+/- Utility

[Tino]

A'ight, my turn.

In economics, when you have a profit-organization, you try to gain as much profit with your company as possible. You want your expenses to be as low as possible, while the gains are as high as possible. This, is something everyone probably knows. As far as I'm concerned, the same concept is applicable when it comes to this particular subject.

Continuing on the economical subject, having more expenses than gains means suffering losses, no matter how small they are. If you have more gains than expenses, you're making profit. If your expenses and gains are equal, you're neither making profit nor suffering losses. I could make it complicated for all of you who know jack shit about economics and talk about balance sheets/scales (whichever term you prefer), but I don't think that's necessary.

This concept immediately suggests that, if a character is being neither a benefit or a detriment to our team, he's being quite average. If a character is being a benefit, he's being above average, and if a character is being a detriment, he's being below average. However, this is way too much of a simplification of the matter. For there are plenty other factors that need to be taken into account. For example, what exactly is average? Are we certain that all characters we think are average actually are average?

Before continuing, I'd like to reply to dondon.

It doesn't make sense that a unit that, although subpar, can do stuff for the team and be considered lower on the tier list the longer that he's around.

You are talking about absolute benefits a unit brings. However, we're always looking at how character's fare relative to eachother. Relativity. This means that if we have three units, where A is one-rounding, B is two-rounding and C is three-rounding, A would be high tier (above average), B would be mid tier (average) and C would be low tier (below average). While C is still contributing through damaging enemies and possibly killing things that were left alive from his earlier battles or earlier battles from B, his offense is still below the average of two-rounding, making him a good deal worse than the others relatively.

Now, we could make this a little more complicated by adding durability to this (note though that this is just looking at numbers, not at actual scenarios, so don't try to dispute what I'm about to say with "but these things happen!"). Now let's say durability and offense are of equal importance for this comparison. We're talking about the same three units. Let's say A gets two-rounded, B gets three-rounded, and C gets five-rounded. Considering this, let's look at how much better (or worse) the characters actually are than the others.

A has two times as much offense as B does.

B has one and a half time as much defense as A does.

Two divided by one and a half equals one and a third, meaning A is 1.33 times as good as B is.

A has three times as much offense as C does.

C has two and a half times as much defense as A does.

Three divided by two and a half equals one and a fifth, meaning A is 1.2 times as good as C.

This already suggests C is suddenly better than B is.

B has one and a half times as much offense as C does.

C has one and two third times as much defense as B does.

One and two third divided by one and a half equals one and a ninth, meaning C is 1.11 times as good as B is.

So A > C > B. Now, we have a few differences. For example, A is 1.2 times as good as C is, while C is 1.11 times as good as B is. Now, do these numbers equate tier gaps? Or are they too close to eachother to guarantee a tier gap? The difference between A and B is definitely large enough for a tier gap. Then let's say the difference between C and B is not enough for a tier gap, our tier list would look like this:

High: A

Mid: C, B

Or you could look at those statistics froma different angle, and say that the average offense is two-rounding, while the average defense is getting 3.33-rounded. In other words, A is two times as good offensively as an average unit, but 1.67 times as bad defensively as an average unit, making him well over average when you average those numbers out, making him high tier material. C is 1.5 times as bad as an average unit offensively, but 1.67 times as good as an average unit defensively. Overall, this makes him slightly above average, but not too much, making him mid tier. Same for B, who is average offensively, and slightly above average defensively, making him mid tier material, although he'll be placed below C.

Same placements, just a different way of looking at it.

With all that said, I guess you'll have noticed that my vote goes to mid tier being roughly neutral utility, while being above mid tier is positive utility and below mid tier is negative utility.

I hope I made sense.

[Gemini-Saga]

My 2 cents:

I only consider negative utility when my unites are hit by a berserk Staff, aside this they are either positive or neutral.

In economics, when you have a profit-organization, you try to gain as much profit with your company as possible. You want your expenses to be as low as possible, while the gains are as high as possible. This, is something everyone probably knows. As far as I'm concerned, the same concept is applicable when it comes to this particular subject.

Continuing on the economical subject, having more expenses than gains means suffering losses, no matter how small they are. If you have more gains than expenses, you're making profit. If your expenses and gains are equal, you're neither making profit nor suffering losses. I could make it complicated for all of you who know jack shit about economics and talk about balance sheets/scales (whichever term you prefer), but I don't think that's necessary.

