to player roles until he’s practically forced to do so. Remember how he stuck with Chacon as his closer until the final week of the 2004 season, even though it was clearly obvious to most fans that Chacon couldn’t hack it as closer by June?

I can’t quantify it easily or provide empirical proof, but Hurdle is really bad as a manager. And I’ve seen more than a few in my nearly 50 years of following baseball.

BTW, you don’t need to keep saying you aren’t pro-Hurdle, Rox Girl. I know you aren’t and I’m not saying you are.

by GoRoxGo on Dec 27, 2008 7:05 PM MST up reply actions 0 recs

I agree with almost all of your points there...

through #3 are particularly problematic in my opinion, and to #8 you can add the same sort of thing with refusing to play Ian Stewart against LHP’s despite him hitting them better than RHP’s. At the same time, I know instinctively not to trust lists that only run through negatives, as there are several positives that Hurdle has done which cancel some of them out, particularly in the last two seasons. We haven’t seen him run up pitch counts or leave starters in too long -Giants pitchers had 276 PA’s last season after throwing 101 pitches, the Rockies 131. He’s shown he’s learned his lessons about sticking too long with a bad thing in the bullpen, replacing both Brian Fuentes and Manny Corpas as closers when they’ve shown signs of collapse the past two seasons, and fairly quickly relegating Luis Vizcaino away from high leverage situations. Matt Herges only pitched in games where the Rockies were already losing or comfortably ahead from the beginning of July on, while Jason Grilli took more leveraged innings. There probably aren’t a whole lot more positives that I could name, team baserunning could be one, but the negatives definitely outweigh them still. I guess I’m just still having trouble getting to 30 runs worth of bad management.

by Rox Girl on Dec 27, 2008 8:50 PM MST up reply actions 0 recs

There's a difference...

… between a manager costing his team when translating RS and RA into wins (the Pythag vs. actual measurement) and a manager costing his team by causing them to score or allow more runs (which will affect both Pythag and actual wins).

I believe the manager has little effect on the first, but can certainly hurt his team significantly on the second.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 7, 2009 8:45 AM MST up reply actions 0 recs

Wouldn't Coors make the Rockies Pythag always look better than their actual record?

At home, they win a lot, scoring and giving up a lot of runs.

On the road, the lose a lot, scoring and giving up much fewer runs.

The margins of victory at home, due to Coors, is much larger than the margins of loss on the road. Thus, even when finishing at .500, they Rockies will likely score more runs than they give up.

In other way of putting it is that a run scored at home is less valuable because it takes more of them to win. Instead of 10 runs per win, you’d use 12 runs per win. If you pro-rate the runs at home down to the same scale of usefulness as road runs (maybe 9.5 runs per win), their Pythag doesn’t look as impressive.

Personally, I think a manager’s effect on Pythag vs. actual wins is way overblown.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 7, 2009 8:43 AM MST up reply actions 0 recs

If the Coors effect affects all teams across the board,

inflating runs scored equally, then why would it change pythag? Pythag is based on run differential….

Ceci n'est pas une signature.

by DbacksSkins on Jan 7, 2009 7:13 PM MST up reply actions 0 recs

The Coors effect doesn't affect all teams equally...

Since it is in part the effects of a team playing their home games at altitude, it’s also a reference to the lag the Rockies experience when they go on the road as the theory goes that they’ll be slower to recover from muscle fatigue, or their pitchers will struggle to readjust to throwing pitches that actually break, or their hitters likewise slow to adjust when facing said pitches. Coors the field affects all teams equally, but the phenomenon known as the Coors Effect is something different.

by Rox Girl on Jan 7, 2009 7:32 PM MST up reply actions 0 recs

I appreciate what you're saying,

but wouldn’t the Rockies recover from muscle fatigue more rapidly on the road, since they’re used to playing with less oxygen?

It’s the same reason in HS our XC team used to go to San Diego and kick all the Californians’ arses — Tucson is at altitude. (~2400 ft)

Ceci n'est pas une signature.

by DbacksSkins on Jan 7, 2009 8:27 PM MST up reply actions 0 recs

Short term yes,

But long term, the muscle fatigue that they don’t recover from while at home, will stack up on the road as well in a six month long season, or so the theory goes. That’s why there’s a debate whether altitude training is really effective for elite athletes who compete frequently throughout the year rather than sporadically, as would be the case with cyclists or distance runners.

