|
Post by 01aztecgrad on Jul 26, 2011 19:45:49 GMT -8
Throwing out tainted data is a very important in science. It is also very important if you want to accurately judge the Padres offensive too. Did you use to think the Rockies were the best offensive team in baseball every year before the humidores? The data aren't tainted. Playing in Petco isn't an anomaly, or an unusual event that can be discarded. Valid data is often adjusted, as when seasonal adjustments are made to sales or unemployment data, but I have never known of anybody who completely discards half of a sample as you always do. If you're going to claim that it's statistically valid to throw out the home games, why don't you provide the correlation or R^2 to demonstrate that Petco has been a significant factor in the variability of runs scored this season? If you're going to argue for throwing out data you should at least support your case with evidence. Or is it enough to just go by the eyeball test with outliers?
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 27, 2011 8:10:32 GMT -8
There is so much variance it is impossible to accurately compare one offense to another by your means Bill. You would have to take into account every opposing team, opposing ballpark, opposing pitcher, opposing defense. If you want to look that deep into statistics be my guest. But you will never find a way out without coming up with some lurking variable in one aspect or another. what we can go by are basic statistics fans and analysts have used for years that have paid off and proven true. The phillies have by far and away a better offense than the Padres. You can come up with every stat you want, but anybody that watches the game of baseball knows that the Phillies have about 3 or 4 players on their team that are better hitters than anybody the Padres have. Looking at games at home and games on the road is very broad divisions. Considering the offensive aspects of the two ball parks, road data is more accurate than overall data, because home data skews the data too much for both the Phillies and the Padres for its unaltered use. Wanting to look at the very obvious half the season breakdown, does not require further breakdowns to look at actual team and pitching matchups as you suggest. If two people's running times are compared, it would be nice to know if one was running uphill half the time and the other downhill half the time. You suggest we use all the numbers when we have other data showing them running on fairly equivalent ground. Throwing out tainted data is a very important in science. It is also very important if you want to accurately judge the Padres offensive too. Did you use to think the Rockies were the best offensive team in baseball every year before the humidores? There are many more variables you haven't considered though Bill. The Padres and the Phillies don't play the same road schedule, don't face the same pitchers, there are many more varying factors than just road vs home. Why don't you account for all of them? Why not account for the ERA of the pitchers in the respected divisions? Who you're facing makes a bigger difference than where you're facing them. But then you would have to disregard the numbers pitched in Philly since it's an unfair offensive advantage. Do the same for the pitchers in the NL West, but disregard the numbers at PETCO since it's an unfair offensive disadvantage. The point is you either take the stats at face value OR you get lost in so many statistics it becomes impossible to compare two teams because the fact is no two teams will ever be on the same level running surface. There is too much variance. There's too much luck, random chance, injuries etc. If you want to tell me that all that random chance evens out, then prove that too. Because there's nothing that says random chance evens out. Comparing one teams road numbers to anothers is just as unfair as comparing overall numbers.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 27, 2011 8:12:23 GMT -8
Throwing out tainted data is a very important in science. It is also very important if you want to accurately judge the Padres offensive too. Did you use to think the Rockies were the best offensive team in baseball every year before the humidores? The data aren't tainted. Playing in Petco isn't an anomaly, or an unusual event that can be discarded. Valid data is often adjusted, as when seasonal adjustments are made to sales or unemployment data, but I have never known of anybody who completely discards half of a sample as you always do. If you're going to claim that it's statistically valid to throw out the home games, why don't you provide the correlation or R^2 to demonstrate that Petco has been a significant factor in the variability of runs scored this season? If you're going to argue for throwing out data you should at least support your case with evidence. Or is it enough to just go by the eyeball test with outliers? Owned. DO YOU EVEN KNOW WHAT YOUR P VALUE IS FOR DETERMINING STATISTICALLY VALID DATA? DO YOU EVEN KNOW HOW HARD I'M TRYING TO CHANNEL MY BASIC LEVEL STATISTICS CLASS RIGHT NOW?! CAN YOU EVEN FATHOM THAT?!? ANSWER ME DAMNIT
|
|
|
Post by AztecBill on Jul 28, 2011 13:05:01 GMT -8
