2047 SBA/Inning (mid-season update)

[RonCo]

Here’s an interesting chart … maybe. As you can probably tell, I’ve been interested in stolen bases and how much ability one has to control their use and timing (my personal goal is to steal “often” but only in properly leveraged situations … i.e. in later innings and in close games). It’s very hard to parse data in this way, but I have been using my little game log script to get medium close. But bottom line, that’s how I try to adjust my strategy sliders, and to a degree the individual player sliders.

Bottom line, though, this chart shows each team’s SB performance by inning—how often they run and when they are successful. I can’t cut it by score at the present time, so I can’t get to where I really want to go. Still, I can look at how different teams are performing later in games—i.e, who is running late more often than early (counting extra innings as “late”). That’s what the last three columns are—total SBA, SBA in innings 7 or higher, and the percent of late SBA against the whole.

So, for example, Phoenix has run only 107 times, but leads the league at 30.5% of those attempts being in later innings. New Orleans is not too far behind, running even less often (88 times, but having 29.4% in later innings. My own YS9 runs more often, and has 25.4% happening later.

On the high-volume scale, Sacramento has attempted 28.7% of their attempts in the later innings—which is high—despite running 336 times overall.

Note that the league average says that 24.3% of SBA (regardless of score) happen in later innings.

There’s more here, too, obviously. The league is succeeding at a 68.1% clip (by my script). Without doing the math, I’d guess that our run environment suggests teams need to steal at something over 70.5% to make the practice valuable—so as a league I’d say we’re losing runs.

San Fernando has been successful at a 76.4% rate, best in the league. Per StatsPlus, they are also leading in running value with 6.88 wSB to Sacramento’s 6.41.

I love it when a plan comes together.

Anyway, there’s more to gaze at, but for now I’ll just leave this here for you to see what you think.

[JimBob2232]

So, for example, Phoenix has run only 107 times, but leads the league at 30.5% of those attempts being in later innings. New Orleans is not too far behind, running even less often (88 times, but having 29.4% in later innings. My own YS9 runs more often, and has 25.4% happening later.

Am i missing something? i dont see this data on teh chart.

[RonCo]

Ah. I scanned over the wrong line.

[usnspecialist]

I love when things work the way I intend them to. I subscribe to the Jeremy Clarkson model, SPEED and POWER (OG top gear reference for you less informed out there).

[GoldenOne]

Guessing you ran those numbers a while ago? Charlotte is at 161 SBs right now but you are showing 144. Our percentage is slightly better now but basically the same.

[RonCo]

Guessing you ran those numbers a while ago? Charlotte is at 161 SBs right now but you are showing 144. Our percentage is slightly better now but basically the same.

Yeah, I ran it off last week's data pull. I went back to see if I might be missing something in the script (which is always possible), and no, it doesn't look like it. Per my count of your games last week, the Cougars stole 17 bases.

[lordtoffee]

That data makes sense, as I haven't tried to set up to run. Then again, the Robins might be going with a makeover.

[cheekimonk]

Believe it or not, RonCo, in my Indy days I parsed an entire season for SB & CS. I came up with a breakpoint of 65% balancing creating vs. erasing runs

[RonCo]

Believe it or not, RonCo, in my Indy days I parsed an entire season for SB & CS. I came up with a breakpoint of 65% balancing creating vs. erasing runs

Well, that’s interesting, but I’m not sure what you mean.

So, let me blabber a bit...

There are two very different ways to ask yourself if stealing bases is worth it. (1) Are you gaining more runs than your opponent, and (2) are you actually scoring more runs by running than you would if you didn’t run. Both of these are complex--because the steal is very much a cost/benefit risk.

For example, here’s the Baserunning data for the BBA today: https://statsplus.net/brewster/wsb/

This answers the first question—SFB has created 6.88 Runs Above Average. Note, however, that this does not say what “average” is. In other words, if the league as a whole is creating negative runs (getting caught more often than they should to make the practice productive), then it’s still possible that SFB is costing themselves runs by running—it’s just that they are costing themselves a lot less than others.

I think this is what you were doing regarding your comment.

