First 2005 Post Mortem

[MFS62]

Well if the hitting is "ok" and the pitching/ Team ERA (top 4 in the NL last time I looked) is "ok", does this mean that we now will re-engage in one of our favorite activities, a discussion of the concept of "clutch"?

Later

[Elster88]

]Neither is the cherry-picking argument (if we take out the 5 best games, etc.).

That's not cherry-picking. Basic statistics. The majority of scores fall within a certain range, and you remove the outliers. What you have left AFTER removing the outliers tells you what a score is likely to be on a certain day.
____________________________
This post had the designation 159) Bob Shaw

[nymr83]

]Shingo Takatsu

how can you put him in the bad category? he threw one bad pitch to one batter because dumbass randolph brought him in (after having not seen a major league batter in over a month) with the bases loaded to face miguel cabrera. that was bad managing not bad player.

[Guest Rotblatt]

Elster88 wrote: ]Neither is the cherry-picking argument (if we take out the 5 best games, etc.). That's not cherry-picking. Basic statistics. The majority of scores fall within a certain range, and you remove the outliers. What you have left AFTER removing the outliers tells you what a score is likely to be on a certain day. ____________________________ This post had the designation 159) Bob Shaw

Um, randomly deciding to remove 5 games--with the number 5 picked completely abitrarily--has nothing whatsoever to do with outliers or statistics.

You don't just say, "I bet there are five outliers!" and get rid of them. You define terms--what offensive output am I measuring? Runs scored? Earned runs scored? Batters reached? OPS? Then you plot all the games out and look for a pattern, from which you find outliers.

So yes, it is cherry picking. Which is fine and still fun to look at, but don't pretend it's statistical.

[Guest ABG]

So...shut down Pedro the rest of the year?

Yes, says I.

[Guest Johnny Dickshot]

I am the last man on Earth you would trust with *performing* statistical analysis, but just for fun, on paper, I took down all the Met score occurances this year.

0 runs
10 times

1 run
15

2 runs
17

3 runs
21

4 runs
17

5 runs
13

6 runs
15

7 runs
8

8 runs
4

9 runs
9

10 runs
4

11 runs
0

12 runs
3

13 runs
0

14 runs
1

15 runs
0

16 runs
1

17 runs
0

18 runs
1

Add it up:
Mean runs: 4.51
Mode: 3
EDIT: Median 4.38

Even though we "average" 4.5 runs a game, in more than half of our games this year we have scored 4 runs or fewer and we're most likely to put up 3 runs on any given date.

What this says to me is that most nights, this is NOT a strong offensive team.

[Elster88]

Rotblatt wrote: ="Elster88"] ]Neither is the cherry-picking argument (if we take out the 5 best games, etc.). That's not cherry-picking. Basic statistics. The majority of scores fall within a certain range, and you remove the outliers. What you have left AFTER removing the outliers tells you what a score is likely to be on a certain day. ____________________________ This post had the designation 159) Bob Shaw Um, randomly deciding to remove 5 games--with the number 5 picked completely abitrarily--has nothing whatsoever to do with outliers or statistics. You don't just say, "I bet there are five outliers!" and get rid of them. You define terms--what offensive output am I measuring? Runs scored? Earned runs scored? Batters reached? OPS? Then you plot all the games out and look for a pattern, from which you find outliers. So yes, it is cherry picking. Which is fine and still fun to look at, but don't pretend it's statistical.

I don't think you know what I mean. You find the range under which the majority of the scores fall. If there are some that are outside that range, you discard them. Those that are going to be discarded (if any), are going to be the extreme highs and exterme lows. This is not cherry-picking. I'm not sure how to explain it better without actually doing the math. I'll get around to doing this out one day.

I was referring to runs scored, since that's much easier to use than figuring out the OPS for each individual game.

]Um, randomly deciding to remove 5 games--with the number 5 picked completely abitrarily--has nothing whatsoever to do with outliers or statistics. You don't just say, "I bet there are five outliers!" and get rid of them.

When did I say this? When did I pick the number 5? I don't know if there are five, less than five or more than five, and never said I did.

