Are Coaching ratings now gone?

[gads]

Originally Posted by I♥Pepper

Awesome! I see feed back. I'm glad this topic is gaining traction. I want to reply to all of you. I just need a little time. I've got a lot of projects going right now. Finally getting some good weather.


A longer sample is not needed. The fact that a stolen base occurred at all within the given parameters, namely a 2 run deficit, proves the point. I have no way to control what happens with a 2 run deficit. Not in OOTP16. I did in OOTP15. I am at the a.i.'s mercy now in that situation.



First point is good. It was a pain in the rear. Streamlining is a good thing. It was made easier by the fact that I could export and import. No need to do it every year. The new presets are a good thing. I want to see settings for all the major situations though. There are 3 bases and 1 batter. One swing of the bat can lead to a max of 4 runs. Therefore there are a minimum of 11 situations concerning run deficits and run leads.>4,4,3,2,1,0,-1,-2,-3,-4,and >-4.The amount of innings covered in each situation could be compressed however.

Second, the settings available do not cover 2 run leads or 2 run deficits.Only +/-1 and +/-3. What is supposed to happen with +/-2. Confusing to me. Also +1 and -1 are not the same. It just gets worse with +2,-2 and +3,-3.

I think +/-3 means -3 to 3. I don't think it means 3 and -3 only.

[Fyrestorm3]

Originally Posted by gads

I think +/-3 means -3 to 3. I don't think it means 3 and -3 only.

Which was his point - and you know, I actually kind of agree with that. My strategies tend to be different when I'm trailing by 2 runs than when I'm up by 2 runs.

So I still say the game is working as designed, based off of your stolen base example, but now I understand what you're looking for. I wouldn't be opposed to it being expanded again, now that I think about it.

[DB930]

I'm on board with the change... as long as the AI can't cheat and look behind the curtain. It's basing it's hirings on rep and strategies, too, right, and not ratings?

[bazeall]

I'm not on board with the change. So many elements of OOTP are options. This could be an option - expose or hide the skills ratings for coaches.

Instead, now if I want a highly rated hitting coach, I have to edit each one to see the ratings.

[RchW]

Originally Posted by bazeall

I'm not on board with the change. So many elements of OOTP are options. This could be an option - expose or hide the skills ratings for coaches.

Instead, now if I want a highly rated hitting coach, I have to edit each one to see the ratings.

What makes a hitting coach highly rated and how does that show in player performance?

If you can recognize a good hitting coach and their work is consistently good, why do they ever get fired?

[I♥Pepper]

Well, gremlins attacked my internet connection. I've been down all afternoon. I want to try to reinforce some of my arguments.

To pick up- where I left off :

Originally Posted by RchW

Can you tell me who the top 5 3B coaches are in MLB? What about the bottom 5 3B coaches in MLB. How would you determine their ratings?

I ask because it is impossible IMO to actually rate in game coaching skills objectively. There is no baseline or level playing field. I'd submit that it's likely the coaches on a bad team are superior to those on a good team but there is no way to quantify it. Player performance overwhelms coaching influence in the majority of games.

My main concern is over the performance of my team while simulating games. I don't want to have to play every single game to ensure my team handles basic situations correctly. I had confidence in the old system. The new system is to vague. I'm not really concerned so much with trying to replicate a real world coach's brain and transform that into digital code. I would like some reassurances, however, that the a.i. does have some solid logic behind it. One way to do that is give me the ability to set up the a.i. that controls my team during simulation to at least the same standard and accuracy as in the past OOTP releases.

Originally Posted by jasonn29tn14

Are you playing as GM Only or are you playing as GM + Manager?

Here s the setup I had when running initial tests :


If you think I should try something different, I'd be glad to run some more tests with any settings you suggest. My OCD is calmed greatly so long as I'm busy trying to fix the problem.

And in response to the more recent replies from DB930, bazeall, and RchW : I would like for all of us to be able to have our cake and eat it too. Maybe options that allow us to personalize the game a little more is the way to go. I personally only want the option to restore the old team strategy U.I. But i am in no way against options that let you personalize the coach evaluation process either. Convincing Markus may be a tall order. And to Fyrestorm3, Awesome! and thank you.

