+/- Utility

[Reikken]

negative utility = opportunity cost - utility(costs aside)

(positive utility = utility - opportunity cost)

If you want to pretend it doesn't exist, ok..... but it's there. The only issue is how you deal with it in tier lists. Sure, you can have tier lists where you assume that the opportunity cost is always kept low enough that neg utility never comes into play, but that's only for those particular lists that it effectively does not exist.

Edited August 16, 2009 by Reikken

[Narga_Rocks]

negative utility = opportunity cost - utility(costs aside)

(positive utility = utility - opportunity cost)

If you want to pretend it doesn't exist, ok..... but it's there. The only issue is how you deal with it in tier lists. Sure, you can have tier lists where you assume that the opportunity cost is always kept low enough that neg utility never comes into play, but that's only for those particular lists that it effectively does not exist.

Wow, I guess you made negative utility a positive quantity. It's weird seeing opportunity cost - utility rather than the other way around.

Anyway, the trouble with doing this is what Dondon and I have been saying. The opportunity cost is the best alternative forgone. The opportunity cost is not the average. If you are going to define negative utility your way, then anyone who is not in the top x units for x slots is automatically a negative. The opportunity cost of deploying that unit is the utility of the better unit it is displacing. If that unit is better, well then surely it's utility would be greater than any utility from the inferior unit. Thus, negative utility. So in a game in which you constantly have say 7 slots and more than 7 units, the 8th best unit will build up so much negative utility over the game for being worse than the 7th best unit but being deployed anyway that after 20 or 30 chapters a unit around for 3 could be obviously worse than unit 8 but end up higher on the tier list because it build up less negative utility over time. That is what happens if you use opportunity cost properly. If you hand wave opportunity cost to make it the average, then you can avoid stupid things happening like the 8th best unit in the theoretical game ranking worse than the 20th best unit. But then it is no longer actually opportunity cost.

And if you can handwave it and make the opportunity cost the average, I can handwave it and make the opportunity cost null. At this point we compare what they give us and whoever gives more ends up better.

[Interceptor]

If you want to pretend it doesn't exist, ok..... but it's there. The only issue is how you deal with it in tier lists. Sure, you can have tier lists where you assume that the opportunity cost is always kept low enough that neg utility never comes into play, but that's only for those particular lists that it effectively does not exist.

Did you read past the first eight words in my post?

Edited August 16, 2009 by Interceptor

[CATS]

Try Wil vs. Vaida or something, though finding an example for FE7 is kind of difficult because even crappy units earlygame turn great by endgame.

How does it apply to Wil vs Vaida? Wil's ranked underneath Vaida, so are you suggesting that were we to pretend negative utility does not exist, that suddenly Wil would be ranked over Vaida? Just trying to understand here as I still feel quite confused by your viewpoint. Apparently you're not pretending that any unit can be used with the same results, yet you also seem to be refusing to acknowledge the lesser results created by using weaker units.

Because acknowledging this fact would mean that in order to place characters on a tier list, we would be contradicting the most basic principle of a tier list.

Eh? What is this "most basic principle" of which you speak?

There is really nothing different between having more negative economic profit and having less positive normal profit. The difference in each between a good unit and a worse unit is the exact same; the only question up for debate is, does the worse unit become more negative or more positive over time?

So it's like a "is the glass half full or half empty" sort of question? Hm. Well, if there's no difference except in what you call it, then I don't see why we're all arguing so much over it......I was assuming that people like you were standing by this "normal economic profit" viewpoint because it would actually have some sort of an effect on tier lists.

Edited August 16, 2009 by CATS

[Reikken]

If you want to pretend it doesn't exist, ok..... but it's there. The only issue is how you deal with it in tier lists. Sure, you can have tier lists where you assume that the opportunity cost is always kept low enough that neg utility never comes into play, but that's only for those particular lists that it effectively does not exist.

Did you read past the first eight words in my post?

Indeed I did, sir. But did you read past the first two sentences in mine?

Edited August 17, 2009 by Reikken

[dondon151]

How does it apply to Wil vs Vaida? Wil's ranked underneath Vaida, so are you suggesting that were we to pretend negative utility does not exist, that suddenly Wil would be ranked over Vaida? Just trying to understand here as I still feel quite confused by your viewpoint.

Yes, if we didn't consider Wil's existence to be negative before he can comfortably double most enemies, then he would be above Vaida. His "good" period begins long before Vaida joins

Apparently you're not pretending that any unit can be used with the same results, yet you also seem to be refusing to acknowledge the lesser results created by using weaker units.

What? You can't do one without the other. Of course I'm saying Bartre doesn't yield as good results as Raven, but Bartre still helps the team.

Eh? What is this "most basic principle" of which you speak?

The tier list is created under the premise of high efficiency play yet it entails selecting negatively efficient units for comparison.

So it's like a "is the glass half full or half empty" sort of question? Hm. Well, if there's no difference except in what you call it, then I don't see why we're all arguing so much over it......I was assuming that people like you were standing by this "normal economic profit" viewpoint because it would actually have some sort of an effect on tier lists.

You and I have been participating in the Karla debate for awhile now on the FE7 tier list, so you should know how the utility viewpoint would affect the placement of characters. Namely, Karla's recruitment cost will not be a negative if fielding Bartre helps the team more than not fielding Bartre.

I'm not going to rehash the Quasar/Interceptor argument over why tiering inefficient characters involve playing them in totally retarded ways. Although I should point out that under that system, characters try to get as close to zero utility as possible, therefore the baseline still happens at zero and goes up from there.

[Rehab]

To some of you, negative utility seems to be sucking exp out of better people, making it impossible to achieve the squeakiest-clean outcome that would be possible if a better unit were used. To me, there would have to actually be something the unit does to make the game more difficult to complete, other units notwithstanding, for me to call them negatively useful. Like, using this unit will result in excessively manly Dracoknioght/Druid/Paladin/Warrior reinforcements bursting out of the worst possible position in every chapter and actively make the game more difficult to complete.

At the end of the day no matter how bad a unit is, they still maim and murder dudes and could have rescue utility. That may not be better than anybody else on the list, but if you think it's worse than nothing, as in not having that unit slot available, you're full of it.