Looking at games at home and games on the road is very broad divisions. Considering the offensive aspects of the two ball parks, road data is more accurate than overall data, because home data skews the data too much for both the Phillies and the Padres for its unaltered use. Wanting to look at the very obvious half the season breakdown, does not require further breakdowns to look at actual team and pitching matchups as you suggest. If two people's running times are compared, it would be nice to know if one was running uphill half the time and the other downhill half the time. You suggest we use all the numbers when we have other data showing them running on fairly equivalent ground. Throwing out tainted data is a very important in science. It is also very important if you want to accurately judge the Padres offensive too. Did you use to think the Rockies were the best offensive team in baseball every year before the humidores? There are many more variables you haven't considered though Bill. The Padres and the Phillies don't play the same road schedule, don't face the same pitchers, there are many more varying factors than just road vs home. Why don't you account for all of them? Why not account for the ERA of the pitchers in the respected divisions? Who you're facing makes a bigger difference than where you're facing them. But then you would have to disregard the numbers pitched in Philly since it's an unfair offensive advantage. Do the same for the pitchers in the NL West, but disregard the numbers at PETCO since it's an unfair offensive disadvantage. The point is you either take the stats at face value OR you get lost in so many statistics it becomes impossible to compare two teams because the fact is no two teams will ever be on the same level running surface. There is too much variance. There's too much luck, random chance, injuries etc. If you want to tell me that all that random chance evens out, then prove that too. Because there's nothing that says random chance evens out. Comparing one teams road numbers to anothers is just as unfair as comparing overall numbers. Someone is running up a hill and you're wondering if there was a tail wind. Look at the Texas Rangers numbers and tell me how good you think their offense and defense is compared to the rest of the league and what their strength is. ERA 4.39 Home #26 out of 30 teams 3.10 Road #1 out of 30 teams. 3.79 All Games #13 out of 30 teams. Scoring4.04 Road #17 out of 30 teams (4.04 - just like the Padres) 6.07 Home #1 out of 30 teams. 5.10 All Games #3 out of 30 teams.
|
|
|
Post by AztecBill on Jul 28, 2011 13:33:25 GMT -8
Throwing out tainted data is a very important in science. It is also very important if you want to accurately judge the Padres offensive too. Did you use to think the Rockies were the best offensive team in baseball every year before the humidores? The data aren't tainted. Playing in Petco isn't an anomaly, or an unusual event that can be discarded. Valid data is often adjusted, as when seasonal adjustments are made to sales or unemployment data, but I have never known of anybody who completely discards half of a sample as you always do. If you're going to claim that it's statistically valid to throw out the home games, why don't you provide the correlation or R^2 to demonstrate that Petco has been a significant factor in the variability of runs scored this season? If you're going to argue for throwing out data you should at least support your case with evidence. Or is it enough to just go by the eyeball test with outliers? That is done all the time. A simple example is surveys. Questions are asked and many many results are thrown out to get a balanced demographic representation of the population. Baseball is like a survey that has half the people describe themselves as Libertarians. Hell ya, you need to throw out a bunch of the data.
|
|
|
Post by 01aztecgrad on Jul 28, 2011 15:48:09 GMT -8
That is done all the time. A simple example is surveys. Questions are asked and many many results are thrown out to get a balanced demographic representation of the population. Baseball is like a survey that has half the people describe themselves as Libertarians. Hell ya, you need to throw out a bunch of the data. Wrong. First of all, if you're conducting a survey in which you want the responses to match the population and you have heavily skewed responses, you probably didn't select your respondents at random, and you should rethink your study design. If you did butcher the first step, you wouldn't throw out surveys to get the proportion of responses you want. You would either: 1) Ask the demographic questions up front, and stop taking the surveys from people whose profile you no longer need. 2) Use all of the surveys to determine how libertarians respond, and adjust the weight you assign those responses to represent the proportion of libertarians in the population. Your philosophy of "when in doubt, throw it out" isn't encouraged in any statistics or forecasting class. Why don't you use park adjusted data to support your theory? Is it because it demonstrates that Petco is only a minor factor in explaining why the Padres are so horrible at home offensively? To any rational person, that would strongly suggest that there are factors much larger than the stadium at play. Of course if you choose to ignore home games, you will never notice the obvious. You also include last year, which is completely ridiculous. Last years offensive statistics are only slightly more relevant to this years team as the 1928 Yankees are to the 2011 Yankees. Who is left from last years regular position players? Headley, Veneble, and Donorfrio? What, aside from the uniform, leads you to believe that this years offensive performance is in any way a continuation of last years? The only reason to use stats across seasons would be to measure the impact of factors that remained constant across seasons, or the impact of a change made across seasons. You could at least justify your methodology if the key players were the same as last year, but that's not even true. You are just cherry picking, as you usually do. Why not start in 1998?
