Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
How did you decide which was the best hitting team and which was the best pitching team? What measures did you use?
Did you account for players playing in each game? If the best hitting team was sitting their top 3 hitters that may make a difference? If the best pitching team was starting their 8th best starter due to injury that may make a difference. If a team ended the season as the best hitting team after trading for 2 hitters in July did you count games played in April? Same for pitching.
Did you account for overall records of the teams? If the best hitting team had the 2nd best pitching and the best pitching team had the 30th best hitting, that may skew the results.
Did you account for defense? Baserunning? Park factors?
How did you decide which was the best hitting team and which was the best pitching team? What measures did you use?
Did you account for players playing in each game? If the best hitting team was sitting their top 3 hitters that may make a difference? If the best pitching team was starting their 8th best starter due to injury that may make a difference. If a team ended the season as the best hitting team after trading for 2 hitters in July did you count games played in April? Same for pitching.
Did you account for overall records of the teams? If the best hitting team had the 2nd best pitching and the best pitching team had the 30th best hitting, that may skew the results.
Did you account for defense? Baserunning? Park factors?
Runs scored per game, and ERA. Unless you can demonstrate that any of the factors you indicated did in fact skew the simplified calculation strongly enough to invalidate a sample of 20 league/seasons, they are not relevant.
Over a reasonable sample, those factors would balance out. If good pitching does in fact beat good hitting, that fact would become obvious without accounting for any of the factors that you are accusing me of ignoring. In other words, I ignored them because they have insufficient relevance unless they pile up over and over again on the same side of the ledger.
You used the example of the best hitting team having the 2nd best pitching and the best pitching team had the 30th best hitting, that may skew the results. But if in another year If the best pitching team had the 2nd best hitting and the best hitting team had the 30th best pitching, that might unskew the results, unless you want to insist (witrhout data) that in a preponderance of incidents, the results are skewed on one direction only. Show me. What probabilistic reason do you have for assuming that all the skews shift in the same direction? If you have none, and also can demonstrate none in fact, you have no basis.
So have fun. Prove, using all your nitpicking, that what I found was wrong, and don't just make wild unsupported accusations that I didn't use a valid methodology. But be damned sure that you don't neglect to enter any factor, or I will accuse you of slanting the data by leaving it out. And also make sure you properly weight each factor, since they won't all have the same degree of impact on the outcome.
I find your defensive reply particularly amusing in light of your current status.
ERA and runs scored aren't a good proxy for relative offense and defense. Take the pre-humidor Rockies and PETCO Padres. The Rockies certainly scored more runs than their talent would imply due to the park they played in. Likewise, the Padres' ERA is lower than in should be because of their park. This is quite obvious and casts serious doubt upon the metrics you chose to determine the best and worst hitting teams.
It's not good science to just assume that these variables don't matter.
I would suggest that you re-run this experiment using wRC+ for the best hitting team and FIP- for the best pitching team. Both metrics are easily found at FanGraphs and eliminate many, but not all, of the problems that I mentioned above with minimal extra work.
The biggest problem that you'll face will still be the specific matchups (the best pitching TEAM won't send up an above average pitcher in every game) and the relative differences between the teams.
I'm sorry that you respond with such defensiveness when I was only suggesting ways to improve your study. Also not a hallmark of scientific thought.
I find your defensive reply particularly amusing in light of your current status.
ERA and runs scored aren't a good proxy for relative offense and defense. Take the pre-humidor Rockies and PETCO Padres. The Rockies certainly scored more runs than their talent would imply due to the park they played in. Likewise, the Padres' ERA is lower than in should be because of their park. This is quite obvious and casts serious doubt upon the metrics you chose to determine the best and worst hitting teams.
It's not good science to just assume that these variables don't matter.
I would suggest that you re-run this experiment using wRC+ for the best hitting team and FIP- for the best pitching team. Both metrics are easily found at FanGraphs and eliminate many, but not all, of the problems that I mentioned above with minimal extra work.
The biggest problem that you'll face will still be the specific matchups (the best pitching TEAM won't send up an above average pitcher in every game) and the relative differences between the teams.
I'm sorry that you respond with such defensiveness when I was only suggesting ways to improve your study. Also not a hallmark of scientific thought.
You did not suggest any ways at all to "improve" my study. You simply nitpicked it to death because nitpicking with statistics is what you do. Furthermore, you did absolutely no study at all to show that results might have been materially different.
You use your metrics, with little extra work, and see if your results are different. You define good hitting and pitching. You do the study, using your criteria. Come back when you have something to say that differs materially from what Ive said. Then, you might actually have something to contribute to the discussion.
By the way (and I don't know why I wasted more time on this) I went back and made a spot check, and found that if I had used your criteria instead of mine, in every case, I would have matched exactly the same teams, or teams that were a close second in team hitting or pitching, still fully satisfying the criteria of "good hitting" or "good pitching". I leave it to yo to labor your whole life away believing that a FIP is of 3.87 is materially different from a FIP of 3.89 to the degree that yields false derivative results that are invalid and "unscientific".
