Tiering Philosophy - It's that time again

[Espinosa]

Edward in FE10 saves around 30-40 turns due to 1-P. Does he deserve to be a couple tiers above Haar?

Clearly not. That's why I thought of complexity in the first place, to come up with an objective solution.

It hasn't been a big problem for the FE10 tier list anyway.

Suppose the first chapter was, on the contrary, highly complex, and Edward still played a key role in efficiently clearing it - would anything change about our perception of his contribution to an efficient playthrough? Probably not a whole lot.

[Chiki]

It hasn't been a big problem for the FE10 tier list anyway.

Suppose the first chapter was, on the contrary, highly complex, and Edward still played a key role in efficiently clearing it - would anything change about our perception of his contribution to an efficient playthrough? Probably not a whole lot.

It hasn't been a problem for that tier list when it should. You guys don't really lay out how you tier units, and certainly don't think over it too much. You simply have a vague idea of how to.

I think it would make Edward's contributions seem far more impressive, but remember that even one chapter with high complexity is not as valuable as many chapters with high complexity. Units that contribute to more chapters are more valuable. I think that matches with our intuitions too.

Think of 5 chapters with 1/5 complexity. Think of 1 chapter with 5/5 complexity. I think those 5 chapters are worth as much as that one chapter with 5/5 complexity. So complexity does not concern a chapter alone, but rather many chapters.

Edited April 29, 2013 by Olwen

[Espinosa]

5 * 1/5 = 1 * 5/5 seems mathematically sound enough, but where you get those numbers from is dubious in itself and ultimately cannot be objectively measured (I raised some important questions for practical application of rating chapters by complexity, which I believe you said you would think over but never really examined properly).

I can agree though that chapters with extra complexity take more time to figure out and so the participants should receive more credit, e.g. Louise is better than both Wil and Rebecca because Louise, if deployed at all, can contribute to harder endgame chapters in Hector Hard Mode, and if you think of Rath then you will find certain unique functions that nobody in the party can really replicate at the point where he is useful.

However, there's no reason why somebody like FE7!Marcus or Titania shouldn't show up in a highly complex endgame map.

I generally find the "cuts turns" thinking problematic. One can be necessary to cut turns (e.g. Titania in chapter 1; though I'm having doubts when you say you can only 5-turn the map without her because I recall reading drafts that took a bit less time), they can be one of the options to achieve the LTC (SM Mia in C9 that we've discussed for way longer than it's worth), or they can simply assist (a lot of Paladins and combat units who fight and ORKO enemies) with one's low-turn completion.

[Chiki]

I don't know why you think it's dubious. Is it the fact that I subjectively rate complexity? It'd probably be possible to calculate it objectively, but unfortunately it doesn't seem possible for me to do that alone.

I don't think that, ultimately, calculating complexity objectively is gonna differ much from my rating.

If Mia is only one of two options to cut a turn, then she gets half credit for the turn cut. There are ways to fix the things you find problematic.

Edited April 30, 2013 by Olwen

[Anouleth]

It hasn't been a problem for that tier list when it should. You guys don't really lay out how you tier units, and certainly don't think over it too much. You simply have a vague idea of how to.

I think it would make Edward's contributions seem far more impressive, but remember that even one chapter with high complexity is not as valuable as many chapters with high complexity. Units that contribute to more chapters are more valuable. I think that matches with our intuitions too.

Think of 5 chapters with 1/5 complexity. Think of 1 chapter with 5/5 complexity. I think those 5 chapters are worth as much as that one chapter with 5/5 complexity. So complexity does not concern a chapter alone, but rather many chapters.

The issue is not the complexity. The issue (which should be obvious) is that Edward is forced deployment and one of only two units available to you. Of course, in such an environment, where the player is practically being forced to use Edward, Edward is going to look good, but I think most people would agree that's not the sort of thing that really makes a unit good. Being forced to use a unit does not make that unit better: the fact that Edward saves a lot of turns is an artifact of the storyline putting you into a position where he is your only viable combat unit, rather than him being a strong combat or utility unit relative to other units (which is what most people would define as good). I don't consider many chapters in FE to be meaningful indicators of whether a unit is good or bad, particularly Prologue chapters which often force you to use only one or two units. In my perspective, those chapters might as well be extended semi-interactive cutscenes. I would certainly prefer it in the case of FE10 2-1, which is a nightmarre without transfers.

