How would you rate the ratings?

[rudel.dietrich]

I also have to admit, I adore threads like this that get behind the curtain so to speak and talk about how the games engine works and how the die rolls work.

On another board I visit there have been some absolutely epic discussions about how the Football Manager engine works and how all of the player ratings, hidden ratings and match engine work with one another to produce results.
They have sometimes spanned over 5,000 posts.

[injury log]

Originally Posted by goalieump413

Despite your opinion that studying whether teams that strike out more score fewer runs is meaningless, it appears that many offensive numbers follow whether a team strikes out a lot or not. And, in a way, you answered my questions yourself, pointing out that teams that do strike out more have lower batting averages. Since there are still 27 (or 24) outs per game, if the ratio of strikeouts per game increases, and batting averages (and other numbers) suffer, then wouldn't that indicate that strikeouts ARE worse outs, having a more profound impact on other offensive aspects?

If your point is that strikeouts are bad, then yes, of course I agree. They're outs, and outs are bad. Your correlation study suggests that more strikeouts leads to worse offense, which I don't think is too surprising (though there are a lot of issues with that kind of study which are a bit too far afield to be worth discussing).

But that's not the question we were originally talking about. We were discussing whether strikeouts were any worse than other kinds of outs. If you have two guys in OOTP with the same Contact rating, those guys make the same number of outs. If they have different AvoidK ratings, they strike out at different rates, but one grounds out or flies out more, the other strikes out more. My point was that those players are of roughly equivalent value, offensively, all else being equal. That is, all those strikeouts don't hurt your offense much more than all those groundouts and fly outs will.

Originally Posted by goalieump413

BTW, I love discussions like this. When I was younger, a buddy of mine and I could talk stats for hours, sometimes never reaching any sort of consensus. No "shoving" here...

Yes, I certainly didn't think you or anyone else in this thread was "shoving" their opinion "down people's throats". But I suspect that comment was directed at me, which I find a bit odd, considering that I wasn't even stating my own opinions in the first place; I linked to an article on a website and basically just summarized what that article says. I hardly feel strongly enough about this issue to care to try to persuade anyone of my opinion, and if I'm continuing the discussion at all, it's only because I think it's interesting.

[injury log]

Originally Posted by rudel.dietrich

I think what you describe as the HR being determined first was how things were coded in versions prior to the OOTP 6.5 rewrite.
I think after 2006 up until today things are different.
Markus has been coy about revealing the die rolls and math behind the games engine.

It's not too difficult to "reverse-engineer" the Contact rating calculation from Power, AvoidK and BABIP (I essentially outlined how to do that in my previous post). When you do that, you find that HR rate is completely independent of K-rate, and of any ball-in-play rates. So the game engine logically needs to make the HR decision first, followed by the K/BB decisions, finally followed by the ball-in-play decision.

[goalieump413]

What fascinates me, as a fairly novice OOTP player, is how well the ratings system paints the picture of player value. When I first started playing, the ratings made some sense, but as an avid baseball fan for decades, translating what I knew about the game into understanding the game's ratings was a little challenging. Discussions like this have helped me gain a firmer grasp of the ratings concepts, so as an acting GM, making better player/personnel movements should be a little more straight forward.

[js1519]

Jut a thought on strikeouts which may have already been mentioned but personally I'd rather take the guy who strikes out less since a ground ball or fly ball is not a guaranteed out. A guy who strikes out isn't putting the ball in play so even if their contact is the same and theoretically will average the same outs, I'll take the lesser strikeout prone player. At least said player will get on base more often, giving him the ability to steal bases and produce runs which is more valuable than a guy who either gets a hit or strikes out the majority of the time. My two cents anyways. Great discussion as others have said though.

[struggles_mightily]

Originally Posted by Questdog

Reaching on errors is possible with two outs or the bases empty.....

Right, as are dropped third strikes. We're talking about marginal stuff here, though, is my point.

[Questdog]

Reaching on an error happens a lot more than reaching on a drop 3rd strike.....

The biggest problem with players that strike out is that they are boring......watching a baseball game is fun, because of the running and diving after balls hit.....strikeouts and walks are tedious.....and the biggest reason that baseball is not the #1 sport in America anymore.....