This concept immediately suggests that, if a character is being neither a benefit or a detriment to our team, he's being quite average. If a character is being a benefit, he's being above average, and if a character is being a detriment, he's being below average. However, this is way too much of a simplification of the matter. For there are plenty other factors that need to be taken into account. For example, what exactly is average? Are we certain that all characters we think are average actually are average?

Talking about economics, I like to use the units with "Nino Arc-type" because they are usually a good investment, I'm always proud of the return.

[CATS]

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).

So I'd agree with Reikken's stance. Only units going down into the lower tiers are bad enough to, on average, always be a significant detriment when actively used.

As for the people who are confounded by the notion of a unit being worse for existing longer, I'd see it more as an issue of that unit generally needing more resources invested in them than the later-joining unit. It's a fact that I can make use of Hawkeye while still using a main team of entirely High and God tier units; Hawkeye can simply be used as a filler sort of unit, to be fielded on larger chapters where there is room for him, and be perfectly helpful. Compare to someone like Fiora, who must be trained in order to be worth anything, and in order to be trained, has to edge other, better units out of a spot on a regular basis.

This is made possible by Hawkeye's high base stats much moreso than the fact that he joins later than Fiora. As another example, take Bartre vs Nino. Bartre clearly is bad and clearly has much, much more availability than Nino, yet he is ranked a tier above her on the list. This is because, despite joining so much later, Nino never has anything useful to offer. Bartre, meanwhile, has use during the early chapters if nothing else.

[Tino]

Talking about economics, I like to use the units with "Nino Arc-type" because they are usually a good investment, I'm always proud of the return.

You must be a bad economist, as Nino-like units come with higher costs than gains.

[dondon151]

Before continuing, I'd like to reply to dondon.

You are talking about absolute benefits a unit brings. However, we're always looking at how character's fare relative to eachother. Relativity.

A unit that is below average can still perform positively. I'll use Bartre as an ideal example again, and let's not consider the fact that he can be replaced by a better character just yet.

A 15/0 Bartre at chapter 20 has 23.5 atk/8.2 AS and 40 HP/7.9 def. That's like under 1.2 levels per map, which is about right if a player is using a character seriously. Some enemy samples (WT already factored):

L7 cavalier (Iron Lance): 28 HP, 15 atk, 9 AS, 7 def

L8 knight (Steel Lance): 27 HP, 18 atk, 2 AS, 12 def

L7 nomad (Iron Bow): 23 HP, 15 atk, 11 AS, 5 def

(If you're wondering about hit rate issues, Bartre has like 77 displayed on the nomad, so it's pretty good)

Now obviously I won't list the Iron Sword cavs because they're crapshoots, but Bartre can just avoid them.

These are all 2RKOs, and Bartre gets like 6RKOd back (4RKO for the 18 atk knight). The only enemies that 2RKO Bartre in this map are a 17+1 atk myrmidon and Cameron. 14/0 Raven ORKOs generics and 15/0 Hector takes no damage from these some of these guys, but it's hard to say in any way that Bartre is actually detrimental in this chapter, though it's much easier to see that he's below average. So how is it that for subpar characters like him, the longer he's around, the worse it makes him?

Though, I must question why you're even using the benchmark of "average." The player who makes optimal character decisions doesn't care about what's average; he cares about what's best. Every character that's not the best has a greater opportunity cost than marginal return; i.e. negative economic profit. But, if a character can kill or use a staff without being killed in return, then he's making a positive normal profit for the team.

Edited August 10, 2009 by dondon151

[Tino]

A unit that is below average can still perform positively.

In one way or another, yes. Being below average doesn't mean you can't perform positively, but that's something different from being negative utility, which is being worse than the average unit.

If it cost you $1,000,000 to develop and launch a certain item on the market, and you know that, if you sell it for price x, you earn $2,500,000 with it, while if you sell it for price y, you earn $3,000,000 with it. Your staff all gave their expectations, and their expectations averaged out to you earning $2,750,000 with the product. Obviously both are incredibly large profits, and it would definitely be nice to obtain a one and a half million profit. Great, in fact. Worth a party. That doesn't mean, though, that the two million dollar profit is lower. It's a profit that's one and a third times as large.

