How to check pitching numbers for fly balls/ground balls

To keyzick,

I also see the shallow depth perspective. This is a great area for attention. Two or three years ago I became completely convinced that the ticket to sure victory was a $6 mil catcher. I beat the data to near death to reach that conclusion! In the end I proved myself wrong, but those kind of insights, flawed as they might end up being, are the fun and reason to play the game!

I've also found some good value with 3's at either 2b and/or ss, enough offense to more than offset their defense. In past years, I'd always keep a 2 or better at ss/2b/cf...not with this set (or maybe I could have done this all along and didn't realize it?).

Anyway, I've really enjoyed this set, whether playing in hitters parks, pitchers parks, lefty slant, righty slant...all of it. Only 4 more months until the next set...

I agree. This has been one of my more favorite sets as well. I was always worried that more efficient pricing would dilute the ability to have fun with a set. But each year there seems to be more and more depth. I have to wonder how long I could go as newb in today's game, given all the nuance involved, and what looks to me (with the benefit of hindsight) like a really steep learning curve.

There's going to be more to include on this topic, but FWIW I asked SOM for the official line on how G/F are ruled in the recording made by the game's Sim Misc tab. I haven't heard back yet, so having a team only nine games into the season, I went ahead and did my own analysis. I can report with 100% certainty the following:

All bunts (sac and sqz) are ruled GB (incl only to say they are in fact recorded at all).
All singles are ruled as GB.
All ground outs, ground out double plays, force plays and infield errors are ruled GB.
All doubles, triples and home runs are ruled as FB.
All fly outs, line outs, foul outs and pop outs are ruled as FB.

I did not have a case of an outfield error made, but I would guess it would be a FB.

The big thing to take away here is that what SOM told me previously, that singles are randomized, does not appear to be correct. ALL singles are ruled as GB. So for any imagination dragons out there, there are no bloop FB that drop in for a single. Every single is a GB "with eyes".

What does this mean for the SOM manager? Whenever you see a player approaching the 1:1 ratio for G/F, if he's any good, he's getting more than an average share of extra base hits. If he's not any good, he just flies/pops/lines/fouls out more than he ground outs. However, while you can make an argument that "goodness" is in the eye of the beholder, generally if you're seeing guys who are in a regular lineup and approaching the 1:1 ratio in the Sim Misc tab, you can likely safely assume that that guy is "good", Adam Dunn notwithstanding.

Good singles hitters with low SLG like Ben Revere can show a poor G/F ratio (assuming here that more GB are bad), but still be a net asset. You just won't be able to ferret that guy out of the Sim Misc tab very easily.

Okay. Time for a serving of home cooking.

I had drafted a team for another experiment that wasn't quite what I was looking for, and after studying the G/F ratios decided this was a perfect situation to pivot to a G/F team.

http://onlinegames.strat-o-matic.com/team/1391179

Pitching and the G/F Ratio

My six starters have the following real-life 2013 G/F as listed in espn.go.com:

Burnett - 1.43
Cobb - 1.35
Ross - 1.34
Liriano - 1.11
Kelly - 1.08
Cole - 1.05

However, after polling about a dozen leagues for their G/F ratios in the SOM Sim Misc tabs I get:

Burnett - 1.72 (7 leagues)
Cobb - 1.96 (8 leagues)
Ross - 1.96 (5 leagues)
Liriano - 1.65 (12 leagues)
Kelly - 1.84 (1 league)
Cole - 1.90 (9 leagues)

I would guess that the overstated SOM lines come from all singles being ruled ground balls. So the advantage here, I think if there is one to be had, is that these cards are especially devoid of extra base hits. I don't have a ratings book, so if someone can provide the gb(A)'s, I would appreciate it. But cards full of walks and singles will definitely not move a lot of runners around the bases. I also have a 1 at SS and a 2 at 2B to help keep those that do reach base (esp first base) honest.

Full disclosure: While these pitching cards have an obscene G/F ratio in real Strat seasons, they are by no means magic cards. Among the season averages, there are always low side outliers. I expect I'll see some of that first hand. Also, the G/F ratio correlates somewhat poorly with wins. Again, these are not magic cards. I'm hoping my salary construction and other Secret Formula type fundamentals will lend a hand in the team's success.

Hitting and the G/F Ratio

FWIW, most managers care zero about the G/F ratio. I polled four or five seasons and discovered that the average number of FB hit in the average season is routinely around 1825. Then I polled the top four offenses (runs scored) for each of those seasons and discovered that those four teams averaged around 1875 FB. So hitting FB in Strat does correlate with scoring more runs. But like all things in Strat, too much of a good thing does not necessarily mean more wins, just like teams that strive for 400 homers or 800 stolen bases don't necessarily win more games than a team looking for a more even-handed strategy.

Having said all that, I think I do have a team of hitters that could be expected to easily record the most fly balls. Will that translate to runs scored and put me in the top four in offense? More importantly, will the total strategy of winning the net G/F game equate to more wins? We'll all find out in real time.

As a last aside, I found another team where I targeted SLG in a pitchers park, something of a counter-intuitive strategy. The team didn't ultimately win the ring, but they made a pretty good showing nonetheless:

http://onlinegames.strat-o-matic.com/team/1385609

.457 team SLG in AT&T isn't too shabby...

***Edit - here are the Double Play chances (LHB/RHB)***:

Burnett - 4/2
Cobb - 12/12
Ross - 12/10
Liriano - 6/6
Kelly - 12/12
Cole - 12/11

Howell - 12/12

"The results were that the average ground ball generated 0.04 runs and caused 0.80 outs while the average ball in air generated 0.23 runs and caused just 0.62 outs. On a runs-per-out basis, balls hit into the air created almost 7.5 times as much offense as balls kept on the ground did."