Bill James already figured this out back in the '80's and published the results in one of his Abstracts, although I could not tell you which one. His conclusion was that there was no identifiable advantage in the post season to having been a team based primarily on offense or defense.
I would add that I agree with jtur about procedures here. filihok, if you wished to discredit jtur's work on this, you should have produced your own report employing the methodology you identified as superior. If your results suggested that jtur's method was unsound because the results were different, then you would have a valid criticism. The way that you went about it was just slinging turds.
By the way (and I don't know why I wasted more time on this) I went back and made a spot check, and found that if I had used your criteria instead of mine, in every case, I would have matched exactly the same teams, or teams that were a close second in team hitting or pitching, still fully satisfying the criteria of "good hitting" or "good pitching".
In 2011 the Phillies led the NL in both ERA and FIP-
2010 Giants lead in ERA, Rockies lead in FIP-, Giants were 3rd, 9% worse
2009 Dodgers led in ERA, Rockies led in FIP-, Dodgers were 3rd, 7% worse
2008 Dodgers led in ERA, DBacks led if FIP-, Dodgers were 2nd, 2% worse
2007 Padres led in ERA, Padres, Dodgers, Rockies and Cubs lead in FIP-
2006 Padres led in ERA, Rockies led in FIP-, Padres were 11th-12% worse
2005 Cardinals led in ERA, Astros led in FIP-, Cardinals were 6th-7% worse
2004 Cardinals led in ERA, Cubs led in FIP-, Cardinals were 6th-7% worse
2003 Dodgers led in ERA, DBacks led in FIP-, Dodgers were 2nd-3% worse
2002 Braves led in ERA, DBacks led in FIP-, Braves were 3rd-8% worse
Your method incorrectly identified the best pitching team about 80% of the time. Some were close (2008), some weren't (2006).
Quote:
I leave it to yo to labor your whole life away believing that a FIP is of 3.87 is materially different from a FIP of 3.89 to the degree that yields false derivative results that are invalid and "unscientific".
I don't believe that I said anything remotely like that
So what, I still matched teams with very good pitching against teams with very good hitting, and got near-.500 results. You just used a different process of selecting teams with very good pitching.
When you matched your selections head to head, what did you get? You didn't do it, because the results were less important to you than the appearance that your elaborate and arcane methodology was better, and that's where you expended all your flailing effort.
So what, I still matched teams with very good pitching against teams with very good hitting, and got near-.500 results. You just used a different process of selecting teams with very good pitching.
When you matched your selections head to head, what did you get? You didn't do it, because the results were less important to you than the appearance that your elaborate and arcane methodology was better, and that's where you expended all your flailing effort.
I didn't do it because I didn't have the time.
Using wRC+ for hitting teams and FIP- for pitching teams (both available at FanGraphs) and W-L records from Baseball-Reference, I came up with the following:
2011 Phillies/Cardinals 3 - 6
2010 Rockies/Reds&Brewers 10 - 6
2009 Rockies/Phillies 2 - 4
2008 DBacks/Cardinals 3 - 4
2007 Padres&Dodgers&Rockies&Cubs/Phillies 14 - 13
2006 Rockies/Mets&Dodgers 5 - 20
2005 Astros/Reds 12 - 4
2004 Cubs/Cardinals 8 - 11
2003 DBacks/Braves 2 - 5
2002 DBacks/Giants 5 - 14 TOTAL 64 - 87 (.424 winning percentage for the better pitching team)
A .424 winning % over the course of 173 games (your original sample) is 73 wins, a bit different than 85 that you came up with.
This ISN'T to say that better hitting teams are preferable than better pitching teams as I didn't factor in many of the things that I pointed out that you didn't factor in. It DOES show that the methodology that you use matters. Which was my point.
Your approach did only NL, which would have a different balance than the AL, and you (arbitrarily?) used multiple teams in some years and not others. One defect in both of our studies, is that if the P team and the H team are in the same division, their head-to-head outcome overweights the sample and would magnify an anomaly. For example, I had '02 Cin/H 5, StL/P 11. If that were reduced to a more typical 2-5, my total would have been .478 instead of .488.
Nevertheless, with two widely different methodologies, both results indicate that Good Pitching does NOT beat Good Hitting.
Last edited by jtur88; 01-01-2012 at 02:59 PM..
Please register to post and access all features of our very popular forum. It is free and quick. Over $68,000 in prizes has already been given out to active posters on our forum. Additional giveaways are planned.
Detailed information about all U.S. cities, counties, and zip codes on our site: City-data.com.
Please register to participate in our discussions with 2 million other members - it's free and quick! Some forums can only be seen by registered members. After you create your account, you'll be able to customize options and access all our 15,000 new posts/day with fewer ads.