[VanillaGorilla]

Interesting discussion, on many fronts.

Originally Posted by injury log

But that's not the question we were originally talking about. We were discussing whether strikeouts were any worse than other kinds of outs. If you have two guys in OOTP with the same Contact rating, those guys make the same number of outs. If they have different AvoidK ratings, they strike out at different rates, but one grounds out or flies out more, the other strikes out more. My point was that those players are of roughly equivalent value, offensively, all else being equal. That is, all those strikeouts don't hurt your offense much more than all those groundouts and fly outs will.

To this point, I think the discussion is a bit semantic and causing dispute.

You are absolutely right that an out is an out. A strike out for an out is no different than a pop up for an out is no different from a ground out for an out is no different than a line out for an out. All of this assuming that no runners advance on any of these outs and no double/triple plays occur on the non strikeout outs.

How often runners advance with non strikeouts and how often runners are erased in double/triple plays and how that affects game winning chances...for the sake of discussion, here, let's call it a wash.

Injury log is right that an out is an out is an out.

However, the application of this to RL is flawed due to reducing the comparison of outs, and outs alone, while electing to ignore how the outs were generated and how that affects game winning chances.

Runners on second and third, down 1 and 2 outs. It doesn't matter if your batter strikes out or lines out. The game is over, same difference. But is the strikeout just as good (or equally bad) as the line out? The line drive can become the game winning hit. The strikeout cannot. Therefore, the line drive is better because it has the potential to win the game.

Runners on second and third, down 1 and one out. A strikeout is better than a line-out DP that ends the game. Do we say that the batter that struck out had a better AB than the batter that torched the ball down the third base line requiring a Brooks Robinsonian dive by the third baseman that brought him to rest on the bag? I don't think so.

When parsing data it must be kept in mind what exactly we have removed from the equation to perform our analysis of the parsed data before we stick it back into the original equation, the context of winning a baseball game, in this case, and draw a conclusion that is not applicable to the original question.

It doesn't matter if you strike out or line out with 2 out and 2 men on and down 1. The inning is over, just the same, either way. There is a big difference in potential outcomes between taking a called third strike vs hitting a ball on a rope.

[VanillaGorilla]

Originally Posted by Questdog

Strike out(s)are boring.

Besides that, they are fascist.

[sc_superstar]

Originally Posted by injury log

Who are you talking about?

Why does it need to be one person in particular? The thread was started as poeples opinions to help someone make a spreadsheet for a draft. Then it turned into a Eye vs Aviod K debate. Like I said both have valid points, it's an entirely subjective and situational matter, and there is data to support both sides when viewed from various points, so its really a moot topic that seems pretty far off the OP

[injury log]

Originally Posted by VanillaGorilla

To this point, I think the discussion is a bit semantic and causing dispute.

However, the application of this to RL is flawed due to reducing the comparison of outs, and outs alone, while electing to ignore how the outs were generated and how that affects game winning chances.

This is not a semantic dispute. The question we are discussing is very clearly defined: if you have two OOTP players who are exactly identical except for their AvoidK rating, is one player significantly more valuable than the other? And the answer to that question is 'no'.

You have reframed the question completely. You are answering the question "if I have runners on base with 2 outs in the bottom of the 9th, would I prefer that my batter swing and hit the ball, or that my batter swing and miss?" Obviously you'd prefer your batter hit the ball. No one is debating that question, because no one would ever disagree about that. The question we're discussing is this: if you have two batters with 50/100 ratings everywhere except AvoidK, who would you prefer to have at bat in that situation? And the answer is: it doesn't matter. If they both have 50 Contact ratings, they both will get hits 26% of the time. One of them will make his outs by striking out more, one will make outs by grounding out more. Either way they make outs at the same rate and you lose just as often.

Originally Posted by sc_superstar

Why does it need to be one person in particular? The thread was started as poeples opinions to help someone make a spreadsheet for a draft. Then it turned into a Eye vs Aviod K debate. Like I said both have valid points, it's an entirely subjective and situational matter, and there is data to support both sides when viewed from various points, so its really a moot topic that seems pretty far off the OP

Usually when someone posts in a thread "there's no need to shove your opinion down anyone's throat" it's a reply to something earlier in the thread, and not a non sequitur. But if it was just a random comment unrelated to anything posted previously, then I can agree that no one should be doing that.