Any analysis along these lines for real life baseball are irrelevant for strat. I real life you look at all the balls off the bat and the ones hit in the air are the ones most likely to be line drives or find a gap for extra base hits while the ground balls are more likely to be singles.

But Strat is NOT the real world. You have the 1915 Pete Alexander on the mound and a 6-8 is roll a 4-10. That is NOT going to have any chance of generating 0.23 runs or any runs for that matter. It is going to be an out. It is going to be an out 100% of the time. There is absolutely 0% chance it is going to find a gap. There is 0% chance the wind is going to catch it and carry it out of the park.

You roll a 6-7 and that is going to be an out. It is going to be an out 100% of the time. There is going to be 0% chance it will find a hole and go through for a single.

You roll a 4-5 and it is going to be a single. It matters not if the imaginary box score description describes it as a line drive single to CF or a ground ball up the middle. The results with respect to it being a single and chances of runners advancing is exactly the same.

To continue using Pete as an example. You look at that card and there are 10 fly balls for LH hitters and 12 fly balls for RH hitters. But unless you get a dice roll off the batter card there are going to be 0 extra base hits off that card.

When Clayton Kershaw strikes somebody out in a real game, I don't look at the television and think "Well of course he did, the roll came off the pitcher's card!"

Just because Strat rolls are deterministic algorithms (no wind carried home runs, no bloop hits, no shortstops tripping over their shoelaces) does not mean those results are irrelevant statistically. On fly outs, the runner on third can score. Some or all runners can advance. On ground outs, runners advance, are forced out or double-played. I agree that it's irrelevant when "a bloop drops in because of the wind." In our Strat world, the card just says "single".

I think it only makes sense that "fly balls" (even imaginary ones) produce more offense. Doubles advance runners two or more bases. Triples advance runners three bases. Home runs clear the base. Fly outs move runners favorably. Ground outs can also advance the runners, but not to the extent that fly balls can. So while you are correct that an out is an out 100% of the time, outs generate offense by moving runners around the bases (and improving production). You can't say that because a 4-10 roll is always a fly out (and an out is an out) there's no need for further analysis. If that 4-10 means the runner on third scores, that is a different outcome than if the 4-10 is a ground out and the runner on third is out trying to score.

Just out of curiosity, given your view, how do you see my experiment playing out? What should I expect to expect?

It is true some outs can move runners over. But I would still ignore ratios because in the strat world they tell you nothing.

If you want to include outs in your calculations of runs a card might generate you can do so. But take it the next step and take advantage of the outs being fixed. A gb()C for example will not just tend to advance runners it always will. Opposite true for the gb()B. And the gb()A will always produce a double play if a runner is on first.

Same is true for OF outs. A fb()A will advance a runner. A fb()C will never advance a runner. So why mess with the ratios? Just as the number of hits on a card are fixed so are the types of outs.

These are fixed results with no relationship to fb/fb ratios. Since you know them in advance time is better spent analyzing such subtle differences on the cards. Bottom line though is any determination that up to say game 81 a strat pitcher has a certain fb/fb results ratio has no relationship to how that card will deliver in the last 81 games. The card from which the outcomes are derived does not change.

Regarding the experiment my expectation is you will put in a lot of work with little to show for it in results. Or possibly worse misinterpret any results (winning or losing) that you might see. But if you enjoy the work then it is worthwhile. Anything that helps you enjoy the hobby more has value.

No matter how we slice and dice it, it looks like I'm not going to win you over!

I wholeheartedly disagree that a Strat G/F ratio tells you nothing. It tells you EXACTLY this:

FB was either a double, triple, home run, fly out A, B, or C, line out, pop out or foul out.
GB was either a single, ground out A, B, or C or bunt.

If you had to field a team and all you could know in advance was the ratio of fly balls to ground balls those players hit, which group would you pick from? Picking the guys who have more results from the FB grouping is a no-brainer. I just don't follow how you see otherwise.

And for all the emphasis you put on pitchers cards yielding similar results, how do you explain that in 10+ league samples, guys like Cobb have close to a 2 to 1 ratio, while guys like Travis Wood consistently have a less than 1 to 1 ratio? It is precisely because the numbers are fixed that they are predictive. So these are not return to the mean results. If they were, and if what you were saying was true, all cards would yield the same ratios over time, but it is very easy to see they do not.

Like I said earlier, this isn't some magic bullet. You could pick your guys by OPS, SLG or any other metric. The G/F ratio is just a good proxy for SLG, which isn't a bad way to build a winning team if you were to focus on simply one theme or metric.

"The results were that the average ground ball generated 0.04 runs and caused 0.80 outs while the average ball in air generated 0.23 runs and caused just 0.62 outs. On a runs-per-out basis, balls hit into the air created almost 7.5 times as much offense as balls kept on the ground did."

Just because Strat rolls are deterministic algorithms (no wind carried home runs, no bloop hits, no shortstops tripping over their shoelaces) does not mean those results are irrelevant statistically.

I don't enjoy agreeing with Valen, who blatantly lied about me on a public forum, but I do...at least on SOM's semiotics. Strat rolls are deterministic algorithms in a probability game effectively functioning as an MLB simulation. It is not an actual simulation of the physical properties of baseball. So the physical dynamics involved in major league baseball are not relevant to SOM simulation. The only thing that is are the actual results SOM's probabilities and algorithms actually simulate. Player names and actual stadiums may add the veneer of Baseball's physical dynamics. But SOM's algorithms and probabilities do not represent those dynamics and are not interested in them.