What I'm saying is, while something may not be a completely negative contribution, if it's below average it's still a negative. If your profits are below your expectations, that means something went wrong.

The player who makes optimal character decisions doesn't care about what's average; he cares about what's best.

That's if we're talking about optimal character decisions, yes. However, tier lists or normal debates never suggest an aim for optimal character decisions. They're merely to see which characters are better than the others.

Every character that's not the best has a greater opportunity cost than marginal return; i.e. negative economic profit.

Basically, what you're suggesting is "this character's the best, the rest can rot on the bench". In other words, let's just keep soloing the game with Titania, since all others only bring negativities with them.

Obviously this doesn't make sense. Negative economic profit is aquired through having more expenses than profit. In other words, those who have the exact same expenses as gains (i.e. average) don't bring anything positive, but nothing negative, either. Those who have more expenses bring something negative. Those who have more profit bring something positive. If you look at just those two, it's even clearer why everything in-between is neither positive nor negative. The average of -1 and 1 is still 0, or neutral/average, for that matter.

Edited August 10, 2009 by Tino

[Narga_Rocks]

Every character that's not the best has a greater opportunity cost than marginal return; i.e. negative economic profit.

Basically, what you're suggesting is "this character's the best, the rest can rot on the bench". In other words, let's just keep soloing the game with Titania, since all others only bring negativities with them.

Obviously this doesn't make sense. Negative economic profit is aquired through having more expenses than profit. In other words, those who have the exact same expenses as gains (i.e. average) don't bring anything positive, but nothing negative, either. Those who have more expenses bring something negative. Those who have more profit bring something positive. If you look at just those two, it's even clearer why everything in-between is neither positive nor negative. The average of -1 and 1 is still 0, or neutral/average, for that matter.

Well, the thing is if you have 10 great characters and 20 not so great, but only 10 slots, why are we comparing over the average? We can only use 10 slots. The 11th best is kicking out the 10th best. The opportunity cost of deploying the 11th best is the profit of the 10th best. The 11th best units economic profit is it's own profit - the opportunity cost of using it, not it's own profit - the average. Subtracting the average just seems wrong to me. For economic profit, you subtract the best thing you could've had otherwise. Which was higher. Therefore, 11th best is negative. Which is why I really don't like the idea of negative utility at all. Still, you can't just handwave it by saying it doesn't help build tier lists and redifine economic profit to make a more useful concept of negative economic profit. If you bring economic profit into this, why not do it the proper way, which is profit - opportunity cost. Negative economic profit is acquired by preventing a better unit from going out there, in this case. That's not suggesting to solo it, just max out deployment with the best, and any of the other units are obviously negative because they are kicking out their betters.

[Vykan12]

Well, the thing is if you have 10 great characters and 20 not so great, but only 10 slots, why are we comparing over the average? We can only use 10 slots.

Did you read Cats' post?

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.

[Narga_Rocks]

Well, the thing is if you have 10 great characters and 20 not so great, but only 10 slots, why are we comparing over the average? We can only use 10 slots.

Did you read Cats' post?

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.

What are you, smash now? Assuming people don't read? I disagree. Of course they aren't always getting used, it just doesn't change the whole opportunity cost thing. As for the RNG, isn't a tier list supposed to mostly assume average levels instead of getting RNG screwed? The not clear who the top 10 are is fine, except a tier list that differentiates has already made that statement. Even if you assume the top 10 aren't known, we often have more than 30 units and yet still only 10 slots, so when you have 40 units I'd like to think at least the top 15 are known, at which point we can already differentiate between the "average" and what we know carries an opportunity cost higher than the rewards. And if there is debate over, say, characters 7 to 15 on how good each is, then all of them effectively give neutral economic profit, since they give a certain utility, but prevent another from giving the same utility, hence, 0 economic profit and neutrality. That's still a far cry from saying anything above the average is positive and the average is neutral.

Edited August 11, 2009 by Narga_Rocks

[Rewjeo]

I meant to put mid tier is roughly neutral, but I accidentally voted other. But yeah, I think people understand the ideas behind that.

[CATS]

Of course they aren't always getting used, it just doesn't change the whole opportunity cost thing.

The thing that I keep seeing is something along the lines of "it's horrible to assume that only the top 10 units are ever used and anyone else who is used has negative utility!" That's what I'm referring to when I say that the top 10 units aren't always getting used, so using units below the top 10 doesn't always create negative utility. If you're referring to something else, then forgive me, but since you're addressing my post, I'm forced to assume.