Edited August 17, 2009 by Rehab

[Interceptor]

Indeed I did, sir.

Then the problem isn't that you didn't read my post, but that you didn't comprehend the point that I was making. The point is that people are never going to agree on what negative utility is, so you're better off taking care of it at the tier goal level.

Tier lists are not character guides -- they lack all sorts of important information necessary for that purpose -- and so accuracy isn't serving any useful purpose. Pure efficient-playthrough lists often result in nonsense character comparisons that don't facilitate any sort of interesting debate (in addition to my example, look at Narga's).

If there was an option in the poll that said "dogmatic obsession with accuracy makes tier lists boring, negative utility should be killed in the crib", I would have probably clicked on that one.

But did you read past the first two sentences in mine?

I did, but didn't see anything useful. Eliminating this argument is not just something that you CAN do, it's something that you probably SHOULD do.

[CATS]

Yes, if we didn't consider Wil's existence to be negative before he can comfortably double most enemies, then he would be above Vaida. His "good" period begins long before Vaida joins

So then, it sounds to me as if the "negative utility doesn't exist" viewpoint simply reverses the effect of availability on tiering. You've complained that using negative utility in arguments "makes availability too much of an issue" (or something along these lines), but how is making availability a total non-issue any better? Now when we have two sub-par but equal units, the one who joins earlier will just be automatically considered better, as he'll be on the higher level when the other one shows up.

What? You can't do one without the other. Of course I'm saying Bartre doesn't yield as good results as Raven, but Bartre still helps the team.

Well, you specifically said you're not doing the first one:

Saying that "using units below the top 10 doesn't always create negative utility" is very different from saying that any unit can be used with the same results, as appears to be your stance.

It's not my stance, lol. I don't know where you could have gotten that thought from.

So if you can't do one without the other?

Anyways, sure, Bartre helps more than an empty slot. But this is where your argument stops, in my eyes, because Bartre doesn't have a slot with his name pasted on it that no one else can take. It isn't a question of bringing Bartre or not bringing Bartre, it's a question of bringing Bartre or bringing someone else. These are simply the facts of the game--I'm not sure whether you're refusing to acknowledge this and saying that it isn't true, or whether you're acknowledging it and then saying that we should just pretend otherwise regardless. I think it's the latter, though.

The tier list is created under the premise of high efficiency play yet it entails selecting negatively efficient units for comparison.

And the acknowledgement of the existence of negative utility contradicts this?

You and I have been participating in the Karla debate for awhile now on the FE7 tier list, so you should know how the utility viewpoint would affect the placement of characters. Namely, Karla's recruitment cost will not be a negative if fielding Bartre helps the team more than not fielding Bartre.

In that case, could the Nino side of the argument not say that Nino reaches level 10 or 15 or something and promotes before Karla joins, and thus Nino's "good period" begins before Karla shows up? Or would that be considered favoritism?

Edited August 17, 2009 by CATS

[Narga_Rocks]

Anyways, sure, Bartre helps more than an empty slot. But this is where your argument stops, in my eyes, because Bartre doesn't have a slot with his name pasted on it that no one else can take. It isn't a question of bringing Bartre or not bringing Bartre, it's a question of bringing Bartre or bringing someone else. These are simply the facts of the game--I'm not sure whether you're refusing to acknowledge this and saying that it isn't true, or whether you're acknowledging it and then saying that we should just pretend otherwise regardless. I think it's the latter, though.

Didn't you pick mid tier? Because that is not an argument for mid tier. That is an argument for "Assuming there's a perfect team choice, anyone who doesn't belong to it has negative utility of some sort". You might notice that you are effectively saying that if you can bring someone better it is a negative to bring the lesser character. Dondon and I are saying it should possibly be ignored.

Besides, with negative utility, if one character isn't very good but still has the stats to be slightly useful but always competes for a slot with his/her betters, then isn't that unit always going to be negative? And if another unit with no redeemable qualities is forced in all it's chapters, and thus never competes for anything, wouldn't that unit not build up any negative ever? So then our theoretical unit that is absolutely irredeemable ends up higher on the list than someone that is actually halfway competent, just not enough to earn that spot on the list.

Say this theoretical forced unit dies in one hit and usually has a 50% hit rate on everything to do 2 damage. (enemies have 30 to 60 hp, say)

The other unit is 2 or 3 3RKOd by everything (when the rest of our group is 5 or 6 RKOd), never doubles, and usually has an 80% hit rate to do 10 to 20 damage (when the rest of our group doubles and tends to ORKO or 3HKO). Doesn't have any unique skills or superior movement or anything like that to make up for it.

The forced horrible unit automatically ends up higher on the list than the other unit because the horrible unit never builds negative (keep it away from enemies, and in this game say a monkey could keep it away from enemies) and the other unit builds negative just for being deployed.

Oh, and if anyone actually agrees that the irredeemable unit should be above the other guy, then I don't know. Say once you get the other guy to a certain level he gets a massive stat increase (say a weapon that boosts stats but only if he's at a certain level) and makes the rest of the game super easy. He was under average for 20 chapters or so and then once he's at a certain level he's the best there is. Say this weapon appears the moment he reaches that level and it's reasonable to have that level when there are 5 chapters left. Now there is a reason to go through his subpar stage. I don't know if that changes anything for anyone who initially wants the irredeemable unit to end up higher.

Edited August 17, 2009 by Narga_Rocks

[dondon151]

You've complained that using negative utility in arguments "makes availability too much of an issue" (or something along these lines), but how is making availability a total non-issue any better?

My goal wasn't to make availability a non-issue; it was to prevent availability from working against perfectly usable units.

Now when we have two sub-par but equal units, the one who joins earlier will just be automatically considered better, as he'll be on the higher level when the other one shows up.

What's wrong with that? I should point out that the unit who joins earlier cannot possibly be more negative than the unit who joins later even when negative utility is considered, since the most efficient way to use them is to not use them at all.

Well, you specifically said you're not doing the first one:

So if you can't do one without the other?

Where am I contradicting myself here?

And the acknowledgement of the existence of negative utility contradicts this?

