How I Envision Efficient Fire Emblem Discussion

Hello YouTube, jumpy23 here with the third part to my "efficient Fire Emblem" series that I wasn't sure I would make. Well, I've decided to. This is a third part, so I would recommend starting from my first video, "A Case for Strictly Defining and Measuring Efficient Fire Emblem Play," which I'll link in the description. And as always, if you'd rather read the script than watch, I'll link that in the description as well. Fair warning, my Ironman play is going to be even worse than usual in this video.

Before I hop in, I also want to shout out ActualLizard, who has just started a series playing through FE3 book 1 with ETC as a benchmark. He's uploaded through Chapter 7 so far and I'll link the first video in the description. It's a great example of applying the theory I've been ranting about in practice, and quite entertaining.

---

The biggest pushback that I received to adopting ETC was "Fuck off, I don't wanna do math every time I discuss Fire Emblem." I feel you there, I don't either. However, I think that we get 90% of the benefit just by agreeing that ETC *could* be used to calculate efficiency, even if nobody ever actually calculates the ETC of anything. Allow me to lead in with an extended analogy.

ETC is a measurement, like weight. We all agree on the units of measurement for weight and how to measure the weight of something if we so wished (yes, Americans have our own units, but we all agree on the conversion factors, so shush). But how often do we actually need to weigh things in order to make decisions about weights? Given two things, we can generally tell which of them weighs more, unless it's close. If we want to reduce the weight of something, we know how to do it. I can even pick something up and probably tell you its weight to within ~10% of the actual value. Yes, sometimes you have to whip out the scale, but usually intuition of what the scale would tell you is all you need.

Similarly, having an intuition of ETC is usually all you need. Valkama pointed out the most important heuristic in the comments on my last video: If you have a reliable clear, ETC is roughly equal to turncount. And ETC punishes resets very heavily; even a 50% reset on turn 1, or a 20% reset on turn 4, costs an entire turn. If you want to get a low ETC for a map, "as quick as you can with a single-digit chance of failure" is usually the right answer. And chance of failure is the easiest possible reliability calculation, so it doesn't get much simpler than that.

When we want to improve ETC, it's easy to grasp the emergent properties of the formula without actually needing to run the numbers.
 * Use accurate weapons, especially on rout maps
 * Don't rely on crits or dodges
 * Prefer extensions to resets
 * Only take risks on turn 1
 * And, yes, reduce turn count
If you have a clear that does all of these, then it's gonna be quite efficient, and you can probably approximate ETC as turncount or turncount + 1. Additionally, it's going to be (in my opinion) a fun way to clear the game, because ETC pushes you to make as many interesting tactical maneuvers as you can without resetting.

Another aspect of the "I don't want to do math" argument seems to be the fear that rough ideas will be discarded entirely. In other words, that you can't spitball how a unit might work, or talk about how you've played a map in the past, unless you bring the numbers and math to back it up. I am absolutely not trying to wipe out this kind of discussion, which I do all the time myself. However, a basic understanding of the principles of efficient play as outlined in the previous paragraph gives us a better framework to evaluate these kinds of arguments.

But when someone feels really strongly about something and really wants to prove that they are *right*, they have the ability to do so by busting out the abacus. Under the current loosey-goosey definition of efficiency where clout and rhetoric are key, it's very difficult to change the status quo opinion except by being a big name YouTuber. Most of the time I personally bother to sit down and do real ETC math is when I'm attempting to do exactly that, but I'm very glad I have this option when it comes up.

---

The other question that I got asked over and over on my last two videos was how does all of this relate to unit tier lists. To me, there's a lot more to efficient Fire Emblem discussion than just tier lists, but I understand why I kept getting that question: for a lot of people, the only time they hear the "efficient" buzzword is when discussing unit tiering, since tiering is such a big part of that semi-casual semi-hardcore Fire Emblem space that, say, Mekkah viewers occupy.