[Chiki]

We're considering the storyline now? That has nothing to do with anything. You certainly need a far better reason than that. Why don't we tier units based on their contribution to the story too? Besides, by that logic, Kurthnaga is bigger than castles in cutscenes whereas the same size as any other unit in 3-E.

We could imagine a scenario where Micaiah wasn't complete crap and could take on the chapter like Edward. This proves that Edward's superiority to other units is what makes him forced in that chapter, and nowhere else (he quickly gets outclassed by Nolan). Edward isn't forced just because you need someone to protect Micaiah and it certainly is possible to beat the chapter without him; Edward is forced because he is just that amazing in Chapter 1-P and makes it possible to beat quickly.

The purpose of an LTC tier list is to consider complexity-weighted turns and reliability. Does Edward save a ridiculous amount of turns? Yes. I don't see why anything else needs to be considered. It's certainly ad hoc to say Edward should be penalized just because the prologue has a few possible units available. Edward is just better than all of them and his position in a tier list shows how necessary he is in such a tier list.

Compare Edward to Titania. Titania is good in early game chapters with 3 to 4 units. Edward is good in one chapter with 3 units. But Titania is never as essential as Edward is, and her usage certainly isn't as forced as Edward's is. It's just a consequence of the storyline that Titania is in every chapter, am I right? So let's penalize her.

Edited May 1, 2013 by Olwen

[Snowy_One]

I think what he's saying is that the reason Edward is even deployed is because of a story-line requirement, not his actual utility. Imagine if Rolf was required to be deployed every chapter in PoR. Since he HAS to be deployed he WILL contribute to LTC and take up resources and the like unless the player hides him in a corner at which point they need to be careful he doesn't get insta-mauled by some baddie or something similar. That doesn't mean he's actually GOOD in any way, shape, or form, just that the player is trying to make the best of a bad situation.

Course, I could be totally wrong on that and he's saying something different entirely.

[Chiki]

Titania is deployed throughout early game as a storyline requirement. That doesn't mean a thing

He just doesn't want to give points to Edward for some unimportant arbitrary reason.

Edited May 1, 2013 by Olwen

[Redwall]

I think everyone is on the same page wrt Edward being a bad unit; the problem is that the SF tier lists justify this prediction in an ad hoc fashion, whereas the complexity idea, although somewhat flawed for reasons that I and others have brought up, at least allows a self-consistent way of predicting Edward being trash.

[Chiki]

I still see no way of testing your predictive model of complexity (you still have not addressed my question of reproducibility) with some actual measurements of how many calculations are necessary, precisely because we don't have the ability to independently measure the relative worth of various types of calculations at this point (e.g. is a Rescue-drop chain equivalent to five rounds of enemy-phase combat for Frederick?).

Your meaning is not clear. Let me put it this way: if asked to choose between a 30%-reliable four-turn clear and a 60%-reliable five-turn clear for a particular level, with no other stage clears influenced, which would your tier list favor? What I would prefer to see in a tier list would be something like what other users have suggested: restrict the player to 100%-reliable (or, for stages where this is not possible, as-reliable-as-possible) strategies; then, seek the lowest turn count under this constraint.

I forgot to reply to this. It's a good post.

Consider chess. Can we objectively verify the number of calculations that takes place in a game between a good player and a bad one? I'm pretty sure there are ways to objectively measure the number of calculations for both players. I heard a study on it just yesterday.

Then consider Fire Emblem. Can we objectively verify the number of calculations that takes place in a chapter with low complexity and a chapter with high complexity?

The answer is yes, of course we can. But it's certainly very difficult to do so. I don't have the resources to pull it off. But I don't think there's any other objective measurement to prevent Edward from rising to the top of the tier list.

I don't agree with the second paragraph at all. There are no strategies with exactly 100% success rate. Do we prefer the 98% chance to clear in 5 turns or the 99% chance to clear in 6? Your problem still remains.

Edward is certainly not bad but not top tier either. He has his uses after 1-P, and even in 3-6 to kill in transformed sub-humans with Caladbolg.

Edited May 1, 2013 by Olwen

[Hawk King]

Olwen there is a major flaw with your "turns saved" definition/example.

(For FE9) It works perfectly fine for Chapters 1-7 since the deployment spots are forced. And saying that Marcia saves 6 turns in Chapter 12, because without her it's an 8 turn clear and with her it's a 2 turn clear, works. But lets assume that Jill is recruited. If Jill is recruited then you can say that Marcia saves just 1? turn in Chapters 13-Endgame because Jill can replicate everything thats Marcia can do except for getting shoved in 17-2.