Well, yeah, because characters that will affect the team negatively are never deployed according to principle. Thus, at a certain point in the tier list, every character below that point will be technically equal - between the options of taking up zero resources and doing nothing or having a greater opportunity cost than normal profit, each one of them will choose the former because it maximizes their efficient use. Even if you stipulate "you can't do that!" it's bound to come up as an argument anyway. Where else do you think the half-assed arguments in the FE10 tier list of "Meg shoves a heron and waits in a corner for the rest of the map > Fiona sucking" come from? A debater will always resort to this kind of bullshit if it helps his argument.

In that case, could the Nino side of the argument not say that Nino reaches level 10 or 15 or something and promotes before Karla joins, and thus Nino's "good period" begins before Karla shows up? Or would that be considered favoritism?

Note that I'm saying that the only opportunity cost that should be ignored is a deployment slot; everything else is up for grabs. The reason behind this logic is linked to the generally accepted assumption that when a unit is being debated, he's assumed to be played, and this assumption implies that the unit can't be replaced.

Anyway, it depends on how good Nino's "good period" really is and if it's better than Karla's. And, if Karla's recruitment cost isn't negative, then does she get credit for it?

Edited August 17, 2009 by dondon151

[Narga_Rocks]

And the acknowledgement of the existence of negative utility contradicts this?

Well, yeah, because characters that will affect the team negatively are never deployed according to principle. Thus, at a certain point in the tier list, every character below that point will be technically equal - between the options of taking up zero resources and doing nothing or having a greater opportunity cost than normal profit, each one of them will choose the former because it maximizes their efficient use. Even if you stipulate "you can't do that!" it's bound to come up as an argument anyway. Where else do you think the half-assed arguments in the FE10 tier list of "Meg shoves a heron and waits in a corner for the rest of the map > Fiona sucking" come from? A debater will always resort to this kind of bullshit if it helps his argument.

Actually, it is perfectly reasonable to talk about Meg and Fiona that way. They have chapters where they don't take slots away from others. Meg has more. Compare what they can do in their free chapters. Although, Meg can't shove a heron in her free chapters since there is no heron in her free chapters. From a truly efficient stand point, they only get deployed in their free chapters, and the winner is whoever does more in their 6 vs. 3 chapters of "availability". Unfortunately, this results in Fiona/Meg > Kyza, since Kyza is always subpar and never gets a free slot. Well, in the first half of part 4 slots are almost free, but if they were never deployed pre-part 4 it's hard to do much there. Kyza gets some shoving in 4-3 desert or something. Maybe some unit might get "burning out the elsleep bishop's uses" utility or something. It does make for some rather uninteresting comparisons and interesting conclusions, though.

Which is why things like negative utility and assuming deployment in all chapters are silly concepts. Assuming deployment is counterproductive to the whole efficiency thing since if we are trying to be efficient then they should be able to choose not to be deployed. To not be silly, they can choose to not be deployed. At which point, negative utility comes in and means if they choose to be deployed they just go down, so nobody will ever be deployed if the opportunity cost of being deployed outweighs the benefits during the chapter. Then anyone with access to free slots automatically gets bumped above the units at 0 utility. Thus, Meg/Astrid/Fiona > Kyza/Oliver/Renning/Stefan/Volke.

In order to avoid that, we do something silly like saying we are trying to be efficient but let's deploy the worst unit in the game in every single chapter anyway, or we can just drop the whole "negative utility for being deployed" thing.

Now we look at things like:

What resources do I have to use to make this person useful?

How much benefit does this unit gain compared to some other unit with the same stuff?

Does having this unit hide in a corner make the chapter easier than trying to let this unit do something?

There are of course lots of other questions, but the idea is there is still an opportunity cost for doing everything else, they still give benefits for doing various things, the costs can still outweigh the benefits, and we can still judge units on who is better.

Availability won't automatically make a unit better. If the cost of raising the unit or making it useful outweighs the benefits received, then it could still be worse. But if they are equally bad, it might be best to focus kills elsewhere and simply run utility. At which point the earlier unit could be better.

Also, they can now choose to be not deployed at some point without messing things up entirely. See, since everyone starts at 0 and works up (mostly), being not deployed at some point now means you gain no more points. So you can deploy a unit while it's good, stop deploying it when it's bad to do so, and thus accomplish the tier list goal of efficiency rather well, but if it's only pretty good for 2 or 3 chapters then bad for the rest (Tormod) we still know that it is worse than a unit that is somewhat good for 7 or 8 chapters (I'm sure some unit fits here) because that unit built up more points. This allows us to see that Kyza > Fiona without forcing us to do something that runs counter to the efficiency goal of the tier list.

The bolded statement is the one thing I want to get across above all else. Assuming deployment in every single chapter of availability seems counterproductive. If we are trying to be efficient in a tier list, shouldn't we look at the most efficient way possible to use a unit? Anything else is mutually exclusive with the tier list goal.

[Reikken]

Indeed I did, sir.

Then the problem isn't that you didn't read my post, but that you didn't comprehend the point that I was making.

I believe I understood your post just fine. This topic isn't about tier goals; it's about the definition of negative utility. Save it for the tier topics themselves, as any specific implementation of negative utility will depend on the particular list's goals. That's what I was getting at.

[CATS]

Didn't you pick mid tier? Because that is not an argument for mid tier. That is an argument for "Assuming there's a perfect team choice, anyone who doesn't belong to it has negative utility of some sort".

Notice the "assuming there's a perfect team choice." I'll just quote myself here:

I maintain that the top 10 units aren't always going to be used, if only because of the RNG; it's perfectly possible for someone such as Eliwood to get crappy levels, while a Upper Mid unit such as Erk or Dorcas gets good levels during Lyn Mode, and ends up getting used instead. More importantly, in some cases it isn't even clear who the top 10 are. In the FE7 list there are only 3 god tier units, then 11 high tier units, and the tiers are not ordered. And a few of these (Pent / Harken and to a lesser extent Ninian) are late joiners, while Matthew won't always be fielded after he hits 20, so you'll be using other units to fill their spots at various points in the game (for example Hector; though clearly not in the top 10 overall, he also clearly provides positive utility during the earlygame).

In other words, afaik, there is no clearly set, objectively determined, universally agreed, etc "perfect team choice." It's more about units who are bad enough that they will almost certainly be among the undesirable choices, which usually entails units who are falling into the lower tiers. It's more difficult to determine exactly who is "good enough" that they are almost certainly being used in any given situation, aside from a few god tier units.