To answer the question of “what is the cutoff at which you have to steal in order to make more runs” requires a different mindset. To answer that question you have to think about each team and each action separately. If, for example, the Nine runs 10 times and gets caught every time, it’s pretty clear that I’ve likely destroyed runs—so, on the whole, team productivity with stolen bases is about adding positive events and subtracting negative events. And really, what you are asking is “what is the run value of a SB and the run cost of a CS?”

Your answer can vary, but I’ve come to like the use of linear weights as pertains to run environment. Tom Tango published these some time ago: http://www.tangotiger.net/customlwts.html

If you buy that idea, then to determine the cutoff point of any particular season, which you need to do because these values can change due to the world around you, you need to determine how many runs a game the BBA is scoring. Then one can use the SB/CS values they calculate from this chart and reverse engineer the answer.

The BBA is currently averaging 4.407 R/G, so this approach would say a SB is probably worth .189 runs and a CS probably costs a team .445 runs. Do the math with those values and the raw cutoff is 70.2%. Per above, the league is averaging 68.1% success, so on the whole (if this concept is right) the league as a whole would have seen more runs scored if no one had ever run at all.

[RonCo]

Using that Tango table, you can calculate the cutoff rates for each environment as such:

R/G	1	2	3	4	5	6	7	8	9	10
SB Value	0.113	0.149	0.17	0.184	0.195	0.202	0.208	0.212	0.216	0.218
CS Value	-0.166	-0.261	-0.342	-0.416	-0.486	-0.552	-0.616	-0.678	-0.738	-0.798
Cutoff	59.50%	63.66%	66.80%	69.33%	71.37%	73.21%	74.76%	76.18%	77.36%	78.54%

So, as you see, when run scoring is higher (teams are scoring with their bats more often), then a team has to succeed more often in order for stealing to be a net positive. Only in an environment where teams were scoring about 2.7 or so R/G would a 65% success rate be a real positive.

[scottsdale_joe]

. . . the league as a whole would have seen more runs scored if no one had ever run at all.

Calvin Coolidge, Dick Van Dyke, and Jerry Seinfeld would agree about not running.
And then there was Harry Reser's jazz band putting it to music in the Kenny and Dennis song.

[cheekimonk]

Sorry...I meant a 65% success rate by your team's runners. I realized that was nowhere near real world, but it's what my numbers spit out.

[RonCo]

I'd be interested in how you calculated that. Is it posted round here?

[CTBrewCrew]

Yellow Springs & Jerusalem discussing SB% successes'. (Nice leg kick @RonCo )

[allenciox]

This is very, very interesting, and prompted me to look at my individual slider settings to see why my steal percentage as a team is so low.
Based on players on my current roster, I have 84 SB against 42 CS --- so exactly 2/3, which is pretty bad.

When I looked at the individual slider settings, I saw that the three players that are really messing up the percentage are all speed=9, steal=8, which I have long known is a recipe requiring careful player control. These three players have combined for 19 SB and 16 CS. Removing them from the team average would leave 65 SB and 26 CS for everyone else, or a 71.4% clip, which isn't too bad. Still, there are a few players in the later category for whom I have reduced their individual slider settings.

Here's the rub. All the players with speed 9, steal 8 have had their individual steal slider all the way to the left at least since the beginning of the season. The only way to fix this I fear is by changing my team slider settings, which, of course are strongly influenced by score and inning ***sigh*** which I guess I will have to do.

Meanwhile, my owner chides me saying that his goal for me this season was to improve my number of stolen bases in 2047 ---- it was 12th in Flick last season and is still 12th in Flick. So he "REALLY" won't be happy with me now.

[allenciox]

Another thing to note is understanding what steal percentage your team "should" be shooting for. I also did an analysis awhile ago (during OOTP 22 Beta) using the simulation module about what steal rate produces more runs than it costs. My result, in a neutral MLB environment, was 70.5% --- if the team was above that rate, then increasing steals while still remaining above that rate would produce more runs. If it were below that rate, reducing steals while improving steal percentage would generate more runs. It is interesting that my analysis and his both produced that 70.5% figure.

But also keep in mind that the goal isn't to produce the highest run differential between you and your opponents --- it is maximizing winning percentage --- and there are situations where these goals conflict. So, if it is in the bottom of the ninth and you are one run behind, an attempted steal might be -EV from a runs perspective but +EV from a wins perspective.