Does defense matter in OOTP6?

[gmo]

Just now I set up a league, but before running the draft I did the csv export. I adjusted the ratings in a spreadsheet and was able to reimport just fine. Maybe it was because it was pre-initial draft. I'm sure I had problems doing this before with the rosters.

Probably already known bug: baserunning instincts did not take from the editted csv file values. As a sidenote to those who love the 2-way players, a few youngsters that were pitchers had some decent hitting ratings/talents, and a few field players had decent pitching ratings/talents as those did not change when importing from csv. Oh, and closer ratings all came in at 52-54 whereas everybody else got the exactly 50s I had set.

What exactly would make a good "average" team for testing? To do home field advantage testing I picked some players I thought might compose a pretty much average team and copied them to make every team the same. I wouldn't want to make every player on every team (pre- any adjusting for testing) a 50 rating in everything. I'd set something like the 1B to slow but good hitting, SS to quick but no power, #1 SP hard throwing & solid, #5 SP quite low ratings, etc.

For IFs the range+arm sounds very likely. Maybe it's not half & half, but it's got to be some combo.

[Skipaway]

Originally posted by gmo
For IFs the range+arm sounds very likely. Maybe it's not half & half, but it's got to be some combo.

Just a wild guess, but by common sense the weight on arm should be highest on 3B, then SS, then 2B.

[gmo]

Originally posted by Skipaway
Just a wild guess, but by common sense the weight on arm should be highest on 3B, then SS, then 2B.

Yeah, that does make sense, but I could live with no difference among the IF positions. I was actually thinking one ratio, like 60/40 range/arm for everyone.

Since IF arm is new to V6 on top of the defense overhaul, I'd assume any tests from V5 might well give results somewhat different from V6.

[Craig Scarborough]

As someone who had a part in this, I can assure you that the fielding model is/was dead-on when it comes to the differences between a good fielder and a bad fielder. Each position was weighed differently and used actual MLB data.

However, that does not mean there isn't the possibility of a bug in the code somewhere. I'll talk to Markus about this - in the meantime, good job all.

[Skipaway]

Originally posted by Craig Scarborough
Each position was weighed differently and used actual MLB data.

I'm sure there is no actual MLB data available for Manny Ramirez at shortstop.

[dougaiton]

Originally posted by Craig Scarborough
As someone who had a part in this, I can assure you that the fielding model is/was dead-on when it comes to the differences between a good fielder and a bad fielder. Each position was weighed differently and used actual MLB data.

However, that does not mean there isn't the possibility of a bug in the code somewhere. I'll talk to Markus about this - in the meantime, good job all.

Thanks for the info, Craig. I think the problem here is that we're trying to simulate conditions that never take place in OOTP, but are important for the game world - namely, how does a team of the best possible defensive players perform compared with the worst imaginable. Moreover, what is the jump between a bad defensive SS and playing anyone there?

[Skipaway]

Originally posted by dougaiton
I think the problem here is that we're trying to simulate conditions that never take place in OOTP, but are important for the game world.

By OOTP, I think you mean real life.

Perhaps it's hard to test all the boundary conditions and strange settings.

[Hank Greenberg]

Im not sure if you held things like GB% static, velocity (K's) and the arm ratings, if you didnt that would change the results atleast a little bit.

[rasnell]

This whole thread is very interesting to those of us who go for the best defense up the middle and greatly sacrifice offense. If it turns out this is a very flawed strategy -- at least under OOTP 6 -- that's a major discovery.

I'll continue following your findings with interest.

[dougaiton]

Originally posted by Hank Greenberg
Im not sure if you held things like GB% static, velocity (K's) and the arm ratings, if you didnt that would change the results atleast a little bit.

They were held static to an extent. I used the same team, same leagues, same arms. k/9 remained very similar through all however many sims.

I think that one thing I'm not smart enough to know is how the defense (which I previously would value on range alone) and offense would offset. The one thing I would say is that if I had a choice between Enrique Wilson (i.e. a very bad defensive SS) at short or Edgar Martinez/Frank Thomas, I'd take the latter.