But the discussion in this thread is directly on point. The OP asked how important the various ratings are. We are discussing specifically how important the AvoidK rating is, independent of other ratings. I'm not sure why you think this is some kind of off-topic tangent.

[Bobble]

Originally Posted by injury log

...But that's not the question we were originally talking about. We were discussing whether strikeouts were any worse than other kinds of outs. If you have two guys in OOTP with the same Contact rating, those guys make the same number of outs. If they have different AvoidK ratings, they strike out at different rates, but one grounds out or flies out more, the other strikes out more. My point was that those players are of roughly equivalent value, offensively, all else being equal. That is, all those strikeouts don't hurt your offense much more than all those groundouts and fly outs will.

Originally Posted by injury log

if you have two OOTP players who are exactly identical except for their AvoidK rating, is one player significantly more valuable than the other? And the answer to that question is 'no'.

The question we're discussing is this: if you have two batters with 50/100 ratings everywhere except AvoidK, who would you prefer to have at bat in that situation? And the answer is: it doesn't matter. If they both have 50 Contact ratings, they both will get hits 26% of the time. One of them will make his outs by striking out more, one will make outs by grounding out more. Either way they make outs at the same rate and you lose just as often.

I may be misunderstanding you. Are you saying that for all intents and purposes, the Avoid K is cosmetic? It just exists to determine which KIND of out was made? Does it give the opposing defense a few more chances at making errors or tire out the pitcher through numerous foul-offs? Not that either of those is a great value...

[injury log]

Originally Posted by Bobble

I may be misunderstanding you. Are you saying that for all intents and purposes, the Avoid K is cosmetic? It just exists to determine which KIND of out was made? Does it give the opposing defense a few more chances at making errors or tire out the pitcher through numerous foul-offs? Not that either of those is a great value...

It's not cosmetic at all - it influences results. It's just that the result it influences - how often a player strikes out instead of grounding/flying/lining out - just isn't very important, as far as offensive value is concerned. AvoidK tells you nothing about what batting average to expect. That's what the Contact rating tells you.

[Righty Groove]

Well, the way I see it:

Avoid-K: ability to make contact, i.e. not wiff. Put the ball in play.

Contact: quality of contact, control of BABIP

So, a player with high contact and low avoid-k will wiff a lot, but have high BABIPs (think Michael Bourn). A player with high avoid-k and low contact will barely ever strike out, good contact%, but will tend to have low BABIP (think Marco Scutaro).

Both influence the outcomes of balls in play, and the "type" of outs a player generally makes, influencing the player's Avg and OBP.

[VanillaGorilla]

Originally Posted by injury log

This is not a semantic dispute. The question we are discussing is very clearly defined: if you have two OOTP players who are exactly identical except for their AvoidK rating, is one player significantly more valuable than the other? And the answer to that question is 'no'.

\The question we're discussing is this: if you have two batters with 50/100 ratings everywhere except AvoidK, who would you prefer to have at bat in that situation? And the answer is: it doesn't matter. If they both have 50 Contact ratings, they both will get hits 26% of the time. One of them will make his outs by striking out more, one will make outs by grounding out more. Either way they make outs at the same rate and you lose just as often.

There is a semantic discord in the discussion, previously, so it is good to clearly define what we are talking about, as you have done, here.

In the context of the simulation, if the avoid K rate has no impact on the probability of creating contact that results in a base running opportunity, but only contacts that result in outs, then it is of lesser value than I thought*. If the avoid K rate does increase contact that can result in the batter being safe, even if only by an error, then the batter with the better avoid K rate is more valuable. I do not know precisely how the order of operations is set up in OOTP for determining the PA outcomes and just how exactly these ratings intertwine in producing the result.

For the sake of discussion, let's say that avoid K has no impact on a batter reaching base in the sim. Let's take your example, and let's say that the sim works as you say it does (Contact determines likelihood of a base hit and avoid K has zero impact on this likelihood). Runner at second, no outs. Lets say one 50/100 contact player produces all of his outs via strikeout and the other produces none of his outs via strikeout. They are not of the same value, since the player that makes contact with his outs can advance the runner to third without collecting a hit and the player that strikes out cannot.