|
|
|
Post by 01aztecgrad on Jul 28, 2011 16:01:21 GMT -8
Someone is running up a hill and you're wondering if there was a tail wind. According to you, the way to account for one person running up hill, and one person running downhill, is to see how they both run on flat ground, regardless of other factors. So the fact that one person ran in Arizona in September into a headwind, while the other ran in New Mexico in April with a tailwind, is insignificant because the ground was flat? Good theory. I'm sure you'd get a lot of support for that.
|
|
|
Post by AztecBill on Jul 28, 2011 16:26:58 GMT -8
Someone is running up a hill and you're wondering if there was a tail wind. According to you, the way to account for one person running up hill, and one person running downhill, is to see how they both run on flat ground, regardless of other factors. So the fact that one person ran in Arizona in September into a headwind, while the other ran in New Mexico in April with a tailwind, is insignificant because the ground was flat? Good theory. I'm sure you'd get a lot of support for that. I am so glad you agree with me that running on flat ground is a much better way to measure than including the hills. Now the question of wind: Unless you are trying to verify a world record time in the 100 meter dash - that is going a bit over board. You may do it if you like. I will just throw out the hill data and call it a day. See my question, in a post above, about the Texas Rangers and please answer: What better rates their offense and defense versus other teams: 1. Their overall numbers 2. Their road numbers 3. Their home numbers
|
|
|
Post by AztecBill on Jul 28, 2011 16:31:16 GMT -8
That is done all the time. A simple example is surveys. Questions are asked and many many results are thrown out to get a balanced demographic representation of the population. Baseball is like a survey that has half the people describe themselves as Libertarians. Hell ya, you need to throw out a bunch of the data. Wrong. First of all, if you're conducting a survey in which you want the responses to match the population and you have heavily skewed responses, you probably didn't select your respondents at random, and you should rethink your study design. If you did butcher the first step, you wouldn't throw out surveys to get the proportion of responses you want. You would either: 1) Ask the demographic questions up front, and stop taking the surveys from people whose profile you no longer need. 2) Use all of the surveys to determine how libertarians respond, and adjust the weight you assign those responses to represent the proportion of libertarians in the population. Your philosophy of "when in doubt, throw it out" isn't encouraged in any statistics or forecasting class. Why don't you use park adjusted data to support your theory? Is it because it demonstrates that Petco is only a minor factor in explaining why the Padres are so horrible at home offensively? To any rational person, that would strongly suggest that there are factors much larger than the stadium at play. Of course if you choose to ignore home games, you will never notice the obvious. You also include last year, which is completely ridiculous. Last years offensive statistics are only slightly more relevant to this years team as the 1928 Yankees are to the 2011 Yankees. Who is left from last years regular position players? Headley, Veneble, and Donorfrio? What, aside from the uniform, leads you to believe that this years offensive performance is in any way a continuation of last years? The only reason to use stats across seasons would be to measure the impact of factors that remained constant across seasons, or the impact of a change made across seasons. You could at least justify your methodology if the key players were the same as last year, but that's not even true. You are just cherry picking, as you usually do. Why not start in 1998? I am perfectly OK with option #2 above. I will take the 81 games the Padres play at Petco Park and adjust the data to count them as only 12 games. Thanks for the idea. (although, between you and me, I have already done that in another post). And I used last year's numbers earlier to demonstrate the effects of Petco Park - not to rate this year's team. In fact, a stat I love using includes all road data since Petco Park opened. 7 years of runs per game on the road shows the Padres 4th in the NL in scoring on the road. Many Padre Fans can't believe it. Than I tell them we were 11th in ERA on the road over those 7 years. That surprises them too, because the general consensus is that we have always been a good pitching team with horrible offenses (since PC opened). But the truth is somewhat different but merely clouded by park effects.
|
|
|
Post by 01aztecgrad on Jul 28, 2011 16:37:57 GMT -8
I am perfectly OK with option #2 above. I will take the 81 games the Padres play at Petco Park and adjust the data to count them as only 12 games. Thanks for the idea. (although, between you and me, I have already done that in another post). Support it with science and you might have a point. You can't, just like you can't support any of your theories. Like I said, and you apparently concede, you have no basis for throwing out half of the sample. It's okay to admit you either don't understand applied statistics, or just don't believe what the data show and would rather go by the "eyeball test". Your "statistical analysis" is a fraud.