[I♥Pepper]

Oh, and one more reply that I missed :

Originally Posted by injury log

Yes, I didn't understand the example either. You instruct the AI to only steal early in games, and only in close games. You sim five games, and the AI only steals one base, and that happens early in a close game. That seems to be evidence that everything is working properly.

My frustration sometimes influences my vocabulary. To say that the simulation is broken is to harsh, I agree. The example I used probably needs to be laid out better in a blog post or a new thread, I've basically hijacked this one already.

The only time I expect the a.i. to steal in my example is with either a 3 run lead or a 3 run deficit. There are also settings for +/-1 run and a tied game. The 2 run situation is mysterious. The 4 and >4 run situation seems to be included in the +/-3 run setting, maybe, I'm guessing. That's the point. I don't want to have to guess.
Also, if you have have any ideas on how to improve the testing, I'm all ears and I'd like to help and volunteer my time. I don't go back to work for a couple more weeks still yet. I love baseball and computers. This game brings it all together.

[byzeil]

Originally Posted by I♥Pepper

I continued my previously started game. Starting new games is now a long process and I didn't want to go through that again for this test. I went to the team strategy settings tab and hit the reset button. I then confirmed that all situations were in fact reset. I then changed all the "stealing bases" settings for all situation to "never" except for one. The Early(1st - 6th) and Close(+/- 3 runs) situation I set to frequent. I then began simulating 1 game at a time until there was a stolen base. The stolen base occurred on about the 5th game.

Bottom of 1st, kc2 to stl0. So with the KC 2 run lead, the ai decided to steal. Where is the setting for 2 run strategy? There isn't one. And I am guessing that it wouldn't have mattered to the ai if I was up 2 runs or down 2 runs.

Players don't always do as the manager wants. 1 stolen base in 5 games does not prove anything is broken. The player went on his own or he misread a sign. This happens in RL and should happen in the game. The game is simulating humans playing baseball not robots doing only what we want them to do.

[I♥Pepper]

Originally Posted by byzeil

Players don't always do as the manager wants. 1 stolen base in 5 games does not prove anything is broken. The player went on his own or he misread a sign. This happens in RL and should happen in the game. The game is simulating humans playing baseball not robots doing only what we want them to do.

Ok, I can roll with that. But, I also have the right, as the manager and general manager and overall king of my fictional universe, to fire anyone who doesn't follow orders. The thing is, I can't tell if he is or isn't following orders because the game strategy settings aren't specific enough. They are just too vague.

[cardsfan05]

I also would like the option of showing coach ratings.

[The Wolf]

Originally Posted by Syd Thrift

This way, you see the guy and you can look up his history, but at the end of the day if he's a bum who rode the coattails of a good team, then, well, the Don Zimmers of the world need to eat too.

Quote of the week. WTG, Syd.

[The Wolf]

Originally Posted by RchW

If you can recognize a good hitting coach and their work is consistently good, why do they ever get fired?

You should ask Charlie Lau, allegedly the best hitting coach ever, that question. Cooperstown Confidential: The tale of Charley Lau ? The Hardball Times

[RchW]

Originally Posted by The Wolf

You should ask Charlie Lau, allegedly the best hitting coach ever, that question. Cooperstown Confidential: The tale of Charley Lau ? The Hardball Times

The question was rhetorical.

[frangipard]

Originally Posted by Matt Arnold

It's not an attempt to dumb it down - I would say the intent is to turn the game away from being a simple spreadsheet and numbers game, and inject some personality.

I had always found coach hiring to be a simple case - when I need a manager or pitching coach, simply go find the guy with the highest ratings. Doesn't matter anything about him other than the numbers. We may as well have renamed them to "Joe Legend" or "Joe Poor".

Changing the system to one where more of the data is hidden to you, and now your coaches have their own personalities and styles now lets you decide a little bit more what makes sense to you. Some are still better than others, but now you can actually make a case that it might make sense to grab a worse guy, but who does best with power pitchers, if that's how you like to set up your pitching staff.

The overall effect is still fairly small overall - hiring a "ground ball" pitching coach won't convert all your pitchers to groundballers overnight, but it now adds a new flavour to the game that didn't exist before.

Matt, I think this has been a clear success at the major-league level; as a game element, I love the choice forced on me by having a LEGENDARY manager who has a strategic approach I don't favor.