As for the RNG, isn't a tier list supposed to mostly assume average levels instead of getting RNG screwed?

When directly comparing two characters, sure. When talking about "these same ten units will ALWAYS be used," then no, one must acknowledge the fact that the RNG can and often will cock-block you in that respect.

The not clear who the top 10 are is fine, except a tier list that differentiates has already made that statement.

Which tier lists have their tiers completely ordered, just out of curiosity?

Tier lists frequently change anyways. Many of the factors that go into them are subjective (i.e. the value of healers vs fighters, thief abilities, flying for a unit such as FE8 Vanessa, etc), and thus it's difficult to objectively prove beyond all doubts that whatever units are definitely the top 10.

[Narga_Rocks]

Of course they aren't always getting used, it just doesn't change the whole opportunity cost thing.

The thing that I keep seeing is something along the lines of "it's horrible to assume that only the top 10 units are ever used and anyone else who is used has negative utility!" That's what I'm referring to when I say that the top 10 units aren't always getting used, so using units below the top 10 doesn't always create negative utility.

I've already said I don't like calling anything less than the top, say, 15, or whatever, as negative utility. My point is, if we are going to relate negative utility to economic profit (tino, not you) then let's do it right. I don't like the negative utility thing as a whole, though. I've already said if they don't hold us back then anything that comes out of a given unit is a positive or neutral at worst. Which is why in general I disagree with simply stating mid tier and being done with it, or low tier or whatever.

If you're referring to something else, then forgive me, but since you're addressing my post, I'm forced to assume.

I think I'm in this situation because I was addressing tino's post and vykan asked if I'd even read yours. So I felt compelled to respond to yours. I actually like this statement of yours:

Only units going down into the lower tiers are bad enough to, on average, always be a significant detriment when actively used.

Which you said was Reikken's stance, so I guess I like that, too. But there are times in which those lower tiers aren't that bad, which is why I especially like the use of "on average", because there are times they in fact help.

As for the RNG, isn't a tier list supposed to mostly assume average levels instead of getting RNG screwed?

When directly comparing two characters, sure. When talking about "these same ten units will ALWAYS be used," then no, one must acknowledge the fact that the RNG can and often will cock-block you in that respect.

The same ten units aren't always going to have the best stats if used, no, I agree. I'm just saying on average, x units are the best, so anything worse is likely stopping us from using something better, on average. Comparing to the average is pointless where economic profit is concerned. Tino cheated a bit by using only two different options of gaining profit. If a company has 10 different options, and there are say 7 above the average (possible if the bottom 3 are really low and the other 7 aren't as much above the average as those three are below) then by Tino's statement all 7 of those options would be positive. A real company wants the best. If they can get it and know which one will give it, then their choice is obvious.

The not clear who the top 10 are is fine, except a tier list that differentiates has already made that statement.

Which tier lists have their tiers completely ordered, just out of curiosity?

Well, the RD and PoR ones are trying to order it. Some of the others just dump them all in tiers and don't order the individual tiers. Which still makes a group of like 8 units between 7th and 14th, or something. At which point, by economic profit all 8 of those would basically be neutral, the top 6 would be positive, everything else would be negative, if the number of slots available is between 7 and 13.

Tier lists frequently change anyways. Many of the factors that go into them are subjective (i.e. the value of healers vs fighters, thief abilities, flying for a unit such as FE8 Vanessa, etc), and thus it's difficult to objectively prove beyond all doubts that whatever units are definitely the top 10.

Yep, which is why it's good that these things aren't set in stone. The real trouble with comparing negative utility to the tier list, though, is because some units take like one or two chapters to stop being a detriment, then are average for like 4, then are awesome for 10 or more chapters afterwards. If that unit gets placed in upper mid, then the whole "anything below mid is negative" would not adequately describe that unit's early situation. There could be a unit in mid that is average for the entire game. Other times a unit is good for a bit and then starts being really bad, or a negative throughout the rest of it's existence, and as such is placed in lower mid. It is then looked at as a negative? This whole utility thing is chapter by chapter, so I can't agree with option A, because units don't compare to other units exactly the same way throughout their entire existence.