|
|
|
Post by AztecBill on Jul 28, 2011 16:39:17 GMT -8
I am perfectly OK with option #2 above. I will take the 81 games the Padres play at Petco Park and adjust the data to count them as only 12 games. Thanks for the idea. (although, between you and me, I have already done that in another post). Support it with science and you might have a point. You can't, just like you can't support any of your theories. Like I said, and you apparently concede, you have no basis for throwing out half of the sample. It's okay to admit you either don't understand applied statistics, or just don't believe what the data show, and would rather go by the "eyeball test". Your "statistical analysis" is a fraud. It is more a more accurate than overall data. You already admitted that. I find it funny that the same people who have no problem including 81 games at Petco Park in a rating, get all bent out of shape because the Padres don't play road games at Petco Park. Even though those road games would be a very small percentage of games. I noticed you didn't answer my question about the RangersI'll ask again. Below is some data about the Rangers team: ERA 4.39 Home #26 out of 30 teams 3.10 Road #1 out of 30 teams. 3.79 All Games #13 out of 30 teams. Scoring 4.04 Road #17 out of 30 teams (4.04 - just like the Padres) 6.07 Home #1 out of 30 teams. 5.10 All Games #3 out of 30 teams. What conclusions would you draw about their team from that data?
|
|
|
Post by AztecBill on Jul 28, 2011 17:01:36 GMT -8
I'll ask again. Below is some data about the Rangers team: ERA 4.39 Home #26 out of 30 teams 3.10 Road #1 out of 30 teams. 3.79 All Games #13 out of 30 teams. Scoring 4.04 Road #17 out of 30 teams (4.04 - just like the Padres) 6.07 Home #1 out of 30 teams. 5.10 All Games #3 out of 30 teams. What conclusions would you draw about their team from that data? I'll draw the conclusion that they were in the WS last year and that they have a decent chance to play in the postseason this year. I'll also infer that somehow you think if both the Padres and Rangers played at Petco, the Padres would have a better record than the Rangers. Further, I'll guess that you'd prefer the Padres' lineup to that of the Rangers. You are not drawing your conlcusions from the data. You are wrong on all your conclusions that involve me. Try again and just use the data listed to see if you can determine the make up of the Rangers team. Most people think they are a great hitting team with OK pitching. Does that Jive with what you see in the data?
|
|
|
Post by 01aztecgrad on Jul 28, 2011 18:04:03 GMT -8
It is more a more accurate than overall data. You already admitted that. I find it funny that the same people who have no problem including 81 games at Petco Park in a rating, get all bent out of shape because the Padres don't play road games at Petco Park. Even though those road games would be a very small percentage of games. I noticed you didn't answer my question about the RangersI'll ask again. Below is some data about the Rangers team: ERA 4.39 Home #26 out of 30 teams 3.10 Road #1 out of 30 teams. 3.79 All Games #13 out of 30 teams. Scoring 4.04 Road #17 out of 30 teams (4.04 - just like the Padres) 6.07 Home #1 out of 30 teams. 5.10 All Games #3 out of 30 teams. What conclusions would you draw about their team from that data? I find it funny that you've admitted that you have no justification to support the exclusion of home games, yet it's still the crutch of your "analysis". I don't care that the Padres don't play road games at Petco, I care that there is nothing to suggest that Petco is the primary factor in their inability to score runs at home. They've been outscored by the opponents 194 to 148 at home this year, in spite of outscoring opponents 215 to 199 on the road. If Petco were the primary cause of the variation, a simple linear regression would demonstrate that home/away explains much of the difference, but it doesn't, and you probably already know that. Unless you have a theory about why Petco has a disproportionate impact on the home team, there is nothing to suggest that something inherent in Petco is the primary cause of the Padres scoring 62 fewer runs at home, while their opponents only score 5 fewer (in 1 fewer game). If it were mostly Petco, then the impact on both teams would be predictable over time. You wouldn't have years where the Padres are greatly impacted and opponents aren't, or vice versa. The only explanation for swings like that are variables that you aren't measuring, which you don't seem to care about. According to you those unmeasured variables magically disappear on the road, even though you don't know what they are, or attempt to prove that they're immaterial. I didn't answer your question about the Rangers because I don't make statistical statements based on the "eyeball test". If you want me to make a guess, I'd guess that their overall ranking of #3 on offense is probably the upper bound, and they are within a couple places of where they should be ranked, but I say that without knowing how close the teams behind them are in terms of runs per game, and how much of a possible impact their home stadium has, even though I suspect that it's minor. I'd guess that the #13 ranking for ERA is the lower bound, and that they are likely within a couple places of where they should be ranked. I am admittedly pulling that theory out of my a$$, but unlike you, I don't try to claim there is a statistical justification for my claim. In actuality, in the absence of evidence to the contrary, their overall ranking is the most justified, and with evidence to the contrary, the correct approach would be to adjust the data rather than to exclude anything. All that is really just a long way of saying that your statistical analysis is a fraud. * I also didn't admit that excluding half of the sample is more accurate than the overall data. I said that if you can demonstrate that it has an impact, you are free to adjust how heavily it's weighted. You don't have any estimate of the impact, because you have no theory or evidence. You are going with the "eyeball test".
