Talk:Pythagorean expectation

Statistics Low‑importance

	This article is within the scope of WikiProject Statistics, a collaborative effort to improve the coverage of statistics on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.StatisticsWikipedia:WikiProject StatisticsTemplate:WikiProject StatisticsStatistics articles
Low	This article has been rated as Low-importance on the importance scale.

Baseball Low‑importance

	Baseball portal This article is within the scope of WikiProject Baseball, a collaborative effort to improve the coverage of baseball on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.BaseballWikipedia:WikiProject BaseballTemplate:WikiProject BaseballBaseball articles
Low	This article has been rated as Low-importance on the project's importance scale.

Clutch

I am curious if anyone has done any serious analysis as to what one can determine about "clutch" performance based on this concept of Pythagorean expectation. There ought to be a theoretical "correct" result if you assume there is no such thing as "clutch" performance. Observing how the actual results differ from the expectation might point us towards certain 'clutch' behavior to look at, if they exist. -- 24.160.126.217 04:34, 18 May 2005 (UTC)[reply]

But there are also other factors that lead to differences between pythag projections and actual team performance, such as bullpen depth, manager decisions, and just plain luck. Javaisfun 04:02, 18 January 2007 (UTC)[reply]

See also the note below on "Systematic Deviations"; I believe it answers the question. Eric Walker (talk) 11:33, 23 November 2009 (UTC)[reply]

Question about Runs Allowed

Hi, I have entirely no expierience with professional baseball, and the term 'runs allowed' makes little sense to me, could someone who would knows what this means clarify this on the page, or create a link to an appropriate page. Thanks. --Neo 21:27, 4 July 2006 (UTC)[reply]

It's the number of runs scored by the other team. I think there's also a technical definition that might differ from this in some instances, used in statistics applied only to pitchers rather than to teams as a whole. Michael Hardy 02:11, 5 July 2006 (UTC)[reply]

Greatest differentials?

Does anyone know what the biggest differential ever is between Pythag expectation and actual record? The Indians were -12 (Pythag = 90 wins, actual = 78 ) last year, which is impressive, but my guess is that there are far more ridiculous numbers. Wickethewok 02:41, 10 February 2007 (UTC)[reply]

Using the Lahman Databse (highly recommended), I calculated the pyth% for every team since 1950. Here are the teams who differed most from their expected wins... (Note: I used the "Pythagenpat" exponent, not 2 or 1.83.)