Even players without a range record range factors similar to bad defensive short-stops. Errors is something else, but still....

I understand that there has to be some replacement value in place, but I can't help but feel that ito assume any player stuck at short will make 400+ assists in a year is far from accurate. MLB data skews defensive ability upwards, because they only play very competent/excellent defensive players at SS in the broad scheme of things.

These are first impressions. I'll run some more tests tonight.

[Hank Greenberg]

A very good test would be to have a 10 team league where every team is fairly equal (batting and pitching, maybe just make 1 copy and export/import 10 times) but have defence cascade down (100-90-80-70-60-50-40-30-20-10) then repeat the same experiement, but hold range static and adjust arm ratings.

[dougaiton]

I know how people here love real-life examples, so here we go. This one is for amusement only, and I know that's its not illustrative of anything, but at the same time is illustrative of everything!:

Using the default rosters, I set everything 'off' including fatigue and had Frank Thomas (or Franky 'the big hurt' Cordoba) play shortstop with no backup on the depth chart, with no range ar short and an 28 arm (which is kinda average for a first baseman, no?) . Here's his line:

Shortstop 162 162 203 478 102 721 40 .945 1348.1 4.81

So, 40 errors is pretty bad, and his FP is worst in the AL.......by 0.04 pts. He had the 4th most chances in the AL (how did he get to all those balls?) and turned the 2nd most double plays. His RF was better than Mike Young, Orlando Hudson, Rich Aurilia 30 points better than Derek Jeter, 40 points better than David Eckstein and 30 pts better than Pokey Reese. Using these rather poor traditional stats, you could argue that Frank Thomas was somewhere between the 8th and 11th best starting shortstop in the AL.

And they say he won't make the hall on his defence!

[Smellyface]

This is all very interesting stuff. If I'm reading all this correctly, the best strategy may be to sign the best hitters available, regardless of position. You could have 6 or 7 1B playing all the infield positions and a couple outfield spots and you may have a great team, if they are all good hitters and you've got good pitching. I know that OOTP may only take the major league range of fielders at each position into account, but the scenario I just described shouldn't be workable.

As for the Big Hurt, don't knock the man... I heard Ozzie Guillen has been teaching him the art of SS behind the scenes....

[Hank Greenberg]

Originally posted by dougaiton
I know how people here love real-life examples, so here we go. This one is for amusement only, and I know that's its not illustrative of anything, but at the same time is illustrative of everything!:

Using the default rosters, I set everything 'off' including fatigue and had Frank Thomas (or Franky 'the big hurt' Cordoba) play shortstop with no backup on the depth chart, with no range ar short and an 28 arm (which is kinda average for a first baseman, no?) . Here's his line:

Shortstop 162 162 203 478 102 721 40 .945 1348.1 4.81

So, 40 errors is pretty bad, and his FP is worst in the AL.......by 0.04 pts. He had the 4th most chances in the AL (how did he get to all those balls?) and turned the 2nd most double plays. His RF was better than Mike Young, Orlando Hudson, Rich Aurilia 30 points better than Derek Jeter, 40 points better than David Eckstein and 30 pts better than Pokey Reese. Using these rather poor traditional stats, you could argue that Frank Thomas was somewhere between the 8th and 11th best starting shortstop in the AL.

And they say he won't make the hall on his defence!

Now put a 100 range SS there and see where he ends up (100 arm too)

[dougaiton]

I'm willing to take Craig's word on the mid-ranges (he could forget far more than I would ever know about fielding stats!), but what interests me is the difference between bad and terrible and good and great.

Well, here's old Frank, range 20, arm 30, and he's actually worse as a bad player than he was as someone with no rating whatsoever (the fp% remains the same, obviously) :
Shortstop 162 162 202 446 91 687 39 .943 1336.1 4.63

Note that as I'm just resetting, the fp% remains the same, but I can't help but think that RF should be changing too!:

As requested:
range 100, arm 100
Shortstop 162 162 204 483 87 726 39 .946 1335.2 4.89

To be honest, I'm more interested in how it is that there seems to be a certain level of defensive performance which is 'adequate' and that anyone can turn in a league average performance. I'm thinking that this must be a bug, and not a design feature. However, that optimum doesn't seem very high.