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Jul 28, 2011 20:51:58 GMT -8
Bill is getting his statistical clock cleaned.
|
|
|
Post by AztecBill on Jul 29, 2011 8:41:05 GMT -8
Bill prefers the Pads' lineup to that of the Rangers. I stated that your conclusions about me were wrong. So don't state I believe something that I already said I didn't.
|
|
|
Post by AztecBill on Jul 29, 2011 9:06:33 GMT -8
I find it funny that you've admitted that you have no justification to support the exclusion of home games, yet it's still the crutch of your "analysis". I have justification. I have stated it. I don't care that the Padres don't play road games at Petco, I care that there is nothing to suggest that Petco is the primary factor in their inability to score runs at home. They've been outscored by the opponents 194 to 148 at home this year, in spite of outscoring opponents 215 to 199 on the road. If Petco were the primary cause of the variation, a simple linear regression would demonstrate that home/away explains much of the difference, but it doesn't, and you probably already know that. Unless you have a theory about why Petco has a disproportionate impact on the home team, there is nothing to suggest that something inherent in Petco is the primary cause of the Padres scoring 62 fewer runs at home, while their opponents only score 5 fewer (in 1 fewer game). Opponents have scored more runs at Petco Park this year then they have earned. That is because a highly disproportionate number of Padres errors have happened at home. The Padres have given up 32 unearned runs so far at home, compared to 19 on the road. Last year for the entire season they gave up 21 UR at home. Defense (esp. at home) has been a big Padres Problem this year. If it were mostly Petco, then the impact on both teams would be predictable over time. You wouldn't have years where the Padres are greatly impacted and opponents aren't, or vice versa. The only explanation for swings like that are variables that you aren't measuring, which you don't seem to care about. According to you those unmeasured variables magically disappear on the road, even though you don't know what they are, or attempt to prove that they're immaterial. You don't need to determine all the causes of something. The numbers clearly show it exists. If you want to start a thread asking why Petco Supresses runs, I would be happy to participate. There have been threads like that in the past. You can search for them via google. You saying that if Petco were the cause of fewer runs than there would be no variation is a complete joke. There are only 162 games a year and only 81 home games. That is far too few to expect no variation. For no variation we would need games into the scores of thousands. I didn't answer your question about the Rangers because I don't make statistical statements based on the "eyeball test". If you want me to make a guess, I'd guess that their overall ranking of #3 on offense is probably the upper bound, and they are within a couple places of where they should be ranked, but I say that without knowing how close the teams behind them are in terms of runs per game, and how much of a possible impact their home stadium has, even though I suspect that it's minor. I'd guess that the #13 ranking for ERA is the lower bound, and that they are likely within a couple places of where they should be ranked. I am admittedly pulling that theory out of my a$$, but unlike you, I don't try to claim there is a statistical justification for my claim. In actuality, in the absence of evidence to the contrary, their overall ranking is the most justified, and with evidence to the contrary, the correct approach would be to adjust the data rather than to exclude anything. So how do you explain tha they are the #1 team in road ERA in baseball? You can't except that maybe the Rangers have the best pitching in baseball but their home ball park (which has 4 runs a game more than Rangers games on the road) may hide that fact? All that is really just a long way of saying that your statistical analysis is a fraud. No it doesn't. It proves that you don't know the history of Petco park and don't understand Park Effects. * I also didn't admit that excluding half of the sample is more accurate than the overall data. I said that if you can demonstrate that it has an impact, you are free to adjust how heavily it's weighted. You don't have any estimate of the impact, because you have no theory or evidence. You are going with the "eyeball test". Here educate yourself about park effects. "Petco Park remains the most difficult environment in MLB in which to score runs, and by a wide margin."
