Evaluating Your Team - Post Facto

[PotKettleBlack]

Does anyone take time to review what worked, what failed, what meh'd and what excited?

If so, anything advanced to take it apart, or just looking at the bats, the pitchers and calling it a day.

Curious as I am putting together my first review of a large number of teams to see what I f'ed up.

[sschu]

Per my previous mail, suggest you look at the following:

Run Efficiency
WP in division
Runs/HR
ER/UER percent
Errors
Blown Saves
Record in 1 run games

If you are really into it you can average OPS (park adjusted) for league pitchers and your hitters and project what you think your team OPS will be. In my experience, OBP and HR rate will be pretty accurate, SLG seems to vary a bit. Then you can evaluate if your team performed per your predictions.

Best of luck, sschu

[supertyphoon]

The lack of fielding stats makes the task of finding out what worked and what didn't much harder. There's a handy stat called defensive efficiency that is useless here which I looked at all the time at another historical baseball sim website. We all know a good defense will help prevent runs scored, but in the final analysis it remains an unknown quantity.

[PotKettleBlack]

[quote:65299b935e="nexrad"]The lack of fielding stats makes the task of finding out what worked and what didn't much harder. There's a handy stat called defensive efficiency that is useless here which I looked at all the time at another historical baseball sim website. We all know a good defense will help prevent runs scored, but in the final analysis it remains an unknown quantity.[/quote:65299b935e]

I have a team that I think could be useful in the evaluation. I would have to dig it out, but about 50+ games in, I subbed out the old Stanky at 2B for a platoon of 1 range guys who hit more, but did not get on base nearly as much. Team fortune turned massively, but I'm not sure if it was enough. I would have to find it to be sure.

[PillPop]

[quote:08d4af8713="sschu"]Per my previous mail, suggest you look at the following:

Run Efficiency
WP in division
Runs/HR
ER/UER percent
Errors
Blown Saves
Record in 1 run games

If you are really into it you can average OPS (park adjusted) for league pitchers and your hitters and project what you think your team OPS will be. In my experience, OBP and HR rate will be pretty accurate, SLG seems to vary a bit. Then you can evaluate if your team performed per your predictions.



Best of luck, sschu[/quote:08d4af8713]

What is Run Efficiency? Is that the percentage of runs per baserunner, something like that?

[sschu]

Run Eff is the measure of how effective hits and OBP are in creating runs.

OBP * SLG * AB = Expected Runs Scored

Expected RS / Actual RS = Runs Eff

You should expect mid 90%s for a good team, if you are below 90%, then you are unlucky or do not have the right formula.

Very fast teams may have an Run Eff above 100$, HR teams will typically be in the low 90s.

sschu

[PotKettleBlack]

[quote:91d1e5d866="sschu"]Run Eff is the measure of how effective hits and OBP are in creating runs.

OBP * SLG * AB = Expected Runs Scored

Expected RS / Actual RS = Runs Eff

You should expect mid 90%s for a good team, if you are below 90%, then you are unlucky or do not have the right formula.

Very fast teams may have an Run Eff above 100$, HR teams will typically be in the low 90s.

sschu[/quote:91d1e5d866]

Whoa. Slow down there, hoss.

First, shouldn't it be Actual Runs / Expected Runs. Example:
http://fantasygames.sportingnews.com/stratomatic/team/team_other.html?user_id=331178

Team scored 1035 actual runs. Team put up 1065.5 expected runs. By going Exp/Act, we get 102.9%. By going the other way, 97.1%. That seems to state it better, that we didn't score as many as we ought to have scored.

Interesting idea, comparing what your expected run production should be to what your actual was. Not sure what I am to take away from 97.1% Run Efficiency. Bad luck? Wrong formula? Correction to the formula might be something more along the lines of switching out a neg clutch guy for a positive clutch guy maybe? Or, do I worry about it since we won 95 games with a Pythagorean of 94?

I'm gonna work through the rest of the tips with this team and another. Will share what I find and any questions, if that's cool.

...
Not clear on what R/HR means, especially in a neutralish park (3 Rivs) and a neutral style division (theme league).
...
About 9% of runs allowed were unearned. Based on my win shares analysis of the team, about 44% of the total winning was due to pitching/defense (considering the build, not terribly surprising) and the split within that 44.4% was 75.3% pitch and 24.7% fielding. This is a departure from the other team I've done this analysis on (67.7% pitch. 32.3% field), which is strange, as I think this was a better fielding team in general. But that might be showing up in the pitching stuff rather than the fielding, even though this process is supposed to tease the two apart.