[Hank Greenberg]

I think part of the problem is GB% not being static team to team (more chances)

[dougaiton]

Originally posted by Hank Greenberg
I think part of the problem is GB% not being static team to team (more chances)

It's the same team, so you would think it would be as close as makes no odds. More errors create more chances, sure, but these guys are committing the same number of errors.

Moreover, the joy of this test should be this: a 0/0 defensive player should be so bad he should stick out like a sore thumb, and the same should go for a 100/100 who should stick out as fantastic, regardless of sample size, number of sims etc. It's a bit like making a hitter a pitcher, or vice versa - they should get killed over 162 games. Neither of these is the case for the extremes of fielding.

[Hank Greenberg]

Originally posted by dougaiton
It's the same team, so you would think it would be as close as makes no odds. More errors create more chances, sure, but these guys are committing the same number of errors.

Moreover, the joy of this test should be this: a 0/0 defensive player should be so bad he should stick out like a sore thumb, and the same should go for a 100/100 who should stick out as fantastic, regardless of sample size, number of sims etc. It's a bit like making a hitter a pitcher, or vice versa - they should get killed over 162 games. Neither of these is the case for the extremes of fielding.

I meant strictly in the # of chances (if the CWS had a ton of GB pitches that would explain all those chances). I think those 40 errors do stick out as a bit of a sore thumb (of course it should be closer to 60)...

What are the 3B and 2B range factors, could this play a factor?

[dougaiton]

Yeah, variables are obviously going to be factored into the creation of the stats, but I think the real interest is not in Thomas's perfomance but rather the fact that Thomas is ranked 20/30, he doesn't improve substantially. Hell, when he's 100/100, he's not even obviously better!

It should be blindingly obvious that its a leftfielder stuck in short, if he doesn't have any range, but if I showed you these three stat lines, which would you guess had no rating and no infield arm of which to speak, and which was had the optimum ratings and arm strenght that any player could possibly develop?

Shortstop 162 162 203 478 102 721 40 .945 1348.1 4.81 (0/20)
Shortstop 162 162 202 446 91 687 39 .943 1336.1 4.63 (20/30)
Shortstop 162 162 204 483 87 726 39 .946 1335.2 4.89 (100/100)

This is an area that is easy to test, because it should be so striking that it defies such things as sample size etc i.e. someone with no range should be terrible. By all accounts, the Thomas without rating/arm was BETTER than the 100/100 Thomas, if we look at the DPs.

Thus, I think it has to be a bug with the player editor or with the way in which OOTP calculates defense for non-ranged players. Whether its indicative of a bigger flaw of the range system is something that someone who knows more about it than me should answer.

[Hank Greenberg]

Originally posted by dougaiton
Yeah, variables are obviously going to be factored into the creation of the stats, but I think the real interest is not in Thomas's perfomance but rather the fact that Thomas is ranked 20/30, he doesn't improve substantially. Hell, when he's 100/100, he's not even obviously better!

It should be blindingly obvious that its a leftfielder stuck in short, if he doesn't have any range, but if I showed you these three stat lines, which would you guess had no rating and no infield arm of which to speak, and which was had the optimum ratings and arm strenght that any player could possibly develop?

Shortstop 162 162 203 478 102 721 40 .945 1348.1 4.81 (0/20)
Shortstop 162 162 202 446 91 687 39 .943 1336.1 4.63 (20/30)
Shortstop 162 162 204 483 87 726 39 .946 1335.2 4.89 (100/100)

This is an area that is easy to test, because it should be so striking that it defies such things as sample size etc i.e. someone with no range should be terrible. By all accounts, the Thomas without rating/arm was BETTER than the 100/100 Thomas, if we look at the DPs.

Thus, I think it has to be a bug with the player editor or with the way in which OOTP calculates defense for non-ranged players. Whether its indicative of a bigger flaw of the range system is something that someone who knows more about it than me should answer.

That is rather bad, maybe it is strictly a SS bug? Try the experiement with CF and 2B and see how that works.....

not that this doesn't need to be addressed