But I think showing all the ratings would add to that "hard choice" factor, not take away from it, especially when you think about filling out your minor league staffs. Do I really care about whether my AA Manager practices smallball or not? The team does not exist to win, it exists to develop players. (Unless a smallball manager improves his players bunting skills or something like that, which it doesn't AFAICT) On the other hand, I could have a meaningful choice between one guy ranked "decent" but "favors veterans" and another who is only "fair" but who "favors rookies." If the manger's "teach fielding" rating affects minor leaguers, that should be visible, too
.
In general, I don't think you're ever going to go wrong by providing meaningful info to the players. Yes, information overload is a real thing, but the way to avoid that is by avoiding meaningless information. IOW, when I go to "staff search," I can see his DOB, his age, and his years of experience. The first two are duplicates, the third is strongly correlated, and AFAICT none have any impact on the game. "Teach fielding" and "favors rookies" do, but are hidden.

In addition the general descriptions smallball/tactician/sabermetric/conventional hide as much as they reveal. Near as I can tell, "tactician" means he likes to substitute a lot; but what do you call a guy who likes to substitute a lot AND bunt often? And isn't favoring the L/R split a "sabermetric" principle?

[byzeil]

Originally Posted by I♥Pepper

Ok, I can roll with that. But, I also have the right, as the manager and general manager and overall king of my fictional universe, to fire anyone who doesn't follow orders. The thing is, I can't tell if he is or isn't following orders because the game strategy settings aren't specific enough. They are just too vague.

If over a significant period of time (many more than 5 games) with your game strategy set not to steal in certain situations players steal then either the coach isn't listening to you or the payers aren't listening to the coach (or both). If one players keeps on stealing I'd take it as the player going on his own and if multiple players steal then the coach isn't listening. If the former then trade the player if the latter fire the coach.

[byzeil]

Originally Posted by cardsfan05

I also would like the option of showing coach ratings.

I'm fine with that being an option but personally I like the new way of doing it.

In previous versions I didn't play with coaches on because it was tedious and it was just a matter of sorting by the different ratings and hiring the guy with the best ratings. Who to hire wasn't a tough decision because I wanted the guy with the best ratings so I sorted by the ratings and hired the guy at the top of the list. Now I hire a guy with a good rep and a personality that sounds like a match and then I have to evaluate the job he does. If I think he does a good job over time then I keep him but if I think he is hurting me I have to make the choice to fire him.

With ratings the only time I'm going to fire a guy is if a guy with better ratings becomes available. No real logic or thought needed regarding how well he seems to be doing just concern for the cost of the contracts.

In regards to evaluating a coach/manager it would be helpful if there was some more info regarding him. Maybe some sort of feedback from the players, like the morale "happy face" but in regards to the coach. Also some feedback from the managers in regards to how the relate/think about the coach and the same from the coaches regarding the manager. Those are things I'm sure a RL GM would take into account when evaluating a coach/manager.

[mpejkrm]

The devs have stated this in the past, but I'll say it this time - not every single thing in the game is going to be an option. Yes, OOTP is great due to its customization. But it would be ridiculous if every single aspect of the game had an option to be turned on and off.

[injury log]

Originally Posted by frangipard

But I think showing all the ratings would add to that "hard choice" factor, not take away from it, especially when you think about filling out your minor league staffs. Do I really care about whether my AA Manager practices smallball or not?

You might care if your objective is to develop coaches in the minors, so they can eventually be coaches for your big league team. And you might also care if you want to use a coach's minor league track record to evaluate him. Some coaches in the game are good, but have unproven reputations, because reputation is based on results, not on ratings.

Originally Posted by frangipard

In general, I don't think you're ever going to go wrong by providing meaningful info to the players. Yes, information overload is a real thing, but the way to avoid that is by avoiding meaningless information.

I agree with your second point here, but not your first. In general, in games, I think if a decision is obvious with complete information, the game is more interesting if it provides incomplete information. Poker wouldn't be a game at all if you could see your opponents' cards. When a decision is not obvious with complete information, then providing complete information can be interesting - that's the situation in chess. But when manager ratings were displayed to users, I think most of us agreed that we were rarely confronted with interesting choices.

Originally Posted by frangipard

In addition the general descriptions smallball/tactician/sabermetric/conventional hide as much as they reveal. Near as I can tell, "tactician" means he likes to substitute a lot; but what do you call a guy who likes to substitute a lot AND bunt often? And isn't favoring the L/R split a "sabermetric" principle?