|
|
|
Post by AztecBill on Jul 29, 2011 9:23:28 GMT -8
Adrian Gonzales is an interesting study. (numbers are BA,OBP,SLG,OPS)
AGON 3 year data (2008-2010) .285 .387 .523 .910 Overall .310 .390 .599 .989 Road .257 .384 .439 .823 Home
I was in discussion with some Red Sox fans about AGon and park effects. He will get to replace his below average (for a first baseman) numbers at Petco Park, with above average (for him) home numbers due to playing his home games at Fenway Park - a hitters park. They expected his overall numbers to be about the same. I expected his road numbers to understate his new results.
2011 AGon stats .352 .414 .560 .974 Overall .397 .450 .593 1.043 Home .308 .378 .526 .904 Road
While it is only part of a season. It is pretty telling. He is putting up home numbers that he never even got close to in San Diego.
AGon Road Numbers .308 .378 .526 .904 Road 2011 .310 .390 .599 .989 Road (2008-2010)
AGon Home Numbers .397 .450 .593 1.043 Home 2011 .257 .384 .439 .823 Home (2008-2010)
So are you going to say that his quad slash line at home this year is a fluke? And that his, well below average, quad slash line at Petco Park over the past 3 years is just a natural fluxuation in data?
|
|
|
Post by AztecBill on Jul 29, 2011 11:05:24 GMT -8
Does it effect opponents' hitting? Below is all teams home ERAs over the 7 years the Padres have been playing in Petco Park. Opposing hitters have been the worst in Petco Park, over that time. No effect? Hardly. Team | 2010 | 2009 | 2008 | 2007 | 2006 | 2005 | 2004 | Average ERA | San Diego | 2.9 | 3.44 | 3.65 | 3.02 | 3.75 | 3.52 | 3.85 | 3.45 | St. Louis | 2.92 | 3.44 | 4.06 | 4.13 | 3.93 | 3.44 | 3.54 | 3.64 | LA Dodgers | 3.81 | 3.1 | 3.01 | 4.24 | 4.11 | 3.94 | 3.71 | 3.70 | NY Mets | 3.07 | 3.98 | 3.78 | 4.2 | 3.76 | 3.45 | 3.73 | 3.71 | Atlanta | 3.18 | 3.43 | 4.44 | 3.96 | 4.33 | 3.82 | 3.59 | 3.82 | Houston | 3.51 | 4.01 | 4.25 | 4.05 | 4.03 | 3.07 | 3.91 | 3.83 | San Francisco | 3.07 | 3.27 | 4.45 | 4.01 | 4.35 | 4.22 | 4.4 | 3.97 | Milwaukee | 4.5 | 4.42 | 3.49 | 4.01 | 4.45 | 3.76 | 4.09 | 4.10 | Chicago Cubs | 4.48 | 3.94 | 3.77 | 4.19 | 4.68 | 4.04 | 3.91 | 4.14 | Florida | 3.88 | 4.47 | 4.3 | 4.78 | 4.07 | 3.82 | 3.74 | 4.15 | Washington | 3.97 | 4.67 | 4.56 | 4.08 | 4.66 | 3.56 | 4.11 | 4.23 | Philadelphia | 3.47 | 4.29 | 3.65 | 4.78 | 4.7 | 4.48 | 4.31 | 4.24 | Pittsburgh | 4.54 | 4.01 | 4.52 | 4.55 | 4.08 | 4.24 | 3.8 | 4.25 | Cincinnati | 3.98 | 4.06 | 4.51 | 4.93 | 4.74 | 5.16 | 4.74 | 4.59 | Arizona | 4.53 | 4.74 | 4.08 | 4.11 | 4.81 | 5.27 | 4.71 | 4.61 | Colorado | 4.25 | 4.41 | 4.83 | 4.34 | 4.72 | 5.18 | 6.27 | 4.86 |
|
|
|
Post by AztecBill on Jul 29, 2011 13:58:13 GMT -8
I added to the previous table the road ERAs of the teams and sorted by magnitude of being a pitchers park. Team | 2010 | 2009 | 2008 | 2007 | 2006 | 2005 | 2004 | Average ERA | Road ERA | Diff | San Diego | 2.9 | 3.44 | 3.65 | 3.02 | 3.75 | 3.52 | 3.85 | 3.45 | 4.56 | 1.11 | Pittsburgh | 4.54 | 4.01 | 4.52 | 4.55 | 4.08 | 4.24 | 3.8 | 4.25 | 5.16 | 0.91 | Houston | 3.51 | 4.01 | 4.25 | 4.05 | 4.03 | 3.07 | 3.91 | 3.83 | 4.57 | 0.74 | NY Mets | 3.07 | 3.98 | 3.78 | 4.2 | 3.76 | 3.45 | 3.73 | 3.71 | 4.44 | 0.73 | St. Louis | 2.92 | 3.44 | 4.06 | 4.13 | 3.93 | 3.44 | 3.54 | 3.64 | 4.34 | 0.70 | LA