That said, yes, I do maintain that if at some point in the game there's a clearly obvious superior team choice (can't really think of when this would be, I guess the early chapters of FE8 it's still pretty obvious who's good and who isn't), a unit who displaces one of those superior units is lowering your overall performance. This is, as far as I can see, perfectly logical and quite true. How it applies to tiers is probably up to interpretation, but the simple fact in and of itself is pretty obvious.

I still can't see what the big protest against this is, other than a general sentiment of people just not liking this fact, which doesn't mean anything as that's just people's personal opinions.

You might notice that you are effectively saying that if you can bring someone better it is a negative to bring the lesser character. Dondon and I are saying it should possibly be ignored.

Yeah, I know that's your position. Why do you want it to be ignored?

Besides, with negative utility, if one character isn't very good but still has the stats to be slightly useful but always competes for a slot with his/her betters, then isn't that unit always going to be negative? And if another unit with no redeemable qualities is forced in all it's chapters, and thus never competes for anything, wouldn't that unit not build up any negative ever? So then our theoretical unit that is absolutely irredeemable ends up higher on the list than someone that is actually halfway competent, just not enough to earn that spot on the list.

Say this theoretical forced unit dies in one hit and usually has a 50% hit rate on everything to do 2 damage. (enemies have 30 to 60 hp, say)

The other unit is 2 or 3 3RKOd by everything (when the rest of our group is 5 or 6 RKOd), never doubles, and usually has an 80% hit rate to do 10 to 20 damage (when the rest of our group doubles and tends to ORKO or 3HKO). Doesn't have any unique skills or superior movement or anything like that to make up for it.

The forced horrible unit automatically ends up higher on the list than the other unit because the horrible unit never builds negative (keep it away from enemies, and in this game say a monkey could keep it away from enemies) and the other unit builds negative just for being deployed.

All perfectly true. In your example, the forcing of that unit is a circumstance imposed on you by the game, and thus it must be factored into tier discussions and comparisons, just like everything else. You might not like it, but that has nothing to do with logical and objective discussion of it. I personally don't like Seth and I want him to go die in a fire, but that doesn't change the fact that he's the best unit. I also imagine that your theoretical forced unit would barely be higher than the non-forced unit, if any at all; if he's this bad, then while he isn't hurting you, he's not contributing anything either, which is about the same as that other guy with better-but-not-quite-good-enough stats just sitting on the bench.

And either way, it doesn't change the basic fact of what happens with slots that aren't reserved for anyone and the units that compete for them, so w/e.

Oh, and if anyone actually agrees that the irredeemable unit should be above the other guy, then I don't know.

You talk as if agreeing with the example is criminal. But why did you not post any logic against it?

My goal wasn't to make availability a non-issue; it was to prevent availability from working against perfectly usable units.

Call it whatever you want. Either way, from what I can see, the effects are these:

Now when we have two sub-par but equal units, the one who joins earlier will just be automatically considered better, as he'll be on the higher level when the other one shows up.

Am I wrong, and if not, then how is this better than the situation which you seem to deplore so highly, where "when given two equally subpar units the one who joins later is automatically better?"

Which also isn't even true btw--it depends on the levels that these units are joining at. If they both join on the same level, then the guy who joins earlier will still get ranked higher regardless of how you look at it. Or at worst, they'd be ranked about the same, with the guy who joins earlier still being placed just above the other one since he'll do better in the unlikely event that you do decide to use him. If the later-joining unit comes in on a higher level with higher stats, on the other hand, then we no longer have "two equal units," so the whole example fails anyway.

Well, yeah, because characters that will affect the team negatively are never deployed according to principle. Thus, at a certain point in the tier list, every character below that point will be technically equal - between the options of taking up zero resources and doing nothing or having a greater opportunity cost than normal profit, each one of them will choose the former because it maximizes their efficient use. Even if you stipulate "you can't do that!" it's bound to come up as an argument anyway. Where else do you think the half-assed arguments in the FE10 tier list of "Meg shoves a heron and waits in a corner for the rest of the map > Fiona sucking" come from? A debater will always resort to this kind of bullshit if it helps his argument.

Indeed. You look at how a unit can contribute to an efficient playthrough. Efficient playthroughs are clearly assumed (correct me if I'm wrong), and thus you look first at what happens on efficient playthroughs. If that makes no difference or distinction, then you can go down to looking at less efficient playthroughs where these bad units are actively used, just for discussion's sake if nothing else (similar to how Lord's seizing is disregarded as an argument purely for the sake of discussion).

[Narga_Rocks]

Didn't you pick mid tier? Because that is not an argument for mid tier. That is an argument for "Assuming there's a perfect team choice, anyone who doesn't belong to it has negative utility of some sort".

Notice the "assuming there's a perfect team choice." I'll just quote myself here:

I maintain that the top 10 units aren't always going to be used, if only because of the RNG; it's perfectly possible for someone such as Eliwood to get crappy levels, while a Upper Mid unit such as Erk or Dorcas gets good levels during Lyn Mode, and ends up getting used instead. More importantly, in some cases it isn't even clear who the top 10 are. In the FE7 list there are only 3 god tier units, then 11 high tier units, and the tiers are not ordered. And a few of these (Pent / Harken and to a lesser extent Ninian) are late joiners, while Matthew won't always be fielded after he hits 20, so you'll be using other units to fill their spots at various points in the game (for example Hector; though clearly not in the top 10 overall, he also clearly provides positive utility during the earlygame).

In other words, afaik, there is no clearly set, objectively determined, universally agreed, etc "perfect team choice." It's more about units who are bad enough that they will almost certainly be among the undesirable choices, which usually entails units who are falling into the lower tiers. It's more difficult to determine exactly who is "good enough" that they are almost certainly being used in any given situation, aside from a few god tier units.

Well, the thing about the option is the way it's worded. It actually doesn't matter that you usually don't know the best units for the specific number of slots. The fact is, the option assumes you do know.