Basically, every unit loses their claim to any contributions as long as there is another unit who is capable of making the same contribution.

So by using this logic, every dancer/ Heron should be at the top of their respective teir list because their actions cannot be replicated by any other unit.

And if you throw the "Complexity" thing in as well FE9 Titania and Marcia would fall significantly because all of their highest "turns saved" Chapters are pretty low on the complexity scale.

[Chiki]

I don't see the big issue. Marcia saves around 10 more turns than Jill does, thanks to being able to double bosses and being able to be shoved.

Due to this, and Marcia not costing as many turns to recruit, Marcia is a lot more likely to be used than Jill in my tier list. Hence her being above Jill.

Edited May 1, 2013 by Olwen

[Anouleth]

We're considering the storyline now? That has nothing to do with anything. You certainly need a far better reason than that. Why don't we tier units based on their contribution to the story too? Besides, by that logic, Kurthnaga is bigger than castles in cutscenes whereas the same size as any other unit in 3-E.

We could imagine a scenario where Micaiah wasn't complete crap and could take on the chapter like Edward. This proves that Edward's superiority to other units is what makes him forced in that chapter, and nowhere else (he quickly gets outclassed by Nolan). Edward isn't forced just because you need someone to protect Micaiah and it certainly is possible to beat the chapter without him; Edward is forced because he is just that amazing in Chapter 1-P and makes it possible to beat quickly.

The purpose of an LTC tier list is to consider complexity-weighted turns and reliability. Does Edward save a ridiculous amount of turns? Yes. I don't see why anything else needs to be considered. It's certainly ad hoc to say Edward should be penalized just because the prologue has a few possible units available. Edward is just better than all of them and his position in a tier list shows how necessary he is in such a tier list.

Compare Edward to Titania. Titania is good in early game chapters with 3 to 4 units. Edward is good in one chapter with 3 units. But Titania is never as essential as Edward is, and her usage certainly isn't as forced as Edward's is. It's just a consequence of the storyline that Titania is in every chapter, am I right? So let's penalize her.

I'm not considering the storyline. My whole point was that certain chapters, like Prologue of Part 1 of Radiant Dawn, are like cutscenes and should be ignored entirely. I don't get where you got that I wanted to consider characters based on their cutscene performance.

The fact that Edward is "better than Micaiah" is not particularly impressive. Being stronger than someone who is considered possibly the second or third weakest combat unit in the game, out of 72 units, in one chapter is not what I'd consider being a "good unit".

I think 4 units is the point where I can take a chapter seriously. So I think that FE13 Prologue, or the starting chapters of FE11, or FE6 are relevant to tier lists. But FE8 Prologue, or FE7 Prologue, or FE9 Prologue, I don't care about at all. It's just that usually, those chapters don't provide excuses to hype up otherwise shit units.

Titania is deployed throughout early game as a storyline requirement. That doesn't mean a thing

He just doesn't want to give points to Edward for some unimportant arbitrary reason.

This has nothing to do with Edward (who is actually one of my favourite characters, who I have argued many times in favour of in the FE10 tier list). I feel the same way about other units in very similar situations, like Brom, Ike, Lyn, Marth... well, there are probably others I have forgotten.

[Chiki]

Saying "I think 4 units is a good point" isn't rational at all. I see no logically necessary reason to think 4 units is a good place to start. It seems like you think so just because you feel so, but clearly that doesn't cut it.

[Anouleth]

Saying "I think 4 units is a good point" isn't rational at all. I see no logically necessary reason to think 4 units is a good place to start. It seems like you think so just because you feel so, but clearly that doesn't cut it.

Well, what's the "logically necessary reason" to use turncounts? Or to disregard contributions like recruitment or seizing? It seems like we don't have any logically necessary reason to do any of these things. Indeed, none of the criteria we use in any tier list is "logically necessary". We use those criteria because we like using them, or because we feel that a tier list with those criteria better captures what makes a unit "good" or "bad".

Edited May 1, 2013 by Anouleth

[Chiki]

Well, what's the "logically necessary reason" to use turncounts? Or to disregard contributions like recruitment or seizing? It seems like we don't have any logically necessary reason to do any of these things. Indeed, none of the criteria we use in any tier list is "logically necessary". We use those criteria because we like using them, or because we feel that a tier list with those criteria better captures what makes a unit "good" or "bad".

A tier list is by definition going to have arbitrary criteria, for a tier list is itself an arbitrary ranking of characters based on criteria which we subjectively prefer.

