Thread: Alternative to total power for evaluating a stats stick eidolon

Disclaimer: this is a long post with what may be a controversial opinion.

The tl;dr version is that I think total power is misleading and that Tiamat is not necessarily the best stats stick as a result.


I have been tinkering a lot with my eidolon setup to get the best stats for my teams. I consistently find total power to be unhelpful though as it implicitly assumes that 1 atk is worth the same as 1 hp. Different situations and different teams will value atk/hp differently, so trying to pin down what the 'correct' trade-off is not that meaningful. Instead, I asked the question of how much atk I can get by sacrificing 1 hp and vice versa using event eidolons. This gave me the following:



Interestingly, the top of the cloud of points traces out an almost perfectly straight line (red line in chart above) and spans a good range of hp/atk levels. This means that for almost any target hp level, there is a 'best' stats stick - one that maximizes atk. If you use only eidolons on/near the red line to hit the target hp level, this will also automatically give you the best atk possible.

A few highlights:

1. Tiamat, commonly considered the best stats stick is NOT on the red line - this means that if you actually need the hp, Tiamat may not be optimal for you.
2. Sakura x Mio, commonly considered the worst stats stick is actually as good as Tiamat based on distance from the red line - this means that if you need the hp, she's actually not terrible.
3. Eidolons on the red line spans the range of 600-800+, so as long as the average hp you are targeting for you sub eidolons fall in that range, there is really no reason to use Tiamat or Sakura as stats stick (assuming you own the eidolons on the red line of course.)

The red line above implicitly assumes that you are running a rainbow team. You can repeat the exercise accounting for the fact that you get a 10% bonus for himes matching the element of the eidolons. Since I main light, I did it for light and got the following:



Barong, Abou, Yule Goat and Nandi are the 'best' light eidolons depending on your target hp level for light. High power stats stick like St. Nick and Sphinx are actually sub-optimal unless you really don't need the hp.

So are you saying Tiamat is bad?

No. I am only saying that if you need hp, Tiamat is bad since you sacrifice more hp than 'necessary'. If you don't own the eidolons on the red line or you're a water main, the situation is obviously different.

Who even cares about hp, don't you just maximize atk?

My personal experience is that hp is very important for things you can't retry like tower, or if you are trying to carry a rag, where long battle means you usually get some consecutive DATA on the same hime eventually. If you have a ton of ascension, you also want your hp high to take full advantage of the heal, since most heals are capped around 20% of hp (ascension raises that cap, but not enough to keep pace with how much it raises your heal.) If that is not your experience, feel free to ignore what I am saying here.

So should I ignore total power rating?

No, it is still useful for comparing eidolons with similar hp. All I am pointing out here is that the implicit 1:1 atk-hp trade does not exists - if you actually tried to mix and match high and low hp eidolons, the best trade you can make is roughly 3.5:1-ish.

What about active abilities?

The analysis above only looks at stats. This is most relevant to AAB builds or short battles where you're not going to be using the eidolons' active abilities anyway. For other situations, it is obviously important to weigh the stats against what actives the eido brings.

Tiamat for example, has a very useful active, so if you're not using Python, Tiamat is still a good choice despite not being on the red line.

What about kaisers and gacha eidolons?

MLB kaisers are currently THE best stats stick no matter what metric you use. MLB gacha eidolons are also superior to event eidolons in terms of stats. Problem is, at LB0, both kaisers and gacha eidolons are actually worse than MLB event eidolons in terms of stats, and most ppl dun have a bunch of MLB kaisers and gacha eidolons, while most ppl DO have MLB event eidolons. You can easily repeat the analysis with gacha eidolons if you want though. I didn't do it 'cos it's not relevant to me.

EDIT:

Did the gacha eidolon analysis, too by request. See a few posts down.

What about defender weapons?

There is a reason why I am not considering weapons at all in this analysis. It is simply because most of the time, you don't have nearly the flexibility of what to use for weapons compared to what to use for sub eidolons. This is especially true if you are going for phantom grid, where simply gathering all the correct types of weapons is hard enough, and you usually have no room to decide whether you want an extra defender weapon in your grid or not - that is pretty much decided by which phantom grid you're going for.

Also, defender values are only as good as your base stats. If your base hp is low, defender isn't going to be that helpful. If you need to trade-off some atk for hp, changing your eidolon setup is your best bet. This is why I find myself consistently gravitating towards 'bad' stats sticks like Python and Barong, 'cos despite their low total power, they give me a decent boost to base hp while minimizing the sacrifice to atk.

What about upcoming events?

I looked ahead and performed the same analysis for Rashape, Yatokami, Nine Shots and Hes. They're actually all right on or very close to the red line (i.e. all excellent stats stick.) However, they're all in the 600-ish hp range, so if you are targeting higher or lower hp, the next several events will not really help you.

Originally Posted by dreamlitz

So should I ignore total power rating?
--

I, at least, entirely ignore "total power." All I use is "Atk"

Really, once you have 5 MLB SSR Eidolons, will you really need any more base HP in this age of double-skill weapons...? Chances are if you need survival, you're better off using Joan instead of more base HP, no?

Originally Posted by Slashley

I, at least, entirely ignore "total power." All I use is "Atk"

Really, once you have 5 MLB SSR Eidolons, will you really need any more base HP in this age of double-skill weapons...? Chances are if you need survival, you're better off using Joan instead of more base HP, no?

But then you have things like dark ult/rag (probably others, idk) where she doesnt even attack until od, and just dots you to death. So far its the only instance I've seen where having a fuckton of hp would do a world of good.

Originally Posted by Nik

But then you have things like dark ult/rag (probably others, idk) where she doesnt even attack until od, and just dots you to death. So far its the only instance I've seen where having a fuckton of hp would do a world of good.

DoTs are based on your maximum health, so most of the time, having more HP is BAD. Because having more HP only means that healing does less.

That said, it's also possible to cap out DoTs, depending on the DoT. In which case having more HP starts being... okay... but still not great.

Very interesting. Would have been interesting to see the results for gacha eidos too. I'd always had the impression that event eidos with high hp got shafted on attack, but seems like that isn't necessarily the case.

Originally Posted by Slashley

I, at least, entirely ignore "total power." All I use is "Atk"

Honestly, I think this is more sensible than looking at total power.

Chances are if you need survival, you're better off using Joan instead of more base HP, no?

I do that for solo content, but for raids, it's hard to (V)MVP with Joan (actually it's just hard to (V)MVP in general, but even harder with Joan...) And helping out the raid with Joan/Andromeda is sort of a thankless job >.>

Originally Posted by Nik

But then you have things like dark ult/rag (probably others, idk) where she doesnt even attack until od, and just dots you to death.

Originally Posted by Slashley

That said, it's also possible to cap out DoTs, depending on the DoT. In which case having more HP starts being... okay... but still not great.

For dark ult, all DoT cap out at 750, so high hp actually really helps. If you're increasing your in-battle hp from 6k to 7k though, dun bother, it just makes it harder for you to heal up as Slashley stated. I went from 10k to 15k by messing around with eidolons and FLB-ing one of weapons, and it helped a lot for dark ult.

Originally Posted by Tanukimo

Very interesting. