Tiering Philosophy - It's that time again

[Chiki]

I'll give a real life example on picking arbitrary criteria. In mathematical logic, we pick certain axioms in order to prove things like 1+1=2, 2x2=4, etc. Of course, the reasoning for those axioms is ultimately arbitrary. But without those we wouldn't be to prove our mathematical operations at all. But the fact that those axioms are ultimately arbitrary doesn't mean that we can pick an infinite number of arbitrary axioms. That is why we use the least number of arbitrary axioms possible.

Apply the same reasoning to Fire Emblem.

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

We take tier listing seriously for the same reason we play video games, and other people watch sports: to pass time. No one gives a crap if you're gonna brag about how cool you are for not taking this seriously. Remember that your spending time here makes you just as serious as we are.

[Narga_Rocks]

We take tier listing seriously for the same reason we play video games, and other people watch sports: to pass time. No one gives a crap if you're gonna brag about how cool you are for not taking this seriously. Remember that your spending time here makes you just as serious as we are.

Actually it doesn't make him serious about it. It just makes his time here a waste whereas our time here is spent on something we are passionate about, hence not a waste.

[Chiki]

Actually it doesn't make him serious about it. It just makes his time here a waste whereas our time here is spent on something we are passionate about, hence not a waste.

I meant serious in the sense that he still seems to want to waste time on an online forum just like we do.

Anyway, this is going off topic. Anyone here want me to try making an FE10 tier list with the 3 criteria? I might do an FE13 one if the FE10 one is well received, as well.

Edited May 1, 2013 by Olwen

[Quintessence]

I meant serious in the sense that he still seems to want to waste time on an online forum just like we do.

Anyway, this is going off topic. Anyone here want me to try making an FE10 tier list with the 3 criteria? I might do an FE13 one if the FE10 one is well received, as well.

Sure, it'd be nice to apply what you have discussed for so many pages.

[Chiki]

Sure, it'd be nice to apply what you have discussed for so many pages.

I've posted an FE9 tier list here: http://serenesforest.net/forums/index.php?showtopic=39246

An FE10 tier list seems a lot harder to make (complex, ha ha lame joke) than FE9. Though I think I'll be putting Jill with transfers above Haar in top tier.

[-Cynthia-]

I don't see how your definition of complexity is "objective". For example, it takes me more calculations to do 1-9 correctly (you must move the BK and Micaiah very precisely to clear the chapter at all) but chapters like...4-P can be quickly and easily completed by putting Haar/Jill with forged hand axes in the general vicinity of the enemies. 1-9 is 1/5 supposedly, while 4-P is 4/5.

Seems to me like complexity is just another arbitrary criterion- especially how something like the mental calculations of each player is wildly inconsistent.

[Kitoari]

We take tier listing seriously for the same reason we play video games, and other people watch sports: to pass time.

to extend the sports analogy, I'm truly amazed when people actually get angry and occasionally violent over the results of their teams

I see no reason to find similar behavior here acceptable.

[Chiki]

I don't see how your definition of complexity is "objective". For example, it takes me more calculations to do 1-9 correctly (you must move the BK and Micaiah very precisely to clear the chapter at all) but chapters like...4-P can be quickly and easily completed by putting Haar/Jill with forged hand axes in the general vicinity of the enemies. 1-9 is 1/5 supposedly, while 4-P is 4/5.

Seems to me like complexity is just another arbitrary criterion- especially how something like the mental calculations of each player is wildly inconsistent.

We assume that there is a perfect player who does these calculations. I've said many times that I assume perfect play.

And it is not so easy to complete 4-P. It takes perfect positioning and such to clear it.

[-Cynthia-]

We assume that there is a perfect player who does these calculations. I've said many times that I assume perfect play.

And it is not so easy to complete 4-P. It takes perfect positioning and such to clear it.

This doesn't address my point at all. 1-9 also requires perfect positioning to clear it and generally takes as many 'moves' as some of the chapters you've listed as more 'complex'.

Still seems like an extremely arbitrary construct to me which seems entirely dependent on whether you think a chapter is complex or not.

[Chiki]

This doesn't address my point at all. 1-9 also requires perfect positioning to clear it and generally takes as many 'moves' as some of the chapters you've listed as more 'complex'.

Still seems like an extremely arbitrary construct to me which seems entirely dependent on whether you think a chapter is complex or not.

First off, I'll probably update 1-9 as 3/5.

The second line just makes me sigh. I don't know how many times I have to explain this.

