Redwall

Avoid and Def are numbers with increasing returns; in Panne's case, since her Avo is so high, in general giving her someone like Lon'qu increases the number of enemy attack attempts (this includes misses) she can survive moreso than giving her someone like Frederick will.

Yeah, I guess my biggest mistake the last time I played through L+ was that I didn't feed enough kills to Avatar; she was like Lv 13 after having completed Chapter 4 and Paralogue 1, and she was also RNG-screwed in everything besides Spd (notably, she only gained two points of Def from base). Seeing you all talk about L+ makes me want to give it another shot -- I'm thinking of allowing Renown awards up to however many Renown points correspond to a completed standard Lunatic playthrough, while banning the use of non-deterministic, high-variance things like Anna shops and event-tile weapons.

wait, what's the magic secret for Chapter 5? If I had to try it again, I'd probably send Fred and Sumia up to the middle deck immediately and take out the Dark Mage, and send everyone else up the right, but the problem with this is that Pavise+ and/or Aegis+ can slow down the rest of the team while they try to take out the Myrm and Barbarian. Also Counter can totally mess up Fred.

the only character you're forced to use in Lunatic is Frederick, and you can phase him out of the lead support slot pretty quickly if you've got a +HP or +Def Avatar. You don't need Avatar, Nowi, or Panne (ostensibly the other "best units" in a leisurely playthrough), as another user (XeKr) on the boards has confirmed. From what I understand, you can even beat Grima without using the Falchion on him, though it requires some ridiculous luck.

Who's Miriel's hubby?

Yeah, hitting the unpromoted caps is something that is likely to happen if you're using the Seal to demote a unit; on the other hand, EXP gain will in general be sped up. If you're playing a no-grind run, you pretty much never want to demote a unit (unless it's to access Manakete or something); if you're allowing grinding, EXPonential growth should allow you to get EXP at a good enough pace for promoted units with numbers in the neighborhood of unpromoted caps

It has the effect of punishing those who grind like one or two characters, yeah. But if you're grinding, you might as well do it right: by training up a whole team, one with enough options to take out a lot of Counter enemies during the player phase. For this reason, I feel like Lunatic+ was designed partly with increasing DLC revenue in mind: when faced with a zillion Counter mooks, it's awfully tempting to just hit up the Outrealm Gate. Although I realize that the game is beatable without DLC, free EXP certainly helps your success rate.

you need a lot of Rescue users (at least three, possibly four), a Dancer, and a lot of firepower. I had Cordelia and Fred pass around a max-forged Beast Killer to take care of the LHS (Rescue Cordelia on turn one so that she kills one Falcon Knight; on turn two she and Fred finish up). Avatar and Lucina took care of the right in a similar fashion. Panne took out two Griffons during each player phase (Dance, Rescue), and I think Chrom handled the last one. I couldn't do the two-turn clear on my LTC attempt since I didn't have enough firepower that time. I might have had enough firepower had I been using Morgan as an actual combat unit (he was my Rallybot), but clearing that chapter in two turns (as opposed to six or seven, which is what I wound up doing) would have prevented me from getting Galeforce in time for Chapter 20.

there are literally two characters who are "mandatory" in vanilla Lunatic: Frederick and a finalboss-slayer. Obviously you'll need other guys to carry you through the main game, but their contributions are not unique; for example, you can beat it without Avatar (as someone else on the boards has done).

The lowest Lunatic+ turncount would probably only be a few turns about the lowest Lunatic turncount (for regular Lunatic, it's probably around 100 with Paralogues 1-4 and 17; my best is 128) since in principle you could reset until everybody had the tamest skills possible (which would of course take way too much time to bother doing). Bosses like Raimi might be problematic, though, since if she has either Pavise or Counter, you won't be able to rush her with Silver Lance like you would in a Lunatic LTC.

I like Morgan, so I'd leave the name unchanged. Honorable mention to Terry, for both genders. It's 2013; let's stop pretending "nerd" is still a bad word.

best: Lucina worst: Nowi

Unless you have some fundamental objection to Renown, go for the awards if you think it will make the playthrough more fun. Ch 8 has the first non-Renown Second Seal.

promote Avatar; the GM class learns a super-good skill at Lv 15. you need not worry about gimping EXP growth with Avatar since Veteran is so good. the Second Seal should be used on Panne to reclass to Wyvern; if you haven't been using Sully, Panne is going to be your number one offensive threat not named Avatar. Chrom is just kind of a disappointing unit; he's good, but there's a reason he's at the bottom of S tier.

The Awakening tier list doesn't assume LTC play; children notwithstanding, there aren't any Galeforce-using units in the S tier, and only one Galeforcer above it (Avatar, naturally). There is also a tacit agreement that giant Rescue chains (central to LTC) are not used. Nonetheless, the player is assumed to play quickly. If giant Rescue chains are not assumed, the difference in character rankings between "quickly", "even more quickly", and "quickest" is probably insignificant. wrt a pairing tier list...I'm not sure what the steward (SDS) has in mind. Currently, children aside from Lucina are not tiered.