And also defining negative utility based off the tier list has other issues. If you make a tier list by placing units as you meet them, then you can have a unit for 20 chapters that you can see isn't as good as many other units, then you stick it in lower mid because it takes time for him to be average or even above average. Then you get a unit that is in 3 chapters and is basically average. At which point you say "well, he's better than the guy in lower mid because by being in lower mid the guy was giving negative utility for 17 chapters and this new guy never got that". Except we are basing a unit's negative utility on the position we just gave the unit. It's circular reasoning. WE placed him in a spot and thus he gets negative utility, therefore some other unit that is almost the same is better despite not helping for any serious length of time and he's also not even as good as the better units anyway.

I went with other.

Edited August 11, 2009 by Narga_Rocks

[Rehab]

Saying a unit has negative utility doesn't make sense to me unless negative utility means you cannot get through the game/meet your objective, be it maximum bexp or ranks or whatever, using them. If they're useful enough to win chapters, they have to have positive utility, just top tier units have obscenely more, so tiering based on efficiency can't have a unit with negative utility, just shit scarce utility.

Edited August 11, 2009 by Rehab

[dondon151]

In one way or another, yes. Being below average doesn't mean you can't perform positively, but that's something different from being negative utility, which is being worse than the average unit.

Why does the character have to be worse than average? Economics never define anything with regard to the average; economic profit is defined with regard to the best and normal profit is defined with regard to the base line of 0.

What I'm saying is, while something may not be a completely negative contribution, if it's below average it's still a negative. If your profits are below your expectations, that means something went wrong.

The thing about your scenario is that if the dude sells for 2.5 mil instead of 3.0 mil, his economic profit is -0.5 mil because he chose the second best option, and that's why it's negative. Not because it's "below average."

That's if we're talking about optimal character decisions, yes. However, tier lists or normal debates never suggest an aim for optimal character decisions. They're merely to see which characters are better than the others.

This has nothing to do with the principles on how utility should be judged. Though, if you think that tier lists don't suggest optimal character decisions, then that's a perfect argument for why we should only consider normal profit, i.e. ignore opportunity costs.

Basically, what you're suggesting is "this character's the best, the rest can rot on the bench". In other words, let's just keep soloing the game with Titania, since all others only bring negativities with them.

That's not what I said at all. In fact, this is my reason why I'm opposed to judging characters on the basis of economic profit.

Obviously this doesn't make sense. Negative economic profit is aquired through having more expenses than profit.

No. Negative economic profit is acquired through being inferior to a better option.

The thing that I keep seeing is something along the lines of "it's horrible to assume that only the top 10 units are ever used and anyone else who is used has negative utility!" That's what I'm referring to when I say that the top 10 units aren't always getting used, so using units below the top 10 doesn't always create negative utility. If you're referring to something else, then forgive me, but since you're addressing my post, I'm forced to assume.

Exactly, hence why we should always judge characters assuming no opportunity costs!

Saying a unit has negative utility doesn't make sense to me unless negative utility means you cannot get through the game/meet your objective, be it maximum bexp or ranks or whatever, using them. If they're useful enough to win chapters, they have to have positive utility, just top tier units have obscenely more, so tiering based on efficiency can't have a unit with negative utility, just shit scarce utility.

Almost exactly what I had in mind, just expressed in a different way.

[CATS]

Exactly, hence why we should always judge characters assuming no opportunity costs!

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. There are units who are bad enough that they will very rarely if ever be worth using or training, even if they do provide limited use in a few situations; for example, Bartre is obviously useful during FE7's earlygame just because there's no one else to be using, but rarely would you ever keep him around for the long term. By Ch. 16 at the latest, his time has passed.

This is still more useful than what a unit like Nino or Vaida has to offer (Vaida is useful only as a filler unit for SoT or VoD, which I find considerably less significant than being helpful during the early chapters), despite the fact that those units have much less availability, so the misguided notion that "comparisons between anyone who isn't in the top 10 are decided just by who's available longer!" is also defeated.

[Quasar]

Voted for mid tier. Technically I think it should be in relation to who is your average unit at given time, but mid tier is a start. To be honest I don't really care as long as a decision is made.

[Narga_Rocks]

Exactly, hence why we should always judge characters assuming no opportunity costs!

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.