I can tell you without doing a single calculation that the two games in Arizona are meaningless when trying to figure out the Mets' average offensive production, but those are the only two I am sure of offhand. As an educated guess, I would say there are more than five. Possibly some of the times we were shutout will be eliminated, too, though those are much more likely to be included.

]but don't pretend it's statistical.

Yes, it is statistics. It is a way of removing error and reducing the influence of extremes from a measure of average. This is a way (obviously not the only way) to figure out what the average offensive output was for a team over a season. I would argue that it is better than simply taking the runs scored as a total. And for me it's more fun than looking up numbers on a website, which is why I plan to do the math out.
____________________________
This post had the designation 159) Bob Shaw

[Guest Johnny Dickshot]

Hey! My little chart is fun! Did you know the Mets score 3 runs or fewer in 45% of their games?!?

[Elster88]

I like your chart, though in a nutshell that's what I was going to do, only using some fancy terms and more specific calculations I learned in one of those boring "stats as a scientific measure" classes way back in freshman year. Give me some of my thunder back.
____________________________
This post had the designation 159) Bob Shaw

[Elster88]

ABG wrote: So...shut down Pedro the rest of the year? Yes, says I.

I like this idea. Let him go back to that beach and sleep for three days or whatever he keeps talking about.
____________________________
This post had the designation 159) Bob Shaw

[Guest Johnny Dickshot]

I did the hard part. Maybe you can make a chart or something (I tried, only to become frustrated that excel doesnt recognize "0" as a variable or something.

I never use excel.

[nymr83]

]You find the range under which the majority of the scores fall. If there are some that are outside that range, you discard them. Those that are going to be discarded (if any), are going to be the extreme highs and exterme lows. This is not cherry-picking.

its cherry-picking because you are removing the best games without removing the worst ones. nowhere did i see out 5 worst games removed as well. saying that our blowouts dont count but getting blown out does isn't fair.

[Guest Johnny Dickshot]

Let's move on from taking away the blowouts (which came up merely as a quick and dirty way to examine the team's scoring).

Those scores are included in the figures I did on the other page, but I think it's easy to see their effect in pumping up the total and mean average scores.

But we still score 4 or fewer in more than half our games and most often score 3.

[Guest Rotblatt]

That's a good start!

Nymr's right--you have to remove ALL outliers. The problem is that we've been shut out 10 times, meaning we have to "own" those games. Which is why it would make sense to try and measure some other offensive output per game--maybe runners reached or total bases or something--preferably with runs as a heavily weighted variable, so that those truly anonomous games (where we hit a ton and score a lot of runs or don't hit at all and don't score at all) ALL get removed.

Rather than do that, though, which seems awfully complicated to me, I'm going to do what JD did with one or more teams (we'll see how long it takes!) so we can compare.

I'm starting with the Phillies and will probably stick with the NLE and maybe Houston.

[Elster88]

The reason you may not remove the shutouts is because they may fall within the range which the majority of the scores fall.

Example:

0, 0, 1, 2, 3, 15

The range here would obviously be 0 to 3, so you remove 15 only.

Obviously when the numbers are crunched for the Mets the zeroes may get removed, they may not.

[Guest Rotblatt]

Okay, so the Phillies average 4.61 Runs/Game. Their Mode is 1. In 53.1% of their games, they scored 4 runs or fewer (compared to 57.5% for us), and, if we're going by mode, they're most likely to score 1 run than 4 on any given date.

Mode's pretty weak, though, so I'd just as soon ignore it.

Anywya, by your criteria, the Phillies aren't much more consistent offensively than we are. Unless I'm missing something.

On to the Marlins.

[Guest Johnny Dickshot]

Well, a 3% advantage in the "4 runs or more" would equal 4 games (that is 3% of 139 games). The Phils have 5 more wins than us.

[Guest Rotblatt]

Marlins average 4.48 runs per game. In 56% of their games, they scored 4 runs or fewer.

]Well, a 3% advantage in the "4 runs or more" would equal 4 games (that is 3% of 139 games). The Phils have 5 more wins than us.

So that 3% defecit is enough to take us from a good offense to a bad one, in your estimation? Remember, I'm only trying to show that we're an average offensive team.