This is related to why I'm opposed to incorporating the sorts of additions you and Olwen are proposing: without confirming an increased amount of explanatory power, making any model more complex is generally not productive. This is not to say that your tier lists do not provide any added explanatory power--only that the explanatory power cannot presently be tested against any independent and objective measurements. Your proposed criteria add subjectivity and complexity to the existing tier list model, but because your (Snowy's) criteria aren't defined in a reproducible way--by which I mean I could look at two units and come up with utility scores different from yours--and because you haven't made any attempt at using them to predict independently measurable quantities, we can't test the predictions in any sort of way beyond "hey Snowy, I disagree with X because...", which is obviously fruitless. The "human error" component of the tier list steward exists, whether in your tier list or in SDS's tier list or whatever, but the difference is that the criteria in SDS's tier list are defined in a way that is actually reproducible, minimizing the number of grey-area decisions to be made. There is no mandate in any of the tier lists or the tier list FAQ that the player use only those teams capable of acquiring the absolute lowest turn count possible; if that were the case, the tier lists would omit F-listers like FE13 Donnel.

1. I'd avoid reclassing him entirely and just phase him out of the lead support spot unless you feel you could benefit from a Rally Spd+Def Cynthia, in which case General Fred and Falco Sumia would be good. 2. Dark Knight is one of the more dependable choices. Dark Flier works for the next few non-Paralogue chapters since there won't be that many bow/Wind users for a while, and the ones that do appear have low accuracy; then again, I don't recall how much of a problem bows are on the Paralogues. Hero might be interesting because of Sol. 3. Dark Mage is probably the most effective, but Mercenary might be more fun. Armsthrift and decent Mag allow Morgan to go nuts with Levin Swords, and he'll be even better once acquiring Sol. 4. Falcon Knight; I don't think Sumia has the offense to make good use of Galeforce on Lunatic, and the better stats of Falcon Knight probably will help more than having access to low-ish magic damage.

No, he won't.

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.

OK, I read through your thread. I also had a glance at your actual tier list. It seems to me that your current model of complexity is neither a) reproducible, since you haven't actually defined it with, say, an actual expression nor b) independently measurable. If f(X,Y,Z) is your complexity function, with X, Y, and Z being certain measurable quantities (like enemy strength, etc.), then while it's true that f(X,Y,Z) is a function of independently measurable quantities, f(X,Y,Z) as a model of complexity can't actually be tested against any actual data. As they say, all models are wrong, but some are useful; as of now, it's not obvious how your f(X,Y,Z) is useful or falsifiable. Absolute LTC requires RNG abuse, as you of course know from your playthroughs. The tier list FAQ assumes no RNG abuse, whether it be in the form of ninja dodges or rigged level-ups. Absolute LTC requires a lot of skill, certainly, but it is not unreasonable for the tier list playstyle to reward risk-averse tactics.

You might want to read this FE tiers are made from efficiency results, which are no less objective than tournament results. The choice to adopt low turns as the metric of choice is somewhat arbitrary, but after adopting this standard, the actual teasing-out of who's-more-efficient-than-whom is not. Um, okay? Michael Jordan was unequivocally a better player than Charles Barkley, but that certainly didn't mean that Jordan would necessarily outperform Barkley on any given night. As this pertains to FE, the law of large numbers gives us a very good chance of preventing the observed average stats of any heavily-leveled character from deviating sharply from the true average. This need not mean that Panne is necessarily going to saving more turns than Gaius in any given playthrough; only that the chances of this happening are high. Bases and availability (both of which present opportunities for saving turns) mean something, too. The balance is much worse than in previous games if you're referring to the disparity between above-average and below-average units, though if you're referring to the usability of below-average units, then I would agree; nonetheless, using below-average units over better ones has a larger opportunity cost than in previous games due to the high emphasis on rout maps and enemy-phase combat. As stated in the tier list FAQ, the metric of usefulness is efficiency; playthroughs adopting the tier list playstyle tend to shoot for low turncounts. High-ranking characters allow for faster completions than do low-ranking characters, though in this particular tier list, giant Rescue/Dance chains are not assumed.

We don't have an ab initio way of determining the One Playstyle to Rule Them All, so yes, at the end of the day, it's fine to pick an arbitrary set of rules that makes the game somewhat fun and that rewards strategy (seeing as this is an ostensible strategy game). The goal of the tier list is to determine which units best fulfill the chosen tier list objective, and nothing more.

Voted Syrena for both. Miriel's current +3 Mag over her is not unsurmountable; the Seraph Robe combined with Veteran can help Syrena snowball into a good unit. I don't know if I'd promote her immediately upon reaching Lv 10, since she could use the extra leveling speed. Then again, I could see Ch 9 being tricky on Lunatic+ due to the swarm of enemies near the start point, in which case promoted Syrena may be useful. I never got past Ch 5, so I can't say my advice is all that credible, but investing in Syrena rather than your other units should have a bigger endgame payoff.

I don't see how three Rally Spectrum users would be helpful at all. She'll not make use of her Lance rank as a Wyvern Rider, but promotion to Wyvern Lord allows her to resume using Lances. Because Sully has Discipline, however, she won't really have to deal with the awkward Ladle/Bronze Axe phase for long. Alternatively, you can just promote her to Paladin to continue making use of her current ranks. Her class-changing choice alone will not make or break the game. I don't even think promotion bonuses influence the kids' base stats. My understanding is that for every three stat points a parent gains from his her or bases, the child character will gain one. Even if promotion bonuses counted, a +4 boost to Atk for, say, Sully would amount to a +1.33 boost to her daughter.