Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Beat articles, power rankings, statistical analysis, etc. goes here.
allenciox
Ex-GM
Posts: 350
Joined: Sun Jun 16, 2019 2:23 pm
Has thanked: 98 times
Been thanked: 144 times

Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by allenciox » Fri Mar 19, 2021 10:47 am

I have been doing a lot of analysis using simulation mode in OOTP 22 beta. The way I have been doing it is by creating a random "2021-MLB-quality" league, Then I ran a bunch of 100,000 game simulations between each pair of teams. After doing that, I hollow out the "middle" team replacing each player on that team by "average" defensive players at every position.

I then identified two teams: the team that has the best offense and the worst offense.

With my control team I then ran 100,000 game sims against the worst and best offensive teams, varying the defensive parameters for each of the positions and seeing how many runs were allowed in each case. On my machine, each 100,000 game sim takes eleven to twelve minutes to run, so this is equivalent to over 617 162-game seasons.

Later on, I might report on results in detail, which are fascinating... but one of the things I could do (and did) was to identify the per game runs allowed comparing an average fielder to both an awful fielder and a great fielder at every position. holding every other fielder on the team to being average.

Once I had that I could look at what happens if I vary two positions at once. The first I chose was SS and 2B.

Now, why did I do this? Well, in MLB, you would expect that for hard balls hit midway between the second baseman and shortstop, that the one with greater range might get to it. If neither have the range to get to it, it gets through the infield. But as long as one of them have sufficient range, it will be an out (barring an error, etc.) Thus you would expect that if you have two bad fielders out there, then that is worse than the run deficit of adding together the run deficits of having bad SS/average 2B or average SS/bad 2B. If you have two great fielders out there then either fielder might be able to get to that hypothetical in-between ball, so the effect of having two great fielders would be less than adding together the effect of having just one great SS or one great 2B.

But in the situation where you have a great 2B/bad SS, or a great SS/bad 2B, you might expect the result to be better than having two average fielders, because some balls neither average fielder would be able to get to, while either the great 2B or great SS WOULD be able to get to it.


So, does OOTP work the same way? Well, for each possible hit, we know that OOTP conducts an "opposition" roll between the hitter and the potential fielder. But is that potential fielder that the opposition roll will be conducted against decided when the ball is hit? Or does it check to see whether multiple fielders might get to the ball and then choose the one from there that will be used for the opposition roll?

I am going to call the first possibility the "null" hypothesis, and the second one, the hypothesis we want to test.


So here is how I set up the simulation:

2B: Average 2B is defined as 120 on each rating (IF range, IF error, IF arm, IF TDBP), while Bad 2B is 40 on each rating and Great 2B is 240 on each rating (all these are using 250 "internal" scale, you can roughly compare it to our 10-point scale by dividing each number by 20).
SS: Average is defined as 140 (or 7 on our scale) for each rating, while bad SS is 60 on each and great SS is 250 (max) on each.

For all sim combination, I did one 100,000 sim against the best offense, and one against the worst offense, and averaged the results together. For each individual sim, the 95% confidence interval around the obtained values is plus or minus .025 runs/game. For the average of the two, it is about plus or minus .02 runs/game.

Here are the combinations and their results:

1. Average 2B, Average SS: allowed an average of 5.592 runs/game. This is the baseline.
2. Bad 2B, Average SS: allowed an average of 6.108 runs/game (deficit of .516 runs/game compared to baseline)
3. Average 2B, Bad SS: allowed an average of 5.974 runs/game (deficit of .382 runs/game: note that a bad SS is better than a bad 2B)
4. Great 2B, Average SS: allowed an average of 5.213 runs/game (bonus of .379 runs/game)
5. Average 2B, Great SS: allowed an average of 5.120 runs/game (bonus of .528 runs/game)

So, if OOTP decides who a fielder will be before a ball is hit, then we would expect the deficit of two bad fielders to be .516 + .382=.898 runs/game. But for two bad fielders we actually get 6.541 runs/game (deficit of .949 runs/game).

For combination of great 2B, and great SS, in that "null" condition, we would expect a bonus of .379+.528 = .907 runs/game but the actual bonus comes out to only .783.

For great 2B/bad SS we would expect .379-.382=.003 deficit, but we get .001 deficit, slightly better (although this is well within the 95% margin of error mentioned above).

For great SS/bad 2B we would expect .528-.516=.012 bonus: we get .020 bonus (again, those two values are within margin of error).

So what does this give us: Well, we see that combining two bad fielders clearly compounds your problems, while two great fielders does not yield as good a result as adding together the individual bonuses for each of them. Both of these are consistent with the hypothesis that OOTP, like real baseball allows multiple players to potentially get to a ball. However, pairing a great SS with a bad 2B or vice versa does not seem to get clearly better results than two average fielders: but the results are in the right direction.

