Simulating different lineups

[Jimbo]

After last week's game Jamie and I were kicking around ideas and how to improve our run output.

 

I threw together a quick simulation program using our real-time stats. I didn't come up with a way to simulate opponent fielding errors, so the numbers are in earned runs scored.

 

As a benchmark, I first used out opening day lineup, then compred other lineups to that.

 

Jamie is thinking about mixing it up, so all the weaker hitters aren't up in a row. Looking at the figures, I would say it would have no effect.

 

We basically are going to average 8 or 9 runs a game no matter the lineup. If we think we will be in a close game, it's possible to tweak it to get that extra run or two.

 

Results

[nullbyte]

Quote:

---

All valpocrewsader lineup

Simulated: 18.12 earned runs per 7 innings.

---

Hahaha.. that's awesome.

[nullbyte]

Short of major developments in human cloning, here are my thoughts...

 

The difference between 8 and 8.75 and 9.66 is pretty significant.. Most managers would kill for 20% of it. That's a huge gain for free, so I think we should take advantage of it in those games we think we can win..

 

And for the others, in the interests of equal playing time and comaraderie, we can take the exact opposite order. If we can pick 3 games that are potentially winnable and 3 that are doubtful, that'd be a good balance the rest of the way. (Even 2 winnable/4 doubtful is fine, since the batting order to this point has been closer to favoring the players that would be at the top of that OBP list.)

 

Any thoughts?

[Jimbo]

One thought that came to me yesterday is to come up with one batting order and "daisy chain" it.

 

Say the 9th batter makes the final out this week, then next week the 10th batter "leads off", followed by 11th, 12th, ...

 

No need to constantly reshuffle the lineup each week. Everyone would be within one PA of each other the final six games.

[brewerjamie15]

Let's do that daisy thing. Good idea.

[nullbyte]

Ooh yeah, good thinking. So now all that remains to be decided is which one to start with http://forum.brewerfan.net/images/smilies/smile.gif

[Jimbo]

Let's do that daisy thing. Good idea.

 

Ooh yeah, good thinking.

 

Daisy Chain: Man Law!

[RightFieldCoder]

Man Law!

[rluzinski]

Hey, how did you do that simulation anyway? What did you use for average base advancements? (for instance, how often does a man on first score from a double). I'm about 80% done with a corny simulator but guessed on some base advancements.

 

Thanks.

[Jimbo]

Russ,

 

Keep in mind this is slow-pitch softball vs. baseball.

 

Basically we play station-to-station. I assumed everyone advances one on a single, two on a double. With two outs, the runners take the extra base.

 

In real time I've been quite liberal in awarding "infield hits".

 

Ground outs, I assume the throw goes to second, if applicable, else to first. Everyone else advances.

 

Fly outs and strike outs - no advancement.

 

I didn't simulate fielding errors nor advancing on throws.

 

I wrote up a program in Visual Basic, stored the probabilites in Access. While I was at it, I drew up a Brewerfan.net Game Day graphic similar to MLB, that way I could watch all the valpos, Robins, razzzors streak around the bases and light up the scoreboard!

[rluzinski]

Got it. I'm sure that's good enough anyway. If you ever want to work with me to write simulator for major league games, please let me know. The fact that I'm using Excel should tell you where my programing skills are. It's BASIC or bust for me http://forum.brewerfan.net/images/smilies/frown.gif

 

While I was at it, I drew up a Brewerfan.net Game Day graphic similar to MLB, that way I could watch all the valpos, Robins, razzzors streak around the bases and light up the scoreboard!

 

That's pretty damn awesome. http://forum.brewerfan.net/images/smilies/smile.gif

[RightFieldCoder]

I don't think I streak anywhere along the bases. It's more chugging for me.