[PillPop]

Please do share, PKB. I'm often at a loss as to why some-- hey, MOST-- of my teams fall short and would appreciate any tools for future analysis.

[PotKettleBlack]

[quote:ed64f25102="PillPop"]Please do share, PKB. I'm often at a loss as to why some-- hey, MOST-- of my teams fall short and would appreciate any tools for future analysis.[/quote:ed64f25102]

Cleaning up the sheet and do not have a way of assigning blame/shame/fame to individual fielders as of yet.
Have a reasonable working of assigning WS to bats versus the other side of the game.
Have a reasonable approximation of breaking fielding from pitching.
Decent approximation of batter values (without credit/debit for fielding prowess) and of pitcher values (without credit/debit for fielding exploits or batting incompetence). I will share it on Google Docs when it's cleaner, as it's an f'ing mess right now.

Issues making it an f'ing mess. Multiple sheets that require information from different pages (team page, Sim Stats L/R, Stats-Team, Stats-Team Fielding). Incomplete data to accurately compute OBP:
[img:ed64f25102]http://upload.wikimedia.org/wikipedia/en/math/5/6/e/56e63a7b1b0724813e09ba0ab1160a22.png[/img:ed64f25102]
and a couple other issues. The fielding data is really deplorable and makes the valuation of slick fielders very tricky indeed. I had an idea to fudge the distribution of WinShares to fielders based on DeanTSC's OffvsDef sheet, but I forgot the mechanism I came up with and I'm way more inventive with spreadsheets away from my computer.

At first blush (having looked at two teams) it does kind of unpack somethings. Like the team I linked above...
without even knowing the D impact of Utley at 2nd, I can say from his hitting that, to do this team again, I'd trade him for a similarly valued second sacker who maybe hits less.

I will continue my love affair with Bob O'Farrell.

I will be using Pete Rose 1969 more in the future. He created half a win share less than Dimaggio's card, at ~60% of the price. The difference is that he got more PAs, specifically 30 more AB, 15 more BB and more HBP or fewer SF. That was leading off most of the season I think. Which brings back something else; the importance of putting your best guys up top. Bonds was a monster as only Bonds Ruth and maybe Gibson can be. He had 701 AB+BB. Rose had 766. Dimag 721. Buck Leonard 763.

Chuck Klein at DH vs Buck Leonard at 1B... About one win with the bat. Leonard has a sterling glove at 1st. Think he makes up the ten marginal runs with his glove? I think so. He's a .502 winner at 100, .532 at 80, .498 at 140...

Scott Brosius. I'm writing him to bad luck. 550 PA. He took a lot of balls in the body, and had FIVE injuries. 36-38, 67-73, 106-113, 115-129, 141-141. He hit the ball well against LHP, which was really his job (.302/.382/.556 vs L... Chuck Klein vs L, Bobby Wallace vs R). But the injuries. If I'd have known, I'd've dropped him after game 106, not sure if that would have been 10 or 20%, but considering that he'd play a game and a half between 107 and 129, someone 80% of him would have produced more than he did during that stretch.


Any rate, I will put up a link to the sheet when I think it's better linked together, minimal work, and more logically arranged.

[sschu]

Yes, my mistake, in a hurry. It should be Actual / Expected, or the effectiveness of the hits&OBP of your team.

This number will vary some, based upon the obvious and not so obvious factors such as clutch, luck, power etc. Is is very much a factor of team speed and opposing OF/C arms. IMHO, middle IF range also makes a difference as more double plays means that your OBP results in fewer runs.

This is why the high BP1B parks need high average team speed to take advantage of the expected additional base runners. Stealing is also good, but their is no substitute for jackrabbits on the base paths.

The margin of victory in ATG is very thin. So when you see new posters say that range and arms are overrated, you know they have not looked at reality. :-)

R/HR is one to look at, what it tells you is are your sluggers hitting HR with guys on base. This makes the case for slugger teams also having high OBP guys at the top of the lineup. Andy liked McGraw in Polo 41 for this exact reason, today there are some good choices like Hamilton or Delahanty that can fulfill this role. Mantle in Polo 41 is gold, lots of walks.

FWIIW, sschu