year Team lg G W L RS RA Pyth% PythW Diff
------------------------------------------------------------------------------------------------------
1993 New York Mets NL 162 59 103 672 744 0.453 73.4 14.38
1986 Pittsburgh Pirates NL 162 64 98 663 700 0.475 77.0 12.97
1975 Houston Astros NL 162 64 97 664 711 0.469 75.9 11.91
1984 Pittsburgh Pirates NL 162 75 87 615 567 0.536 86.8 11.79
2005 Arizona Diamondbacks NL 162 77 85 696 856 0.403 65.2 11.75
2004 New York Yankees AL 162 101 61 897 808 0.551 89.2 11.75
1984 New York Mets NL 162 90 72 652 676 0.484 78.3 11.67
1967 Baltimore Orioles AL 161 76 85 654 592 0.544 87.7 11.66
2006 Cleveland Indians AL 162 78 84 870 782 0.551 89.3 11.34
1954 Brooklyn Dodgers NL 154 92 62 778 740 0.524 80.7 11.30
1970 Cincinnati Reds NL 162 102 60 775 681 0.560 90.7 11.26
1972 New York Mets NL 156 83 73 528 578 0.461 71.8 11.15
1955 Kansas City Athletics AL 155 63 91 638 911 0.335 51.9 11.10
1999 Kansas City Royals AL 161 64 97 856 921 0.464 74.7 10.67
1993 San Diego Padres NL 162 61 101 679 772 0.440 71.3 10.34
1961 Cincinnati Reds NL 154 93 61 710 653 0.539 83.0 10.01
1970 Chicago Cubs NL 162 84 78 806 679 0.580 93.9 9.95
1980 St. Louis Cardinals NL 162 74 88 738 710 0.518 83.9 9.92
1972 San Francisco Giants NL 155 69 86 662 649 0.509 78.9 9.91
1997 San Francisco Giants NL 162 90 72 784 793 0.495 80.1 9.88
1966 New York Yankees AL 160 70 89 611 612 0.499 79.9 9.88
1955 Detroit Tigers AL 154 79 75 775 658 0.577 88.8 9.80
2004 Cincinnati Reds NL 162 76 86 750 907 0.409 66.2 9.77
1955 Cincinnati Redlegs NL 154 75 79 761 684 0.550 84.7 9.75
1953 New York Giants NL 155 70 84 768 747 0.513 79.6 9.56
2001 Colorado Rockies NL 162 73 89 923 906 0.509 82.5 9.50
1977 Baltimore Orioles AL 161 97 64 719 653 0.544 87.6 9.38
1981 Cincinnati Reds NL 108 66 42 464 440 0.524 56.6 9.37
1974 San Diego Padres NL 162 60 102 541 830 0.313 50.7 9.31
1997 Houston Astros NL 162 84 78 777 660 0.575 93.2 9.22
1962 New York Mets NL 161 40 120 617 948 0.306 49.2 9.19
1978 Cincinnati Reds NL 161 92 69 710 688 0.515 82.8 9.15
1998 Kansas City Royals AL 161 72 89 714 899 0.391 62.9 9.10
1972 Baltimore Orioles AL 154 80 74 519 430 0.578 89.1 9.06
1962 St. Louis Cardinals NL 163 84 78 774 664 0.571 93.0 9.04

Cleveland 2006 was the 9th-biggest differential since 1950. They had the most expected wins of any team in the top 10, however. - Goo Paine 04:23, 10 February 2007 (UTC)[reply]

"Systematic Deviations"

Re: There are some systematic statistical deviations between actual winning percentage and expected winning percentage, which include bullpen quality and luck. That there are deviations from anything but luck is an unproven and probably incorrect assertion.

Over the past 44 years, the percentages of teams with, as of that year, a streak of any given length are these:

Streak     Actual     Chance
Length      Pct.       Pct.
----------------------------
   12:     0.00%      0.03%
   11:     0.09%      0.05%
   10:     0.09%      0.10%
    9:     0.44%      0.20%
    8:     0.55%      0.39%
    7:     1.35%      0.78%
    6:     2.15%      1.56%
    5:     4.37%      3.13%
    4:     8.42%      6.25%
    3:    15.14%     12.50%
    2:    28.61%     25.00%
    1:    56.06%     50.00%

That does not seem to make much of a case for anything but chance deviations.

(The period 44 years is because the database I used runs from 1955 on, that being the first year certain stats, such as IBB, began to be officially kept; those are not relevant here, but explain the years used--1966 was the first year in which a 12-year streak could exist in those data. Since a team has a 50% chance of being over or under in a given season, "streak" probabilities are simply powers-of-2 divisors.)

That being so, I urge a change to something like: There are inevitable deviations between actual winning percentage and expected winning percentage owing to chance, but there do not appear to be systematic factors (that is, factors attributable to the characteristics of particular teams); the incidence of "streaks" of under-performing and over-performing predictions varies little from that expected from chance.

I will wait to see if there is any comment here before cutting in such a change.

Eric Walker (talk) 11:31, 23 November 2009 (UTC)[reply]

2010 Example

I don't like the 2010 example. I don't think it makes sense to track running totals of the difference from expectation throughout the year. What is that supposed to show? Plus, a three-game difference in Pythagorean expectation between the two teams is not that large.DavidRF (talk) 03:56, 28 April 2012 (UTC)[reply]

Luck

"There are some systematic statistical deviations between actual winning percentage and expected winning percentage, which include bullpen quality and luck"

Why is an article about statistics claiming luck is systematic? I'm no expert but this strikes me as wrong. 24.50.189.182 (talk) 05:47, 24 August 2019 (UTC)[reply]