So, then I looked at the OF: notably CF and LF. Here are the results:

1. bad LF + bad CF: gets a deficit of .521 runs/game where a deficit of .51 runs is expected assuming independence: check
2. great LF + great CF: gets a bonus of .532 runs/game where .509 runs are expected: check again
3. great CF + bad LF: gets a bonus of .11 runs/game where .077 are expected: clear third check
4. great LF + bad CF: gets a deficit of .032 runs/game where a deficit of .057 is expected: game, set, and match.

All of these are beyond the margin of error, except for the first, and it is close, and in the expected direction.

In summary, it appears that OOTP, like real baseball, allows for a good fielder to "help out" a worse, neighboring fielder, but the effect is very small.

GoldenOne
Ex-GM
Posts: 3317
Joined: Wed Mar 14, 2018 1:36 pm
Location: South Riding, VA
Has thanked: 728 times
Been thanked: 485 times

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by GoldenOne » Fri Mar 19, 2021 11:14 am

Good stuff!!
Brett "The Brain" Golden
GM: Nashville Goats 2034-2039 (The Plan® was working when I left!)
GM: Charlotte Cougars 2040-2052
GM: Rocky Mountain Oysters 2053-2057
2056 BBA Champions!

"Tonight, we take over the world!"
-- The Brain

sjshaw
Ex-GM
Posts: 587
Joined: Sat Mar 30, 2019 4:09 pm
Has thanked: 292 times
Been thanked: 105 times

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by sjshaw » Fri Mar 19, 2021 12:55 pm

Thank you!
GM, Louisville Sluggers, end of 2038 - current

Image

crobillard
Ex-GM
Posts: 2936
Joined: Wed Mar 05, 2014 7:38 am
Has thanked: 297 times
Been thanked: 240 times

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by crobillard » Fri Mar 19, 2021 1:07 pm

Cool feature. Thanks for the info!

User avatar
Dington
GB: Recruiting & Development Director
Posts: 4811
Joined: Mon Mar 30, 2020 12:06 am
Has thanked: 2073 times
Been thanked: 1147 times
Contact:

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by Dington » Fri Mar 19, 2021 1:51 pm

Good info. Pretty much what one would assume, but good to have stats to back it up.
Image
Nashville Bluebirds GM
HOW I BUILD A WINNING TEAM <---Click
Kuwait City GM 2042-43
2043 UMEBA United Cup Champion*

User avatar
Jwalk100
GB: FL Pacific Division Director
Posts: 3039
Joined: Sat Jan 11, 2020 9:42 pm
Has thanked: 1817 times
Been thanked: 771 times

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by Jwalk100 » Fri Mar 19, 2021 4:05 pm

I saw this in SOBR Ratings Theory. Was that you @allenciox?
Image
ImageImageImageImageImage

User avatar
RonCo
GB: JL Frontier Division Director
Posts: 19807
Joined: Sat Nov 14, 2015 10:48 pm
Has thanked: 1981 times
Been thanked: 2901 times

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by RonCo » Fri Mar 19, 2021 6:04 pm

Interesting way of looking at it.

What you’re asking, in OOTP speak is “does a shortstop with great range make a play in zones 4MS, 4M, or 4 MD?” (and/or “does a second baseman with great range make a play in zones 6MS, 6M, or 6MD?”). In other words, is a shortstop can make a play in zone 4M, then he can possibly “save” a play the second baseman should have made.

* That “Should” is a big word there. The question ultimately is whether OOTP would demerit a poor range second baseman if the shortstop didn’t make the play in question. I think the answer is “yes.” But I admit I don’t know. Zone Rating in OOTP is weird, and I'd want to understand how they use that "impossible" range before I'd be comfortable saying anything with certainty.

Anyway, the answer to that question about shortstops in 2B zones and vice versa is “yes, but …” which kind of buttresses your argument, but maybe not so much.

Here’s what I mean.

I went to 2045 and grabbed the results file I put out, then pulled out all ground balls and sorted them by the zones they were fielded in and the fielder who fielded them. This includes fielders who started double plays. This gives me every play in every zone, and who made them. I then highlighted plays made in shared zones.

First, to refresh, here are the zones OOTP uses, with the exception that center field gets split into “R” and “L” segments.

zones.PNG

Here’s the chart of ground ball locations and fielders in the BBA for 2045:

2045-GB-by-Zone-Fielder.PNG

So, for example, shortstops made 128 plays on slow rollers hit up the middle to the second base side (zone 4MS), and 276 plays on balls hit harder, and 135 on balls deep behind second base. Similarly, second basemen made a total of 151 plays in zones on the shortstop side of the base. That sounds like it corroborates your approach. The problem, of course, is that there’s no real way to establish what infield shifts are doing.

I mean, shortstops made ten plays in zone 4 last year (and three in zone 4D). Those almost have to be due to some kind of shift, right? Seriously, I cannot see a game in a normal situation where a shortstop makes a play on balls hit to those zones.