Again this is just the theory that’s been put out there, and while I feel it has some merit, I think the adjustment to pitch differences might be the bigger issue.

by Rox Girl on Jan 7, 2009 8:45 PM MST up reply actions 0 recs

I would agree

that I would think the differences in pitch behavior would be more significant.

Ceci n'est pas une signature.

by DbacksSkins on Jan 7, 2009 9:25 PM MST up reply actions 0 recs

Right...

Park factors, we think, are multiplicative. So let’s say, like back in the day, Coors produces 1.5 times more runs.

At neutral parks, let’s say the Rockies lose 5 to 4 every game. At home, to keep that proportion the same in the Rockies’ favor, let’s say they win every game 7.5 to 6. They are a .500 team, as makes sense, but average scoring 5.75 runs per game and allowing 5.5. Their Pythag record would be over .500. That’s because we’re counting a run at home as important as a run on the road, when it’s worth less. One way around that is you calculate home Pythag separately from road Pythag, which would result in the Rockies being the same number of Pythag wins above .500 at home as they are below .500 on the road. Or park-adjust the runs all onto the same scale before computing Pythag, which would adjust to 5 to 4 wins at home, making their RS=RA overall and a .500 Pythag overall.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 9, 2009 11:26 AM MST up reply actions 0 recs

To be clear, I'm not talking about a Coors Hangover Effect.

I’m talking about the conversion of actual RS and RA numbers at home and away being converted into an expect win total.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 9, 2009 11:28 AM MST up reply actions 0 recs

That

is also a factor. That’s generally the Coors Effect I’m more intimately familiar with.

Ceci n'est pas une signature.

by DbacksSkins on Jan 9, 2009 9:22 PM MST up reply actions 0 recs

Oh wait. You meant....

;-)

Ceci n'est pas une signature.

by DbacksSkins on Jan 9, 2009 9:23 PM MST up reply actions 0 recs

And conversely, this might imply Pythag underrates teams palying at pitcher-friendly parks

There are likely other issues at work, though.

Beyond the Boxscore // Calling BJ Upton lazy is lazy.

by Sky Kalkman on Jan 9, 2009 11:29 AM MST up reply actions 0 recs

But that's not quite what's been happening...

Take 2006, for instance, when the Rockies Pythag record was 81-81. The Rockies scored an average of 5.63 runs at home, and allowed 5.10 park adjusted that’s 5.23 to 4.74, which works out to exactly the 44-37 record they got at Coors that year. Their entire under-performance relevant to the Pythag occurred on the road, where they were “supposed” to go 36-45 but wound up going 32-49. An adjustment like you say makes it a little closer, a predicted record of 80-82, but it doesn’t explain the difference. I think it’s largely luck, more than anything, but I wonder if what you refer to as the “Hangover effect” is the real cause, since the lag from the Pythagorean record since Hurdle or the humidor has been almost entirely on the road.

by Rox Girl on Jan 9, 2009 1:44 PM MST up reply actions 0 recs

If it were really caused by some hangover effect from Coors,

then you would expect to see it in the Dbacks, as well, since Chase Field is the 2nd highest stadium in altitude in MLB.

Ceci n'est pas une signature.

by DbacksSkins on Jan 11, 2009 6:08 AM MST up reply actions 0 recs

No, I mean, I understand the traditional theory of park factors.

But I’m not convinced that the same multiplicative function applies to run differential itself. All I was trying to say was that, since there’s a marginal increase required in adding additional runs, (not that this is groundbreaking, but rather, a problem with statistics in general, not park factors in particular) I’m not quite sure how useful it is to use a statistical % change in aggregate runs to predict run differential in a specific game. My sense of it is (based on absolutely nothing but my own instinct) that the run differential doesn’t necessarily change based on the park — but rather, more runs are scored in total, instead of adhering to a certain proportion, i.e., instead of Team A winning a game over Team B 2-1 at PetCo, or 3-2 at a neutral park, Team A beats Team B 4-3 at Coors. Using the proportional increase rather than the aggregate based on, say, the PetCo score, would yield a predicted score at Coors of 6-3 or 4-2, but I don’t believe that’s how it actually happens.

Using multiplicative park factors seems to me to be much more useful in quantifying past performance in aggregate RS rather than predicting future performance or outcomes. A bit of a blunt tool, so to speak.

Ceci n'est pas une signature.

by DbacksSkins on Jan 11, 2009 7:51 AM MST up reply actions 0 recs