From what I'm reading, he's not saying anything remotely close to that. He's saying look at normal profit, not economic profit. How do you get "any unit can be used with the same results" from that? Look at a unit, see how well it does. Other units do better, therefore they go higher on the list. See how I differenciated between units without ever including negative utility? We can use this concept for any of the characters. You don't need to create some arbitrary 0 line and make everything beneath negative. Just see how much they profit us. Clearly some will profit us more than others.

There are units who are bad enough that they will very rarely if ever be worth using or training, even if they do provide limited use in a few situations; for example, Bartre is obviously useful during FE7's earlygame just because there's no one else to be using, but rarely would you ever keep him around for the long term. By Ch. 16 at the latest, his time has passed.

And don't you think their "normal profit" would be less than other units at some point? We don't need to call it negative utility to see who is better and who is worse. I said to smash once, I don't need to see a "-" to know a unit is worse than another unit.

This is still more useful than what a unit like Nino or Vaida has to offer (Vaida is useful only as a filler unit for SoT or VoD, which I find considerably less significant than being helpful during the early chapters), despite the fact that those units have much less availability, so the misguided notion that "comparisons between anyone who isn't in the top 10 are decided just by who's available longer!" is also defeated.

And their normal profit would be even less than Bartre's, then. It is simple enough to ignore opportunity cost and just look at normal profit. The only issue is that it rewards availability a little much at times, but that's fine since negative utility has the tendency to punish availability far more than it should.

[CATS]

From what I'm reading, he's not saying anything remotely close to that. He's saying look at normal profit, not economic profit. How do you get "any unit can be used with the same results" from that? Look at a unit, see how well it does. Other units do better, therefore they go higher on the list. See how I differenciated between units without ever including negative utility? We can use this concept for any of the characters. You don't need to create some arbitrary 0 line and make everything beneath negative. Just see how much they profit us. Clearly some will profit us more than others.

So, when comparing Neimi to Eir Innes, do you simply assume that Neimi is fielded for every chapter before Innes joins and totally ignore the fact that a better unit could've been taking her spot?

[dondon151]

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.

There are units who are bad enough that they will very rarely if ever be worth using or training, even if they do provide limited use in a few situations; for example, Bartre is obviously useful during FE7's earlygame just because there's no one else to be using, but rarely would you ever keep him around for the long term. By Ch. 16 at the latest, his time has passed.

Bartre in chapter 20 is 2RKOing and being 4RKO'd back, which isn't bad at all. Bartre's weakest time is earlygame when he's doubled by more enemy types and his durability isn't that good because of poor defense.

This is still more useful than what a unit like Nino or Vaida has to offer (Vaida is useful only as a filler unit for SoT or VoD, which I find considerably less significant than being helpful during the early chapters),

Vaida's performance isn't that bad - she doubles almost everything in chapter 31 and is ridiculously durable against non-magic units, and she's like 1 or 2 atk away from ORKOing enemy generals IIRC.

so the misguided notion that "comparisons between anyone who isn't in the top 10 are decided just by who's available longer!" is also defeated.

I don't think you understood what I meant. If we were to take 2 subpar units of equal quality, the one that is available longer would be considered worse. The units you listed don't conform to these prerequisites (and you said so yourself).

So, when comparing Neimi to Eir Innes, do you simply assume that Neimi is fielded for every chapter before Innes joins and totally ignore the fact that a better unit could've been taking her spot?

Yes. Do you simply assume that after Innes joins, he's fielded for every chapter and totally ignore the fact that a better unit could've been taking his spot?

[CATS]

Yes.

So then, you're choosing to ignore a blatantly obvious fact. Why?

If we were to take 2 subpar units of equal quality, the one that is available longer would be considered worse. The units you listed don't conform to these prerequisites (and you said so yourself).

Can you give an example of this within the actual tier lists?

Edited August 14, 2009 by CATS

[dondon151]

So then, you're choosing to ignore a blatantly obvious fact. Why?

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.

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?

If I were to suddenly reverse my stance and only rank based on economic profit, I would then be ignoring the blatantly obvious fact that units can, you know, do stuff that aids toward completion of the game.

Can you give an example of this within the actual tier lists?

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.