The assumption that tends to come up in these posts is that in an ETC-informed tier list, units are only judged on a single context, usually the most efficient known playthrough. In other words, there are two approaches to unit tier lists: one which sorts units by their contributions to some canonized route, and another that ranks units on how they could *potentially* help you. Sometimes LTCers do make lists of the former type after completing an LTC for fun, and people are concerned that I envision the same thing applying to low-ETC runs. I guess the assumption here is something like, "While we didn't really have a metric for efficiency and everything could basically be argued to be equal, you had to take all these different contexts into account, but if you can prove that Context A is better than Context B, you can completely disregard B." I've seen some other assumptions like that an ETC tier list could not ignore recruitment cost.

For example, in the most efficient known contexts, Lute does literally nothing in FE8. There are only four units for whom this is true (the others are the usual culprits -- Marisa, Amelia, Ewan). So if we adopt the 'canonized route' methodology for an ETC-informed tier list, Lute would be a bottom 4 unit in FE8.

This is definitely not how I would run a unit tier list. The core purpose of unit tier lists is, in my opinion, to facilitate discussion of units. Specifically I think there are three questions to answer: "What can this unit do," "How good is this unit at doing this," and "How central is this to beating the game?" Narrowing contexts too much fails to answer these questions in a satisfactory way because you are considering only a small subset of answers to the first question, "what can this unit do?"

The way I phrased this in unit tier lists I have run is as follows: "We consider all contexts in which a unit can reasonably be used, weighted by efficiency. A minor contribution that reliably costs no turns is valuable. A major contribution that reliably costs only a few turns is valuable. A minor contribution that reliably costs only a few turns, or a major contribution that reliably costs a handful of turns, can be somewhat valuable. Slower strategies need proportionately greater impact / reliability to be worth heavily weighting." Yes, this ends up being subjective, but that's good enough for something as silly (fun, but still silly) as a unit tier list.

I think this system produces fairly good results in practice. Returning to the Lute example, when she *is* used, she has some unique upsides and replicates some upside of other fairly decent units, at the price of a slightly costlier training arc. The efficiency Discord's tier list for FE8 has Lute well far away from the bottom, somewhere around other mid filler units like Garcia, and some units with unique but few contributions like Colm, which feels about right to me.

An interesting positive application of ETC to tier lists comes in the realm of "unique contributions." Unique contributions, things that can only be done by a specific unit, are generally given more weight in a tier list. With a vague understanding of efficiency it can be very difficult to actually have unique contributions, but ETC can crystallize being the 'best' at something into something meaningful. To give a real example from a few days ago: the efficiency discord was debating Panette's tiering and discussing her wrath/vantage performance on rout maps. Technically this can be replicated by Kagetsu, with 10 less crit, which would make it not a unique contribution. However, doing some quick napkin math, we saw that Kagetsu has a significant chance of death on a late turn thanks to his non-100% crit chance, which meant that Panette saved multiple ETC turns and this *was* a unique contribution.

And just to illustrate the ease of using ETC intuition, I'll demonstrate the napkin math we did. We spitballed 10 do-or-die 90% crits over the course of a 6 turn clear. .9^10 is .35 chance of success, so that's a 65% chance of failure. Amortize that to turn 4, and you get 7.4 turns of reset penalty with Kagetsu. It took almost exactly as long to say this paragraph as it took to do the math during the conversation. No this isn't the exact number but it's clearly high enough to prove that Panette is uniquely good at vantage/wrath setups. This example is for high tier units, but this logic applies for lower tier units as well and I could give several examples: Joshua can double Caellach reliably; Sheema is your most accurate Parthia user; Benny's Fierce Mien increases hit rates.

---

Compared to my last two videos, I'm not sure there's much of a conclusion to be had here. This was more of an exploration of the ways my proposed definition of efficiency can affect Fire Emblem discussion. But I hope some of them sounded like interesting applications to you. Thanks for watching.