But even if you don't know the specific best units and order, we have already grouped people in different tier levels. Take the PoR tier list. That one is actually ordered even within the levels, but say it wasn't. There are 20 units between the top and the bottom of upper mid tier. There are never more than 20 deployment slots in PoR. Ever. Therefore, when there are x deployment slots in a chapter and x of those 20 units can be selected, the line between positive and negative utility would be between mid and upper mid. If there are times in which there are y deployment slots and <y of those 20 units are available, then for those chapters the dividing line could be between lower mid and mid. Calling mid tier roughly neutral utility would be incorrect because most of the time it is negative.

That said, yes, I do maintain that if at some point in the game there's a clearly obvious superior team choice (can't really think of when this would be, I guess the early chapters of FE8 it's still pretty obvious who's good and who isn't), a unit who displaces one of those superior units is lowering your overall performance. This is, as far as I can see, perfectly logical and quite true. How it applies to tiers is probably up to interpretation, but the simple fact in and of itself is pretty obvious.

Well, I did a thing on diminishing marginal returns earlier. The point is, once you have 7 or 8 really good characters and there are 13 slots, another 4 or 5 really good units don't really have much to do anymore. 4 or 5 decent units in most maps can cover what those same 4 or 5 units are doing anyway, so calling them negative simply because their stats aren't as good seems silly since they are doing the same thing anyway. In some cases all 13 units need to be good and anything less will slow you down, but that doesn't happen very often. It is quite possible for a low tier unit to fill a sufficient role that it isn't an issue to deploy it instead of some upper mid unit. Maybe not for every single turn in every single map, but enough that it is very much positive overall. That's why I take issue with applying a blanket negative utility to anyone under a certain level.

I still can't see what the big protest against this is, other than a general sentiment of people just not liking this fact, which doesn't mean anything as that's just people's personal opinions.

Just because a unit that is around for 20 chapters and slightly subpar is worse than a unit that is around for 3 chapters yet majorly subpar under this standpoint.

That's really what I have against simply judging a unit and saying "wow, you are bad, negative utility. You are automatically worse than this other unit that you outperform in every way simply because you are around longer."

I don't actually know what I'd say for unit A being around for 20 chapters and being slightly subpar for all of them compared to that slightly subpar unit for 3 chapters. It really depends on how subpar "slightly" is. If they are only subpar enough that they are still helpful just others are more helpful, I don't see why being able to do something for longer is considered worse than the guy who can only do it for 3 chapters. If by slightly I don't actually mean slightly but should be saying "rather" subpar, or "very subpar", then I can maybe accept that being around for less time might just mean something. Except that the "rather" subpar unit is still doing something for longer. It might not do as well as the other guy, but for 20 chapters it is rescuing or shoving or gumming or whatever. At this point, if we have say 11 deployment slots but could finish the game rather well with 8 raised characters then doing all that utility stuff allows us to focus exp on those 8 units better. Really, it might not be as good as one of the other options, but if it can do that for 6 or 7 chapters before it's level falls too far behind the enemies then surely it's doing better than the similar unit that is only around for 3 chapters.

So I still have issues when the units are close in performance level simply because they still have more time to do something, anything. Calling it negative utility and saying they build up so much more just seems unfair when they can do more towards an efficient playthrough than Mr. 3 chapters.

You might notice that you are effectively saying that if you can bring someone better it is a negative to bring the lesser character. Dondon and I are saying it should possibly be ignored.

Yeah, I know that's your position. Why do you want it to be ignored?

Because the conclusions stemming from that standpoint don't make much sense when considering an efficient playthrough. Therefore something must need fixing, and this seems as good a way as any.

Besides, with negative utility, ... just for being deployed.

All perfectly true. In your example, the forcing of that unit is a circumstance imposed on you by the game, and thus it must be factored into tier discussions and comparisons, just like everything else. You might not like it, but that has nothing to do with logical and objective discussion of it. I personally don't like Seth and I want him to go die in a fire, but that doesn't change the fact that he's the best unit. I also imagine that your theoretical forced unit would barely be higher than the non-forced unit, if any at all; if he's this bad, then while he isn't hurting you, he's not contributing anything either, which is about the same as that other guy with better-but-not-quite-good-enough stats just sitting on the bench.

And either way, it doesn't change the basic fact of what happens with slots that aren't reserved for anyone and the units that compete for them, so w/e.

Oh, and if anyone actually agrees that the irredeemable unit should be above the other guy, then I don't know.

You talk as if agreeing with the example is criminal. But why did you not post any logic against it?

Sorry, I guess it could do with some rewording. That other unit is rather bad compared to what it competes against, so it isn't criminal to say he's worse than Mr. Useless. I also didn't delve into enough of the other possibilities that the forced unit can do.

Really, I should focus on the units that are 5RKOd instead of 6RKOd and that 3HKO the enemies instead of 2HKO the enemies.

13 units are the really good ones that are 6RKOd and 2HKO the enemies (while doubling, so 1RKO).

10 units are the sorta good ones that are 5RKOd and 3HKO the enemies (while doubling, so 2RKO).

1 unit is 3RKOd and 3 to 5 HKOs enemies but doesn't double.

So, assuming all of them do the same things where everything else is concerned then obviously my tier list looks like this:

high tier

those 13

mid tier

those 10

low tier

that 1

Now say every chapter has 8 deployment slots.

Where do we stick Mr. "I suck but I'm forced" (aka Mr. Useless)?

I don't have a big problem with putting him in the top of low instead of the bottom, I suppose, but I really think that the whole "negative utility" thing as I understand it from an economics standpoint should place this guy at the top of mid. That I have a problem with.

Am I wrong, and if not, then how is this better than the situation which you seem to deplore so highly, where "when given two equally subpar units the one who joins later is automatically better?"

Which also isn't even true btw--it depends on the levels that these units are joining at. If they both join on the same level, then the guy who joins earlier will still get ranked higher regardless of how you look at it. Or at worst, they'd be ranked about the same, with the guy who joins earlier still being placed just above the other one since he'll do better in the unlikely event that you do decide to use him. If the later-joining unit comes in on a higher level with higher stats, on the other hand, then we no longer have "two equal units," so the whole example fails anyway.

Well, I think we are assuming that if the earlier unit is around for say 10 more chapters, then the later unit has an appropriate level for the difference in chapters. Then our issue is #2 is automatically better than #1.