There's also an 'unorthodox' category of manager. It's very possible the categorizations could be improved, and that's a worthwhile discussion to have.

[Avindian]

I would like the coaching changes more if it were easier to see whether coaches favored veterans or rookies in the screen where you're signing them. Clicking on each coach can be a bit of a chore, especially when you're like me and sign personnel for all of your minor league teams.

[frangipard]

Originally Posted by injury log

You might care if your objective is to develop coaches in the minors, so they can eventually be coaches for your big league team. And you might also care if you want to use a coach's minor league track record to evaluate him. Some coaches in the game are good, but have unproven reputations, because reputation is based on results, not on ratings.

Wait ... so the reputation is only based on results? I thought it was a loose approximation (i.e. like a scouting report) of his ratings. If I put Joe Maddon in charge of a team of stiffs, that will destroy his reputation? Buck Showalter is legendary in the game ... how did that happen with his pedestrian career W-L?

In the same way, if I have a light-hitting AAA affiliate and want to get them to develop power by assigning a hitting instructor who emphasizes that, are they just going to drag his ratings down because of what I gave him to work with? That seems kind of counterproductive.

I agree with your second point here, but not your first. In general, in games, I think if a decision is obvious with complete information, the game is more interesting if it provides incomplete information. Poker wouldn't be a game at all if you could see your opponents' cards. When a decision is not obvious with complete information, then providing complete information can be interesting - that's the situation in chess. But when manager ratings were displayed to users, I think most of us agreed that we were rarely confronted with interesting choices.

Meaningful information =/= complete information.

You're absolutely right that the latter kills the game. But the former is what gives it life.

I agree completely that there needs to be "fuzz" between the ratings and the data we are given. But there shouldn't be such a mystery about what ratings exist. If I decide that I'm scheme-indifferent, and just want to get coaches that works well with young players, I should be able to look for those guys; but other than by clicking on dozens of profiles, there's no way to find those guys. And AFICT the text descriptions don't make a distinction between "he's crappy with young players, but at least okay with vets" and "he's okay with young players but awesome with vets" ... they seem to be purely a relative thing.

Or say I want to build a smallball team, and decide that defense and baserunning are part of that. For both, there are just occasional mentions that I have to click to see. Defense I know matters, because in the editor there's a "teach defense" rating ... but there is no "teach baserunning" rating ... is that just mythical?

Do the tendencies affect player's ratings? If I staff my whole system with smallballers, will bunt ratings improve, or do I need to draft guys with those skills to begin with?

Again, I'm NOT saying we need completely accurate information, where we know exactly how the engine is working ("A spreadsheet game"). But I do need to know what things at least matter.

To make a very concrete example: I am starting a fictional league, and I'm going to put the parent club of the org I will run in an extreme pitcher's park. So the approach I might take is to draft guys with tools -- great speed, good throwing arms -- and stock my system with the best hitting instructors money can buy, and smallball managers who teach good defense and baserunning.

That's a coherent organizational philosophy, right? The kind of thinking and strategy the game should (in time, with luck) reward. But here's the thing: I'll have to click on a hundred or more profiles just to find six coaches who are good at both "teach defense" and "teach baserunning," when I should just have sortable columns for that. Again -- maybe the descriptions are "fuzzy" just like scouting is for players; but I should still be able to quickly locate the kinds of guys I'm looking for. Moreover, I'm not even sure "teach baserunning" is a real thing. I don't know if those hitting instructors will suddenly start looking stupid because of the stiffs I gave them to work with. And finally, I don't know if having a series of smallball managers who call for a lot of bunts will ever help young players develop those skills, or if they'll just keep on popping up to the pitcher, and I should look for guys who can already bunt.

I don't want complete information. I do want meaningful information so I can make decisions.

There's also an 'unorthodox' category of manager. It's very possible the categorizations could be improved, and that's a worthwhile discussion to have.

I think the smart way to do it would be to have the tendencies be

1) Descriptive of the sliders, not coded-in. If I go into the editor and move Joe Maddon's sliders for bunting/stealing/etc all the way over, he should change to "smallball."

2) On 2-3 different axes: smallball vs sabermetrics, uses bench vs. uses starters, etc.