Years ago (25-30, maybe), the average winning score on any particular day, week, season was 5. The losing was 3. Memory serves, these were long-standing numbers.
I just ran the numbers for September 3, 4 & 5 of this year, a total of 46 games. The average winning score was 5.6 and the losing 2.6.
Seems odd to me in that a higher winning average indicates better hitting, but a lower losing average indicates better pitching.
Anyone know when these averages changed? How about an explanation of what’s behind the changes?
When you are comparing things statistically, if you just use “averages” you are doing it wrong. It is entirely possible for you to get a 5.6 to 2.6 average over 3 days, with an underlying distribution of scores that has an expected value of 5-3. It could be that the things have changed (and a possible explanation is that winning teams have gotten better and losing teams have gotten worse) but it also possible that nothing has, and to figure out if it has you need information on the spread of scores.
It’s very unlikely the average winning and losing scores were 5 and 3 years ago. It’s much more likely they were around 5 and 3 and you’re just remembering round numbers. (If the averages were reported with one decimal, then there’s presumably only a 10% chance the digit after the decimal point was a zero so only a 1% chance both were.) Maybe the averages were 5.4 and 2.8 which are pretty close to the numbers you got.
In any case with only 46 games this weekend, your averages are much less precise. I get average winning scores of 5.6 and 2.7 with standard deviations of 3.0 and 1.9. Now these numbers are clearly not normally (Gaussian) distributed but those standard deviations indicate your one-weekend means would be in the range 2.6 to 8.6 and 1.1 to 4.9 (+ or - one standard deviation) only 2/3rds of the time.
Just to point out the variation in averages over a short time note that the Mets won 18-5 on Sunday. 18 is a huge outlier of a score; the next highest score was 12. There were also 6 shutouts during those three days which I believe is also quite unusual, but haven’t checked leading to low losing scores.
What you point out simply wouldn’t be unusual over a three day period even if your remember numbers are exactly right.
I don’t know how accurate your stats really are, but one explanation is expansion plus the lack of a salary cap. Expansion allows more players to become major leaguers who otherwise would not. The lack of a salary cap allows for rich teams to pile up the talent while poor teams get stuck with the guys who, without expansion, wouldn’t even be major leaguers. The talet discrepency from team to team now is greater than it was 30 years ago.
**Small sample size. **46 games is going to allow for a great deal of variation.
Confirmation bias. Like others have said, you’re remembering whole numbers from decades ago and comparing them to more accurate numbers from this weekend.
When did averages become predictors? Answer: Never.
I just looked up the average high temperature for 9/7 in my area, and the average high temperature is 79. I looked over several decades worth of info for 9/7, and the high temp was 79 only twice in 30 years.