My understanding is that the contact rating determines the chances of contact being made on a given pitch that results in a hit. A player with a better avoid K rating will foul off 2 strike pitches instead of striking out and, therefore, get more subsequent chances to put a pitch in play than the batter with the same contact rating who strikes out on the 2 strike pitch.

If I am incorrect on this, the fact remains that the player with better avoid K rate is fouling off more two strike pitches, thus forcing the pitcher to throw more pitches. If for this difference, only, the player who has the better avoid K rate than the player who is identical to him in all other ratings, is more valuable as a sim player.

Originally Posted by injury log

Usually when someone posts in a thread "there's no need to shove your opinion down anyone's throat" it's a reply to something earlier in the thread, and not a non sequitur. But if it was just a random comment unrelated to anything posted previously, then I can agree that no one should be doing that.

But the discussion in this thread is directly on point. The OP asked how important the various ratings are. We are discussing specifically how important the AvoidK rating is, independent of other ratings.

Agree completely with the above. This is a good thread.

[Bobble]

Steering the discussion back towards something that helps me personally , who would you rather have:

CON GAP POW EYE K's
_50__75__50__50__50

or:
CON GAP POW EYE K's
_50__50__50__50__75

[snepp]

The top player without question.



Edit: I take that back, I just may question my choice.

Edit 2: Brain cramp, I re-take that back. Definitely the top player.

[Righty Groove]

Originally Posted by Bobble

Steering the discussion back towards something that helps me personally , who would you rather have:

CON GAP POW EYE K's
_50__75__50__50__50

or:
CON GAP POW EYE K's
_50__50__50__50__75

Depends on the scoring enviroment. League modifiers all at 1.000 (default)? I take the top player. But neither is very good.

[snepp]

Here's a fun little exercise I threw together. Not overly complex, but good enough to support what injury_log is saying I think.

I created a two team league, neutralizing everything possible (parks, no injuries, no position player fatigue, etc). They played 100 games, using 4-game series with a 4-man rotation. All starting pitchers had identical ratings, all relievers identical ratings, all hitters had identical ratings with the exception of avoid K and BABIP. One team was given a 75 for avoid K, the other team 125. BABIP was adjusted so that their resulting contact ratings would be identical.

The end of season triple-slash lines were...

.257 / .315 / .403 - the high strikeout team
.251 / .315 / .398 - the low strikeout team

The hit types...

904 H 180 2B 26 3B 93 HR - high strikeout
869 H 136 2B 19 3B 111 HR - low strikeout

The walks, strikeouts, runs...

451 R 266 BB 822 K - high strikeout
466 R 300 BB 494 K - low strikeout

[VanillaGorilla]

Originally Posted by snepp

Here's a fun little exercise I threw together. Not overly complex, but good enough to support what injury_log is saying I think.

I created a two team league, neutralizing everything possible (parks, no injuries, no position player fatigue, etc). They played 100 games, using 4-game series with a 4-man rotation. All starting pitchers had identical ratings, all relievers identical ratings, all hitters had identical ratings with the exception of avoid K and BABIP. One team was given a 75 for avoid K, the other team 125. BABIP was adjusted so that their resulting contact ratings would be identical.

The end of season triple-slash lines were...

.257 / .315 / .403 - the high strikeout team
.251 / .315 / .398 - the low strikeout team

The hit types...

904 H 180 2B 26 3B 93 HR - high strikeout
869 H 136 2B 19 3B 111 HR - low strikeout

The walks, strikeouts, runs...

451 R 266 BB 822 K - high strikeout
466 R 300 BB 494 K - low strikeout

Awesome. I love it when I learn something. I did not know that contact is a resultant figure of Avoid K AND BABIP. It's amazing what you can find in the manual when prodded by an interesting forum discussion!

So, to the point of the OP (and the response that suggested using contact/AvoidK as a base), no matter what weight you choose to give to various ratings to get a one-stop shopping figure for offensive rating, you do NOT want to use Contact AND Avoid K because the avoid K is already figured into the contact number and you are devaluing the more important engine driving number of BABIP as a result.

TYVVM for this post.