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.

[-Cynthia-]

One way to stop Edward from reaching the top of the tier list objectively is to say that saving turns over multiple chapters matters less than saving turns over a number of chapters. Titania is never as useful in one single chapter as Edward is in 1-P, but she's also useful for ~10-12 chapters where he has mostly minor contributions outside of a chapter or two.

There's also probably a better way to define complexity than 'number of calculations' if you can't actually measure the number of calculations. Number of enemies, positioning of enemies, map objective(s). But, part of chapter complexity may be the units we have available (if we don't have any staff users FE11 becomes more complex for instance). This creates a difficulty when tiering units obviously- does Lena make the chapter less complex or is the chapter not complex because of Lena?

[Chiki]

One way to stop Edward from reaching the top of the tier list objectively is to say that saving turns over multiple chapters matters less than saving turns over a number of chapters. Titania is never as useful in one single chapter as Edward is in 1-P, but she's also useful for ~10-12 chapters where he has mostly minor contributions outside of a chapter or two.

There's also probably a better way to define complexity than 'number of calculations' if you can't actually measure the number of calculations. Number of enemies, positioning of enemies, map objective(s). But, part of chapter complexity may be the units we have available (if we don't have any staff users FE11 becomes more complex for instance). This creates a difficulty when tiering units obviously- does Lena make the chapter less complex or is the chapter not complex because of Lena?

Why does it matter more? If all chapters in a game were equally complex, I don't see why 5 turns in one is more important than saving 1 turn in 5. The burden of proof is on you: you have to demonstrate why it's more important instead of simply saying it is.

We can measure the number of calculations, and there is another way to do so. We can take a large sample size and find the average number of calculations for each and every chapter. That's a perfectly objective way to measure complexity.

[-Cynthia-]

Why does it matter more? If all chapters in a game were equally complex, I don't see why 5 turns in one is more important than saving 1 turn in 5. The burden of proof is on you: you have to demonstrate why it's more important instead of simply saying it is.

We can measure the number of calculations, and there is another way to do so. We can take a large sample size and find the average number of calculations for each and every chapter. That's a perfectly objective way to measure complexity.

And the burden on proof is on you to determine why it is more important to save more turns in 'complex' chapters than 'simple' ones.

You don't actually have a large sample size about the average number of calculations for each chapter, but you're still assigning complexity scores for each chapter.

[Chiki]

And the burden on proof is on you to determine why it is more important to save more turns in 'complex' chapters than 'simple' ones.

You don't actually have a large sample size about the average number of calculations for each chapter, but you're still assigning complexity scores for each chapter.

I think it's less arbitrary than to claim "well, it's more important to save turns in many chapters as opposed to save a crap load of turns in one chapter, just because. I won't explain why, it's just because." Ultimately, I won't deny that the reason both are important is arbitrary, but it's more rational to use an objective, well thought out term for tier lists as opposed to saying "just because." We should ultimately prefer something that's objective and falsifiable as opposed to something that's subjective.

You guys will have to pick which one you prefer. The subjective one vs. the objective option.

....The point is that we can get a large sample size and measure the number of calculations if I had the resources to conduct tests. Therefore it is objective.

Edited May 2, 2013 by Olwen

[Redwall]

Have you considered simply including a surveymonkey poll in your tier lists as a way of acquiring others' stage-complexity ratings, and then editing the displayed complexity ratings in your tier list to match the averages from the survey? While obviously not as good as in vivo measurements of brain activity, it seems necessary to make the rating system work objectively.

[Chiki]

It's a good idea. I doubt anyone would be interested though.

[Snowy_One]

Maybe this is just me being dumb but... The point of counting turns is that they're a measurable proxy to unit quality, right? Well, maybe this is one of the situations where a unit of low quality can save a lot of turns, but isn't a quality unit? Seems to me to be straight-forwards. Edward has a situation where he can be unusually useful, but outside of that situation, he's pretty meh. As a result, he's not ranked highly as his actual quality isn't that good despite him being able to shave off a lot of turns in one instance.

Who am I kidding though? You've probably ignored me a while ago it seems, if only to focus on this debate.

[Chiki]

Why is one instance in which Edward saves 5 turns less important than 5 instances in which Jill saves 1 turn?

[Snowy_One]

Because, least as I understand it, turns saved is merely a measure of unit quality, not the definition of it. Jill may not save as many turns (least for the sake of argument), but she's still a solid unit of high quality. Edward is an meh unit of meh quality who, in one instance can save a lot of turns. If turns = quality, you'd be right (least in this theoretical setup) but turns =/= quality. Instead turns ~ quality and the tier list is meant to measure the latter. If PoR Sothe can unlock one door that saves 5 turns, is he suddenly a higher-quality unit than PoR Zihark who doesn't save any? Least in this hypothetical question he doesn't save any. I don't think anyone would say so short of feeling that the only point to a list was turn reduction.

[Chiki]

Here is your argument. Turns saved entails unit quality.

If Edward saves 5 turns in one instance, it entails that he is a unit of extremely high quality in that instance. Let us say that he is rated 5/5 for unit quality in this instance. We then take the average of unit quality for the total of every instance, which is 1/5.

If Jill saves 1 turn in 5 instances, it entails that she is a unit of okay quality in these instances. Let us say that she is rated 1/5 for unit quality in these instances. The total average of her unit quality for the total of these instances, then, is 1/5, as it is the same for every 5 chapters.

It is totally irrelevant, then, if turns saved entails unit quality.

[Snowy_One]

That's not my argument. My argument is that unit quality is something measured most objectively by turns saved. However, there are other factors that can change the outcome of the turns saved outcome. (Edit: Note, I said 'objective', not 'accurate' or 'best', so I don't want to hear that turns saved = quality). Edward happens to benefit from one of these qualities that allows him to net a lot of turns saved without actual quality. Since the tier-list measures unit quality and not turns saved, Edward's turn-saving doesn't count.

Is Edward's turn-saving a consistent thing? Does he continue to provide these reductions over the various chapters? Or is it a one-time thing that he just happens to be in the right position for? An errant as it were?

Edited May 2, 2013 by Snowy_One

[Chiki]

That's not my argument. My argument is that unit quality is something measured most objectively by turns saved. However, there are other factors that can change the outcome of the turns saved outcome. (Edit: Note, I said 'objective', not 'accurate' or 'best', so I don't want to hear that turns saved = quality). Edward happens to benefit from one of these qualities that allows him to net a lot of turns saved without actual quality. Since the tier-list measures unit quality and not turns saved, Edward's turn-saving doesn't count.

Is Edward's turn-saving a consistent thing? Does he continue to provide these reductions over the various chapters? Or is it a one-time thing that he just happens to be in the right position for? An errant as it were?

I have no idea what you're saying. I'm trying to find an objective criteria in order to prevent Edward's turnsaving from launching him up into the top of top tier. Are you agreeing or disagreeing with me?

[Snowy_One]

I don't think I'm doing either really. Just objecting to the notion that he should rise up, especially since Edward's contributions aren't really consistent across multiple chapters/teams.

[dondon151]

I have no idea what you mean by the first two lines, but I will respond to the third. Why on earth would it be better "if we keep the definition shenanigans to a minimum?" There is nothing more important than defining our concepts in a non-ad hoc way in arguments. Why would we rely on our intuitions to vaguely define terms for us? There is no rational debater alive in the world who would claim that definitions aren't important, and that we shouldn't argue over them. If we can't establish solid, agreed-upon definitions, how on earth can we have a rational debate?

1) Everyone will know exactly what they mean once I define them--and once we list those definitions out after agreeing on them.

2) They won't use them incorrectly if we define them properly.

3) People argue over everything unnecessarily, but in the end, a democracy wins once we can agree upon some definitions.

confusion over definitions is inevitable precisely because the nature of fire emblem tiering is subjective. plastering an objective facade over what is a subjective business isn't really going to get you anywhere. certainly, there are some facts that are more objective, so to speak, than others. one is that everything is a means to an end, and that end is turncounts. another is that units must be used in order to be tiered. these rules have been around for a long time and are easy to grasp.

there is no way to fix the problem of people disagreeing with your (and i mean, specifically your) method of tiering. it's actually almost amusing that you cannot fathom why others would disagree with your definitions even after they've been agreed upon. it's precisely because the definitions that you're using are still not objective! how does the expression go again? putting lipstick on a pig?

or, this is like what happens when a bunch of people try to create a standard all at once - you end up with multiple "standards." some guy is just going to create his own version of a tier list that uses slightly different definitions than yours does. now, i guess you could say that this isn't a problem, but then you carry on as if your tier list were the only one that matters, so this stance is obviously self-contradictory.

The issues with the definitions you guys use is that none of them are well-defined. And once you guys do define them, you seem to add various ad hoc adjustments such as stuff about direct reliance and indirect reliance without even being able to articulate your view coherently. They aren't logically necessary things to consider at all, so I don't see why we should even waste time arguing about them.

Here are the non-ad hoc, very simple definitions for the terms I use.

Complexity: the number of calculations to complete a chapter efficiently. I believe this is logically necessary, to prevent contradictions like Edward from rising to the top of the FE10 tier list. Furthermore, I believe that this is preferable over a subjective notion like "difficulty."

Turns saved: A unit saves turns if and only if the highest possible turn count for a chapter without them is higher.

Reliability: The chance that a unit will do actions with reliability. This is measured by percentages.

i don't see anything wrong with using ill-defined definitions. tiering is innately an imprecise science, as i have repeatedly stated. any random dude can play through a fire emblem game and immediately form his own impression on which units are good and which are bad. you could just as well judge units by "feel" and it would be just as legitimate (well, not really, because of the RNG, but consider a hypothetical situation where stat growths are not subject to randomness).

so you might think that your definitions leave no room for error. on the other hand, i see a plethora of problems. here's a non-exhaustive list:

1. you don't define what a "calculation" is.

2. you don't define what qualifies as "efficiently." do you care about only one possible outcome?

3. if you care about multiple possible outcomes, how do you weight the outcomes?

4. why is complexity measured on a discrete scale instead of a continuous scale?

5. what happens if multiple units possess the ability to "save" a turn?

6. what happens if multiple combinations of multiple units possess the ability to "save" a turn?

7. do you even consider scenarios where a unit saves turns only if the unit who saves the most turns isn't used? why or why not?

8. even reliability can't always be accurately calculated because of the sheer quantity of possible outcomes. how do you account for that?

Note that I am not adding anything logically unnecessary with crap like "reliance on other units" or "double-counting." I am taking a minimalist approach, whereby I only consider the terms which I consider to be logically necessary and nothing else. These are the only things which I would like to use when comparing units to each other in tier lists. So I don't take points from FE9 Mia for taking bexp if there's no opportunity cost, as the effect on reliability is trivial, whether it be direct or indirect. There is also nothing about "uniqueness," which I think is inherently contradictory. If we had a million Marcias in FE9, I see no reason for them to go into bottom tier, or for them to go into low tier and the rest of the units in bottom tier. For it would be silly to put them in low tier when the very term has negative connotations, while there is nothing intuitively negative about having a million Marcias. I do not find it logically necessary to penalize units based on uniqueness--only award if necessary (such as the case of Volke being one of the only two thiefs in FE9).

this "unnecessary crap" forms the basis of simple macro- and microeconomic principles. you may consider yourself minimalist, but i'm the one who's not making things up.

But if it were the case that Marcia could not promote immediately, then we would have issues in Chapter 12, where Marcia saves 6 turns alone. She would not be able to kill the boss, and that could cost us a turn or two--I'm not sure. So if Marcia misses out on this bonus experience, the other units and your strategies have disappointed her--because Marcia relies on your other units and your strategies to get this bonus experience and shave one or two turns. You could make the counterargument that Thany relies on Rutger alone, and without him she is useless and has no way to get back up to par without him, unlike Marcia, who can promote with a couple of chapters of bexp. But remember that Rutger is replacable, and he is not the only unit you get in the game. I believe even mounted units like Lance can be rescued (I don't know if Thany can rescue Lance, though) so there are other units which can replace Rutger. Dieck could be one of them. I'm no expert on FE6 though, so feel free to correct me.

espinosa 5-turned chapter 12 with a shitty marcia. i really don't believe that the difference between a 20/-- marcia and a 20/1 marcia is 6 turns.

[Snowy_One]

This whole debate has caused me to look at one particular aspect and ask myself if tier lists should factor in consistency into their tier-placement. Ergo, if a unit can provide the same service, or lack of, on a constant basis with no/few spikes or valleys in their performance. Basically, we look merely at the rough mean-average of how many turns a unit shaves off as opposed to unusual spikes. Assume two units who join at the same time. Despite being statistically different both shave off 2 turns on average per chapter. Unit A, the statistically stronger one, shaves these turns off every chapter. Unit B, the weaker, shaves off an average of 1 turn per chapter though, but on one chapter can shave off 11 turns. Their average turns-saved is the same as a result, but their mean turns saved is vastly different. Eliminating notable spikes and valleys could correct that.