I should also note that if you’re measuring defensive runs saved (and weighting Bad, Average, and Great as you are), that the number of plays in each zone are important—meaning shortstops have more natural plays than second basemen, so there’s a warpage there.

It’s interesting to see the 56 zones, and how 3B perform in the short and normal depth ranges—which makes sense. Note, too, however, that third basemen made 119 plays in zones that were dead shortstop zones. This, to me, is almost certainly due to shifts, not range. At least I’d hope so.

Same kind of thing in the outfield

2045-FB-by-Zone-Fielder.PNG

Certainly those two shared gaps might suggest overlapping ranges, but I’m not sure about that. Could that be outfield shifts? Some of it almost certainly is, but I have no idea how much. The intriguing data in that aspect here are plays that the RF and LF made in the 8R and 8L zones. Those were rangy plays, but how much did shifts aide them.

I note that the generic percentage of plays in those shared 78 and 89 zones are quite similar, which makes me think that plays in the outfield are not really as overlapped as one might … hope?

Don’t know.

Bottom line: I’m intrigued by what you’re doing, and I like the basic framework of the idea. It makes logical sense, anyway. So I’m buying the base concept. But given how I think OOTP defense works, I’m not buying the magnitude of the findings, and I'm pondering that "makes up for deficiencies" bit simple based on how how I understand the algorithm works. I note, I could always be wrong. :)

Ultimately, I think the question is: “how does OOTP determine the fielder of any particular ball in play, given its zone?” I say that because I'm fairly confident the game decides where the ball is hit first, and then has to go through some algorithm to determine who attempts to field it. This idea would mean that the fielders could--and probably do--essentially work on their own, but that, yes, a great range provides value in zones adjacent to their normal ranges. I’m not sure I’d say he “makes up for other defender’s deficiencies,” though maybe that’s just dicing semantics. I suppose that great defender could make a defender appear better than his ZR says he is, though.

Alternatively, if you have great range, I suppose you could say that playing next to another guy with great range makes you look worse than you are. :) I mean, how many plays did Brooks Robinson steal from Mark Belanger?
GM: Bikini Krill
Nothing Matters But the Pacific Pennant
Roster

User avatar
RonCo
GB: JL Frontier Division Director
Posts: 19807
Joined: Sat Nov 14, 2015 10:48 pm
Has thanked: 1981 times
Been thanked: 2901 times

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by RonCo » Fri Mar 19, 2021 6:20 pm

Though I guess it makes sense that your magnitudes are so large given the ratings for average and poor players are probably really better labeled bad and extremely bad.
GM: Bikini Krill
Nothing Matters But the Pacific Pennant
Roster

User avatar
CTBrewCrew
GB: FL Heartland Division Director
Posts: 5110
Joined: Wed Sep 04, 2019 11:20 am
Location: Milford, CT
Has thanked: 923 times
Been thanked: 1316 times

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by CTBrewCrew » Fri Mar 19, 2021 7:57 pm

Interesting stuff @RonCo & @allenciox - Might try to Visualize something out the data you produced. :)
Image

User avatar
Dington
GB: Recruiting & Development Director
Posts: 4811
Joined: Mon Mar 30, 2020 12:06 am
Has thanked: 2073 times
Been thanked: 1147 times
Contact:

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by Dington » Fri Mar 19, 2021 9:50 pm

I'm just gonna teach my guys to hit the ball over the fence so your fancy fielders can't touch them.
Image
Nashville Bluebirds GM
HOW I BUILD A WINNING TEAM <---Click
Kuwait City GM 2042-43
2043 UMEBA United Cup Champion*

User avatar
RonCo
GB: JL Frontier Division Director
Posts: 19807
Joined: Sat Nov 14, 2015 10:48 pm
Has thanked: 1981 times
Been thanked: 2901 times

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by RonCo » Fri Mar 19, 2021 9:55 pm

Probably wise.
GM: Bikini Krill
Nothing Matters But the Pacific Pennant
Roster

allenciox
Ex-GM
Posts: 350
Joined: Sun Jun 16, 2019 2:23 pm
Has thanked: 98 times
Been thanked: 144 times

Re: Yes, your awesome CF. SS CAN make up for adjoining defensive deficiencies!

Post by allenciox » Sat Mar 20, 2021 9:26 am

You know, for ease of "analysis" purposes, we should create a new "game": "Baesball: Three true outcomes". The rules are as follows:
1. There are no fielders, just a catcher, a pitcher, and a hitter.
2. Walks, strike outs, and home runs are as they are now.
3. Any time a player hits a ball into the field of play, it is automatically an out.
4. No base stealing allowed.

OOTP could generate an extremely accurate model. Players don't even need to run the bases, so injuries would be practically non-existent.

Post Reply Previous topicNext topic

Return to “League Features”

Who is online

Users browsing this forum: No registered users and 8 guests