[Guest Johnny Dickshot]

Astros
140 games, 592 RS

Mean: 4.22
Median: 4.55
Mode: 4
4 runs or fewer: 57%
3 runs or fewer: 41%

[Guest Johnny Dickshot]

]So that 3% defecit is enough to take us from a good offense to a bad one, in your estimation?

It is my intuitive suspicion that the Mets "average" offense most nights is actually below average, but let's try and step away from what "we're trying to prove" and see what the numbers show us.

[Guest Johnny Dickshot]

Nationals
142 games, 554 runs

Mean: 3.9
Median: 3.0
Mode: 3
4 runs or fewer: 66%
3 runs or fewer: 50%

[Guest Rotblatt]

This is kind of grueling for me, so I'm taking a break. I keep fucking up and then having to figure out where I did.

So far, I'm not convinced that there's anything significant in the differences, but let's see what you come up with!

[Guest Johnny Dickshot]

Braves
141 games, 674 runs

Mean: 4.78
Median: 4.95
Mode: 5
4 runs or fewer: 50%
3 runs or fewer: 35%

[Guest Johnny Dickshot]

Somebody better than me at math has to apply a standard deviation to these figures, which would take care of the outliers by showing how far out of the standard bell curve they lie.

I'm taking the rest of the day off too.

[Guest Johnny Dickshot]

of the completed teams so far:

% of games scoring 4 runs or fewer:

Braves: 50%
Phillies: 53.1%
Marlins: 56%
Astros: 57%
Mets: 57.5%
Nationals: 66%

[Guest Rotblatt]

We should also see how closely that percentage correlates to RS/game.

Braves: 50% (674 RS, 4.78 R/G)
Phillies: 53.1% (658 RS, 4.67 R/G)
Marlins: 56% (634 RS, 4.50 R/G
Astros: 57% (592 RS, 4.23 R/G)
Mets: 57.5% (631 RS, 4.48 R/G)
Nationals: 66% (554 RS, 3.90 R/G)

Generally, it looks like teams that have scored more runs have more a higher percentage of games where they've scored more 4 runs or more.

[Guest Rotblatt]

Cubbies:

60.3% of four runs or fewer. They average 4.48 R/G--same as the Mets.

edited for typo

[Guest Rotblatt]

Milwaukee

55.3% at 4 runs or fewer. 4.51 Runs per game.

San Diego

57.1% (4.29 R/G)

So far:

Braves: 50% (4.78 R/G)
Phillies: 53.1% (4.67 R/G)
Brewers: 55.3% (4.51 R/G)
Marlins: 56% (4.50 R/G)
Astros: 57% (4.23 R/G)
San Diego: 57.1% (4.29 R/G)
Mets: 57.5% (4.48 R/G)
Cubs: 60.3% (4.48 R/G)
Nationals: 66% (3.90 R/G)

Well, that makes 2 teams worse in RS who have scored 5 runs or more more often than the Mets, with only 1 team better or as good as them with a worse percentage.

That being said, the difference to me seems pretty slight. Marlins with a 1.5% advantage (2 games), Brewers with a 3 game advantage, but both of them have scored slightly more runs overall than we have, so it's not too surprising. Astros with a slight advantage (under a game) but given how few runs they score, it's a suprise.

Cubbies have been really inconsistent and hit better at home, so no real surprise there.

Nats are just a bad team offensively, so again, no surprise.

And I'm done for the night.

[Guest Edgy DC]

It's just easier, cheaper, and less risky to first look at upgrading the bad positions toward average, instead of the average positions toward excellent. Excellence is also the occasional residue of mere goodness.

That said, Cliff Floyd's disappearance today definitely makes me re-think where the Mets are on the plus side of average.

[Guest Spacemans Bong]

="Edgy DC"]It's just easier, cheaper, and less risky to first look at upgrading the bad positions toward average, instead of the average positions toward excellent. Excellence is also the occasional residue of mere goodness. That said, Cliff Floyd's disappearance today definitely makes me re-think where the Mets are on the plus side of average.

He only disappeared today for you?