Edited August 14, 2009 by dondon151

[Quasar]

Meh, I forgot that I already begun to hate negative utility once. I'm deserting from mid tier and joining forces with no negative utility. This has nothing to do with which is the most logical form for tier lists but purely based on what is most interesting to argue about. All methods have their flaws, there's no perfect choice that everyone will like and thus no negative utility and every unit being fielded in every chapter where available is the best option in my opinion. I'd also add that their supports will always be fielded but that might be pushing it. Supports are overrated anyway.

Ignoring logical fallacies of all methods the reasoning to why this is more entertaining is in the flaws negative utility brings to debating. In most cases bad characters gather negative utility outside of their forced period, thus the most logical way to play them when arguing them is "can't be negative while forced and once it's over we'll just bench him/her". Great, why not just scratch one third off from every list, there's practically no difference how well sub-par people play if they're only fielded in a single chapter and "don't care, do some damage and go die". Removing negative utility you'd have to examine the combat prowess and/or general utility of the character until the end and can't end the case in a single sentence.

Another example is a unit good early on but gets weak later. Let's see, theoritical example of Marcus vs. Zealot. They both get obsolete around the same time and negative utility forces them to the bench. Now this is not the case in reality but let's say Zealot would be significantly better than Marcus after they go to "negative". This would be completely disregarded and Marcus would steamroll over Z. If their performance as a whole would be looked at it wouldn't be so obvious that Marcus is the winner. Also clearly more interesting than Marcus > Zealot before Zealot joins, Marcus=Zealot when Zealot joins and after they get benched -> Marcus > Zealot.

Some of this might already be said, I didn't exactly read all of the thread.

[Reikken]

I made a post in regards to that before on another site...

For units like Dorcas, and even more extremely, Marcus, they're actually good starting out, for a length of time, until later, where they become meh/bad. To evaluate them, you can either consider the part where they're good and assume they get dropped after that, or assume they get used for the entire time, or some combination of both. I think that it should be a combination of both, but with how much weight the good part is given depending on the magnitude of that "good"ness. Like for Marcus, it's huge, so his placement is essentially 100% used early and then dropped when no longer good. Dorcas, while far from Marcus, is also quite high up there in this area. He's good, takes no unit slots for a good while, and it lasts for several chapters. However on the other extreme, like Wil for example, he has just one chapter (he's good because he's forced in the chapter and so not taking a slot, and he covers a part of the map that no one else can really cover), and not much in it either. He helps kill a couple of pegasus knights or something, and that's it. That's insignificantly small, so almost no weight is given to it at all.

I know doing it this way can get rather iffy for things that fall midway between the two extremes, but fortunately not much of that exists.

[General Spoon]

Negative utility doesn't exist. Anyone can be a sacrifical lamb.

[Interceptor]

I've voted that negative utlity does not exist, not because I think that all units have positive utility, but because I think that the concept of negative utility as applied to FE tier lists is a hamfisted way to solve a problem that doesn't really exist.

Tier goals are sufficient for anything you want to do, and if they aren't, you did it wrong. If we're making a list that is supposed to be ranking units based on efficient game completion with low turn counts, as an example, it's straightforward enough to tier units based on their contributions towards that goal (so long as efficiency is appropriately defined). There's this magic phrase called "opportunity cost" that will tell me everything that I need to know. It may be complicated, but the basic idea is clear: the vast majority of actions in this hypothetical scenario can be adjudicated just by looking at what they cost you in terms of the tier goals.

I have no problem with wacky criteria for ranking in tier list, because sometimes a straight-up "efficiency" goal will result in match-ups that don't lead to interesting debates (see, for example, the hilarity of trying to compare Fiona to Lyre in an FE10 efficiency tier list, since the only correct move for both of them is not to deploy them at all). But the point here is that the criteria need to be clearly stated up-front.

You could, for example, take that hypothetical "efficient completion and low turn counts" tier list and tack on a sub-requirement that stipulates that all units being compared are deployed in every situation where it is possible to do so. Mandating deployment is obviously in clear contradiction with the goal of efficient play, but it allows you to compare shit-tier units against each other in more interesting ways than "Meg shoves a heron, then hides in a corner for 12 turns". It's up to the people making the tier list if they want to sacrifice accuracy in the name of expanding the range of interesting match-ups (which is what would happen; let's call a spade a spade).

tl;dr: the fact that this thread exists and few can agree suggests that the concept of negative utility is not useful for tiering. We already have perfectly good ways to rank units in the form of clear tier goals, and opportunity cost.