Well, yeah, because characters that will affect the team negatively are never deployed according to principle. Thus, at a certain point in the tier list, every character below that point will be technically equal - between the options of taking up zero resources and doing nothing or having a greater opportunity cost than normal profit, each one of them will choose the former because it maximizes their efficient use. Even if you stipulate "you can't do that!" it's bound to come up as an argument anyway. Where else do you think the half-assed arguments in the FE10 tier list of "Meg shoves a heron and waits in a corner for the rest of the map > Fiona sucking" come from? A debater will always resort to this kind of bullshit if it helps his argument.

Indeed. You look at how a unit can contribute to an efficient playthrough. Efficient playthroughs are clearly assumed (correct me if I'm wrong), and thus you look first at what happens on efficient playthroughs. If that makes no difference or distinction, then you can go down to looking at less efficient playthroughs where these bad units are actively used, just for discussion's sake if nothing else (similar to how Lord's seizing is disregarded as an argument purely for the sake of discussion).

But if we do that, then a unit that is forced in a couple of chapters is better than a subpar character that isn't ever forced. See, the whole "If that makes no difference or distinction" line isn't always true. Take RD, Kyza is never forced. Kyza is very much subpar. Pretty much everyone can see that there are many better options for deployment than Kyza. Fiona is not forced but is not competing for a slot in any part 3 chapter. There are x slots, and x possible units for those slots. There is no cost whatsoever in deploying her. With negative utility in mind, all she has to do is use a torch item once in 3-6 and she's now better than Kyza. She has a slight positive, he has 0. There is a difference now, so there is no reason to compare them in a less efficient playthrough because Fiona is better than Kyza in the only playthrough that is supposed to matter: the efficient one.

The fact that Kyza might be better in a less efficient playthrough should be irrelevant because she's better in the optimal playthrough.

This also highlights my issue with negative utility.

[Zkirsche]

"wow, you are bad, negative utility. You are automatically worse than this other unit that you outperform in every way simply because you are around longer."

It depends. See, the unit that's around for longer might need to use resources to become better than the unit who joins much later on. They are taking these resources from units who can actually use them well. This is negative utility. What has to be decided is if the resources the earlier unit gets outweighs the margin in which the later unit is beaten b the earlier unit.

That's the way i see it, anyway.

Edited August 17, 2009 by kirsche

[Narga_Rocks]

"wow, you are bad, negative utility. You are automatically worse than this other unit that you outperform in every way simply because you are around longer."

It depends. See, the unit that's around for longer might need to use resources to become better than the unit who joins much later on. They are taking these resources from units who can actually use them well. This is negative utility. What has to be decided is if the resources the earlier unit gets outweighs the margin in which the later unit is beaten b the earlier unit.

That's the way i see it, anyway.

I agree it depends.

If the unit needs massive amounts of bexp or important stat boosters or something to become usable then usually that would mean it would be worse. What I meant by that statement, though, would be simply getting the unit kills and achieving a normal level and then be able to outperform in every way. Just imagine Rolf vs. Lyre if Lyre started in part 4. Up her stats a bit and give her S strike to start so that she is where she would be for part 4 if we did try to use her in part 3, of course, but I'm pretty sure a normally raised Rolf will Nuke that Lyre enough to count as "outperform in every way". Obviously Rolf needed to find a way to kill things in part 3 and we had to use "resources" defined by enemy kills to get him there, and he is "taking" those kills from others, but considering it improves his performance as well I think it's okay. However, since he was being a negative (deploying him rather than better units) until at least 3-11 and so that's 9 chapters of "negative" and 2 chapters of "neutral" and by part 4 he's likely now doing "okay" for 2 chapters. That's a lot of "negative" built up, and Lyre is only sucking for 2 chapters. How can Lyre possibly build up enough negative in 2 chapters for Rolf to actually win this thing? (Oh, and if I'm wrong about the raised Rolf nuking Lyre in part 4 then this comparison doesn't apply. Also assume that there is no 4-E.)

[dondon151]

Yeah, I know that's your position. Why do you want it to be ignored?

Why do you want to ignore the fact that virtually any unit helps toward game completion?

Now when we have two sub-par but equal units, the one who joins earlier will just be automatically considered better, as he'll be on the higher level when the other one shows up.

Am I wrong, and if not, then how is this better than the situation which you seem to deplore so highly, where "when given two equally subpar units the one who joins later is automatically better?"

Uh, because the earlier joining unit has been helping you complete the game for a longer period of time? With regard to the rest of this section that I didn't bother to quote, the later joining unit will have the same parameters as what's expected of the earlier joining unit, i.e. equal.

Indeed. You look at how a unit can contribute to an efficient playthrough. Efficient playthroughs are clearly assumed (correct me if I'm wrong), and thus you look first at what happens on efficient playthroughs. If that makes no difference or distinction, then you can go down to looking at less efficient playthroughs where these bad units are actively used, just for discussion's sake if nothing else (similar to how Lord's seizing is disregarded as an argument purely for the sake of discussion).

The discussion of the "less efficient playthrough" will inevitably regress back to a discussion of the "most efficient playthrough" where the debaters are trying their hardest to not let units with negative utility be deployed or take up combat positions or whatever.

It depends. See, the unit that's around for longer might need to use resources to become better than the unit who joins much later on. They are taking these resources from units who can actually use them well. This is negative utility. What has to be decided is if the resources the earlier unit gets outweighs the margin in which the later unit is beaten b the earlier unit.

That's the way i see it, anyway.

I said earlier that the only opportunity cost that I propose we ignore is the deployment slot. Every other resource still has to be considered and weighed.

[-Cynthia-]

Hmm, are you suggesting that we ignore slot situations? I'm not so sure about this, because it seems to ignore that units sometimes push out others for spots, whereas forced units allow you to deploy someone "extra".

[Rehab]

Yes, ignore slot situations. They doesn't necessitate classifying units in a way that doesn't make sense in context. That's not negative utility, again, that's lesser utility.

[Quasar]

Since the problem is that

a) there are a bunch of morons who seem to think that worse units should be cannon fodder or face negative utility

and

B) there is a league of idiots who seem to think that playing worse units doesn't affect the credibility of efficiency

the one immediate solution is to create separate efficiency tier lists for cases a) and B). Seriously though, I don't think a single solution can just be agreed on - it's just a matter of view after all. I'd sure like to discuss characters in non-negative utility lists with people sharing that view and folks who like "x doesn't necessarily get played" and "is worse than your average unit and isn't fielded" can enjoy shorter and more vague arguments in their own lists.