But that doesn't mean we should add more and more meaningless arbitrary criteria. It certainly isn't a good thing to use more than we need. I've been advocating a minimalist approach for a while now.

Consider drinking wine. Surely it isn't bad for the first 3 glasses, but if we keep adding glasses of wine over and over again, at one point it's going to get bad.

[Kitoari]

A tier list is by definition going to have arbitrary criteria, for a tier list is itself an arbitrary ranking of characters based on criteria which we subjectively prefer.

But that doesn't mean we should add more and more meaningless arbitrary criteria. It certainly isn't a good thing to use more than we need. I've been advocating a minimalist approach for a while now.

Consider drinking wine. Surely it isn't bad for the first 3 glasses, but if we keep adding glasses of wine over and over again, at one point it's going to get bad.

Drinking has real world consequences though. Subjectively valuing certain things more than other things doesn't.

[Chiki]

Drinking has real world consequences though. Subjectively valuing certain things more than other things doesn't.

It doesn't have to have real world consequences if you do it in private.

The analogy remains. Just because we're being 10% irrational doesn't mean we should be 100% irrational.

[Redwall]

Drinking has real world consequences though. Subjectively valuing certain things more than other things doesn't.

The real-world consequences predicted by a tier list structured around complexity-weighted turn counts include, but are not limited to, the reason you, Anouleth, and Olwen place little value on Edward's 1-P contributions: it's just not an interesting chapter, as Anouleth puts it. I agree with him on that. The problem is that the tier lists, as they are presently constructed, don't describe which chapters are interesting and which are not.

In a tier list structured around turn counts alone, rigorously applying the criteria would mean placing Edward in a much higher tier than the SF tier list would. The conventional argument against high-tier Edward is that all your other units in 1-P suck worse, and that he becomes, if not bad, then certainly unremarkable after a certain point. However, this line of thought and the devaluing of his contributions is not consistent with valuing turn counts.

Interceptor would counter this by saying "OK, but turn counts are not the point; the playstyle is the point." But working in this line of thought would still necessitate putting Edward in a high tier, as there is no stated reason in the tier list rules to value saving 50 turns in 1-P over, say, saving two turns across each of the next 25 levels.

While we can agree that Edward should not be high tier, the SF tier list's framework offers no explanation for why this is. The explanation falls outside the bounds set by the tier list's rules.

Edited May 1, 2013 by Redwall

[Kitoari]

I actually didn't comment on Eddie if I remember correctly. I was around when Thany was the unit of the day, aka the double-counting fiasco... which makes it seem like this topic is going nowhere, but I digress.

As for the Eddie debacle, every framework presented this far isn't satisfying me. Something just seems off, so I'll refrain from commenting.

Edit: Wait, what's the TC if Eddie is used but not Micaiah or Leo?

Edited May 1, 2013 by Kitoari

[PKL]

I think everyone is on the same page wrt Edward being a bad unit; the problem is that the SF tier lists justify this prediction in an ad hoc fashion, whereas the complexity idea, although somewhat flawed for reasons that I and others have brought up, at least allows a self-consistent way of predicting Edward being trash.

Speak for yourself. Not everyone thinks Edward is a bad unit. Hint: he really isn't. HM does hit him though.

[Redwall]

Speak for yourself. Not everyone thinks Edward is a bad unit. Hint: he really isn't. HM does hit him though.

Missing the point completely.

[Chiki]

I actually didn't comment on Eddie if I remember correctly. I was around when Thany was the unit of the day, aka the double-counting fiasco... which makes it seem like this topic is going nowhere, but I digress.

As for the Eddie debacle, every framework presented this far isn't satisfying me. Something just seems off, so I'll refrain from commenting.

Edit: Wait, what's the TC if Eddie is used but not Micaiah or Leo?

Double-counting "fiasco?" Eddie "debacle?" Care to explain?

Edited May 1, 2013 by Olwen

[Kitoari]

Well, my use of somewhat dated and whimsical terms for arguments should indicate I'm not taking this terribly seriously. I'm not sure what there is to explain there, I meant what I said.

it's also just that tiering philosophy is SRS which is kind of silly in my opinion

[Anouleth]

Speak for yourself. Not everyone thinks Edward is a bad unit. Hint: he really isn't. HM does hit him though.

I haven't really thought about Edward in FE10 in a long time. When I called him "shit", I was overstating it: but Edward is definitely a mediocrity, he definitely has serious issues, and I can see why people might think of him as a bad unit. I don't know if I'd put Edward in Mid tier, as the current list does; I might put him at the top of Lower Mid