Dodgers | 3.81 | 3.1 | 3.01 | 4.24 | 4.11 | 3.94 | 3.71 | 3.7 | 4.29 | 0.59 | Milwaukee | 4.5 | 4.42 | 3.49 | 4.01 | 4.45 | 3.76 | 4.09 | 4.1 | 4.69 | 0.59 | Washington | 3.97 | 4.67 | 4.56 | 4.08 | 4.66 | 3.56 | 4.11 | 4.23 | 4.81 | 0.58 | Florida | 3.88 | 4.47 | 4.3 | 4.78 | 4.07 | 3.82 | 3.74 | 4.15 | 4.54 | 0.39 | Atlanta | 3.18 | 3.43 | 4.44 | 3.96 | 4.33 | 3.82 | 3.59 | 3.82 | 4.20 | 0.38 | San Francisco | 3.07 | 3.27 | 4.45 | 4.01 | 4.35 | 4.22 | 4.4 | 3.97 | 4.25 | 0.28 | Cincinnati | 3.98 | 4.06 | 4.51 | 4.93 | 4.74 | 5.16 | 4.74 | 4.59 | 4.71 | 0.12 | Philadelphia | 3.47 | 4.29 | 3.65 | 4.78 | 4.7 | 4.48 | 4.31 | 4.24 | 4.26 | 0.02 | Chicago Cubs | 4.48 | 3.94 | 3.77 | 4.19 | 4.68 | 4.04 | 3.91 | 4.14 | 4.05 | -0.09 | Arizona | 4.53 | 4.74 | 4.08 | 4.11 | 4.81 | 5.27 | 4.71 | 4.61 | 4.43 | -0.18 | Colorado | 4.25 | 4.41 | 4.83 | 4.34 | 4.72 | 5.18 | 6.27 | 4.86 | 4.50 | -0.36 |
The Padres had the best home ERA but gave up more than a run a game more on the road. Note: home ERAs tend to be better. In the NL over that timeframe, home teams scored 0.4 runs more per game.
|
|
|
Post by 01aztecgrad on Jul 29, 2011 15:03:15 GMT -8
I have justification. I have stated it. You don't need to determine all the causes of something. The numbers clearly show it exists. If you want to start a thread asking why Petco Supresses runs, I would be happy to participate. There have been threads like that in the past. You can search for them via google. I never said you have to determine all of the causes. You have to determine the significant causes. You don't, and never have. You just throw out all of the data from half of the sample because you don't like what the data show. You need scores of thousands of games to have an adequate sample? Where did you learn that? You know even less about statistics than I thought. It's very rare that you need "scores of thousands" of observations of anything to demonstrate the strength of a relationship with a high degree of certainty. When there is a large variation from year to year, even with only 81 observations a year, it means that there are significant factors aside from the stadium at work. Those need to be measured, not ignored or hidden. I understand the theory of the Park Effect. Unfortunately you don't understand that it completely discredits your assumption that you can just throw out home games. The entire reason to try to determine the park effect is to find out the impact of the stadium to allow for adjustment of data. I already told you that you are free to adjust the statistics at Petco if you have a valid theory and evidence. You don't. You just want to throw out all of the data. Just so you know, you would also have to adjust the stats for every other team in every other stadium in addition to adjusting the Padres. Consistency is somewhat important in statistics. Just like when you posted the thread trying to prove that Trevor Hoffman performed better than Mariano Rivera with inherited runners, you are again posting a link to analysis you don't understand. Maybe later I'll go through the website and let you know if there are any problems with this guys analysis, as I did with the other website you posted that was discredited in about 5 minutes. Once again, your analysis is a fraud.
|
|