[Interceptor]

I believe I understood your post just fine.

Truly? Because what you just said in your post, what you were "getting at", is what I've been saying already. So, I guess were agreeing with me...

This topic isn't about tier goals; it's about the definition of negative utility. Save it for the tier topics themselves, as any specific implementation of negative utility will depend on the particular list's goals. That's what I was getting at.

This isn't an economics forum, this is General Fire Emblem, where talking about negative utility means we're talking about its impact on tiering. There is no reason to be discussing it otherwise. The OP asks the question "How should +/- utility be determined?", looking for consensus on the issue. Again, for the purposes of tiering: look no further than options #1 and #3 as evidence.

My position is that the answer is not to find a consensus on what negative utility means WRT tier lists, but to recognize that it generally introduces really awkward situations that make debate silly, aka find a way to eliminate it (my preference is via tier goals, so everything is clear up front).

I realize that this idea is anathema to accuracy nazis, but I don't particularly care.

[CATS]

Well, the thing about the option is the way it's worded. It actually doesn't matter that you usually don't know the best units for the specific number of slots. The fact is, the option assumes you do know.

Ofcourse it matters. That option is invalid when there is no clear perfect team choice. In that case, I'm leaning more towards the mid tier = neutral side of things. If we have a god tier of 10 units, then two tier gaps between the rest of the units in the game, and 10 deployment slots available for every chapter, and all 10 god tiers join in the first chapter, then yes, you have an obvious perfect team choice. But that's not how it is.

But even if you don't know the specific best units and order, we have already grouped people in different tier levels. Take the PoR tier list. That one is actually ordered even within the levels, but say it wasn't. There are 20 units between the top and the bottom of upper mid tier. There are never more than 20 deployment slots in PoR. Ever. Therefore, when there are x deployment slots in a chapter and x of those 20 units can be selected, the line between positive and negative utility would be between mid and upper mid. If there are times in which there are y deployment slots and <y of those 20 units are available, then for those chapters the dividing line could be between lower mid and mid. Calling mid tier roughly neutral utility would be incorrect because most of the time it is negative.

And in that case I'd more likely choose upper mid = neutral utility. Unfortunately that isn't an option, and I haven't played FE9 anyway so w/e. Regardless, I hardly think this discussion should be centered around the definitions and validity of the poll choices. Rather, it's about whether or not the existence of negative utility is a fact, and if it is, then why should it be ignored.

Well, I did a thing on diminishing marginal returns earlier. The point is, once you have 7 or 8 really good characters and there are 13 slots, another 4 or 5 really good units don't really have much to do anymore. 4 or 5 decent units in most maps can cover what those same 4 or 5 units are doing anyway, so calling them negative simply because their stats aren't as good seems silly since they are doing the same thing anyway. In some cases all 13 units need to be good and anything less will slow you down, but that doesn't happen very often. It is quite possible for a low tier unit to fill a sufficient role that it isn't an issue to deploy it instead of some upper mid unit. Maybe not for every single turn in every single map, but enough that it is very much positive overall. That's why I take issue with applying a blanket negative utility to anyone under a certain level.

I don't see how this means anything.

By the time that everyone becomes good, your chosen team of units has generally progressed in levels to the point that no one else aside from some later joining prepromos is worth fielding (and those filler prepromo guys get points on their own for being usable without requiring any resources invested in them). Yes, Bartre and Raven can perform similarly against hordes of really crappy enemies in the lategame. But both of them need lots of training earlier in the game to reach that point, and Bartre sucks alot more during that earlier time, so using him is still alot worse than using Raven no matter how similar they may be later in the game. And it's true that they're never equal, even very late. There are still promoted enemies mixed in with the lulzy ones, where the differences between Bartre and Raven are still very apparent.

Just because a unit that is around for 20 chapters and slightly subpar is worse than a unit that is around for 3 chapters yet majorly subpar under this standpoint.

Only if the latter unit is able to do something useful during those 3 chapters that they're around which the other guy can't match in any of the 20 chapters that he's around. Otherwise they're still worse. Bartre is still rated above Nino despite Nino joining much later and both of them sucking enough to not be actively used.

Calling it negative utility and saying they build up so much more just seems unfair when they can do more towards an efficient playthrough than Mr. 3 chapters.

If they can do more towards an efficient playthrough, then they're obviously not worse. It's only negative utility if you use them poorly and inefficiently, for example fielding Wil all the time and trying to feed him kills on every turn, even if it slows you down in order to do so. You don't have to assume that this is what's happening. This whole thing stemmed from the Karla debaet in the FE7 tier list, which is a special case as far as I can see. It's one of the few instances (possibly the only one?) in which you do have to assume that this is happening. You are required to feed Bartre levels and a promotion item; there is no way around it.

I don't have a big problem with putting him in the top of low instead of the bottom, I suppose, but I really think that the whole "negative utility" thing as I understand it from an economics standpoint should place this guy at the top of mid. That I have a problem with.

Why? What's the logic against it?

Well, I think we are assuming that if the earlier unit is around for say 10 more chapters, then the later unit has an appropriate level for the difference in chapters. Then our issue is #2 is automatically better than #1.

I repeat, if you have #1 (early joiner) and #2 (later joiner) joining on different levels, then they're not equal in any sense of the word. A higher level coupled with (more importantly) higher base stats is an advantage in and of itself, as demonstrated by Seth / Marcus / Titania / etc. #2 is better because he has higher base stats and is thus more useful without requiring resource investment. It has nothing to do with availability. Indeed, if you put them on the same level, then #1 wins precisely because he joins earlier. And if #2 is on a higher level and wins, then making him join earlier would obviously just make him win by even more. In no instances is it shown that greater availability is a disadvantage in and of itself. So I don't really see what your example is aiming to prove.

But if we do that, then a unit that is forced in a couple of chapters is better than a subpar character that isn't ever forced. See, the whole "If that makes no difference or distinction" line isn't always true. Take RD, Kyza is never forced. Kyza is very much subpar. Pretty much everyone can see that there are many better options for deployment than Kyza. Fiona is not forced but is not competing for a slot in any part 3 chapter. There are x slots, and x possible units for those slots. There is no cost whatsoever in deploying her. With negative utility in mind, all she has to do is use a torch item once in 3-6 and she's now better than Kyza. She has a slight positive, he has 0. There is a difference now, so there is no reason to compare them in a less efficient playthrough because Fiona is better than Kyza in the only playthrough that is supposed to matter: the efficient one.

The fact that Kyza might be better in a less efficient playthrough should be irrelevant because she's better in the optimal playthrough.

This also highlights my issue with negative utility.

What is your problem with this scenario? Do you just plain dislike it?

It's not like discussion of what happens when the units are actively used couldn't still happen. It's just that one would also have to acknowledge what they are able to contribute without taking away from anything else.

Why do you want to ignore the fact that virtually any unit helps toward game completion?

Don't deflect. "Any unit helps toward game completion" isn't being ignored. Yes, anyone helps more than an empty slot. So what? You don't have empty slots. The question is, why pretend that you do?

Uh, because the earlier joining unit has been helping you complete the game for a longer period of time?

The tier list is about efficiency, not merely completing the game. You can easily complete the game even if it takes you 3000 turns and all characters other than your main lords, healers and jeigan are killed. Using Bartre over Raven doesn't help you complete the game efficiently, quite the opposite, and I believe it's quite clear that tier lists assume efficient play. Again, if I'm wrong, correct me.

The discussion of the "less efficient playthrough" will inevitably regress back to a discussion of the "most efficient playthrough" where the debaters are trying their hardest to not let units with negative utility be deployed or take up combat positions or whatever.

If that discussion is relevant, then yes, ofcourse. If not, and neither unit is contributing anything worth mentioning to an efficient playthrough, then presumably this would've already been acknowledged before moving on to discussion of inefficient playthroughs. What's your point?

[dondon151]

Don't deflect. "Any unit helps toward game completion" isn't being ignored. Yes, anyone helps more than an empty slot. So what? You don't have empty slots. The question is, why pretend that you do?

If the most efficient way to play a unit is to not play him at all, then you are ignoring any possible contributions he may have to the team. The reason why I'm responding to your question with a question of my own is because neither issue can be simultaneously addressed with whatever principle of utility that tier lists will be ranked upon. One of them has to be ignored.

The tier list is about efficiency, not merely completing the game. You can easily complete the game even if it takes you 3000 turns and all characters other than your main lords, healers and jeigan are killed. Using Bartre over Raven doesn't help you complete the game efficiently, quite the opposite, and I believe it's quite clear that tier lists assume efficient play. Again, if I'm wrong, correct me.

Efficient play and optimal character choice aren't the same thing. And if they are, then even considering the negatively efficient characters is contradictory to the principle of the tier list.

If that discussion is relevant, then yes, ofcourse. If not, and neither unit is contributing anything worth mentioning to an efficient playthrough, then presumably this would've already been acknowledged before moving on to discussion of inefficient playthroughs. What's your point?

My point is that the debaters will not move on to a discussion of inefficient playthroughs. Why would they? It will still be a contest of who can be less negatively efficient, and the discussion will inevitably regress to talking about an efficient playthrough.

I'd like to bring up again whatever Quasar or Narga_Rocks said earlier about how diminishing marginal returns make a character's overall positive or negative utility rather small. This means that the overall effect of availability on a unit's position in the tier list is not that significant; i.e. having 2x the availability doesn't mean that it can only be balanced out with 2x the performance.

Edited August 19, 2009 by dondon151

[Dat Nick]

I think we've strayed way off course with this unit slot shit, because as stated in my post, going by that logic would mean everybody in Mid tier is negative utility :/

Speaking of my post, I think we could do well to look back on it.

Negative utility is pretty much any unit that can't pull it's own weight on the field rather than making worse use of a unit slot because otherwise anybody not in High tier would be negative utility. Let's use Minerva in FEDS as a random example: She needs to be shielded from enemy fire as Sniper, but she makes up for needing to be fenced with solid offense for a good portion of the game: Her positives outweigh her negatives, so we can consider her a good unit who is pulling her own weight fine, i.e positive utility. A good example of negative utility is primarily found in underlevelled characters: Let's use Rolf as, again, a random example. Rolf needs to be shielded which is not only a bigger issue in PoR than in FEDS since you have a lot of ubergods who can soak up enemy hits like nuts in PoR so actually needing to be shielded stands out a lot more: Rolf's chip damage for most of his existance is so pitiful that it's not worth all the hassle it is to fence him in as anybody who's doing so could have actually been, you know, doing stuff. For most of the game your team is better off without trying to baby Rolf than with it: Making Rolf negative utility.

And believe me, I know from personal experience: Trying to use Ryan in FE3 makes the game MUCH harder than it actually is.

A summarized version: Let's go back to actual unit performances without giving two shits who uses a slot or not. Est and Fiona aren't godawful units because they use slots, after all.

Yes, anyone helps more than an empty slot.

See, this is where the whole "negative utility argument doesn't exist" argument crumbles: Because this simply isn't true. Using FEDS examples again, take Gordin. Gordin's doing nothing but helping in C1-3 because he doesn't eat up a slot, fine. But the further the game goes, the more trying to keep Gordin alive and trying to utilize his shit chip damage [5 base str and 20% growth and never doubling ever? Yuck] becomes a billion times more of a hassle than it really is. And it never gets better: Gordin will just suck up a ton of resources and not make any use of them as he's going to suck the whole game, demand protection from your other units who could have had better things to do like kill this guy or wall X guy who's about to die instead, and guess what? His damage is still balls and it's just not worth the time, EXP, and effort it takes to field him past C3 [if he's kept as archer]: I.e he's pretty much negative utility past chapter 3.

Don't even try the "but everyone else like Cord/Zag/Wolf use EXP lol" argument because it just doesn't fly: Cord/Zag/Wolf/Ogma/Dolph/Catria/Minerva/Radd for fick's sake all actually improve and make some use out of the EXP they're given and can eventually become positive contributions...will Gordin ever be able to outweigh his negatives? Not without a crapload of boosters, no.

Edited August 19, 2009 by Miyamoto Powers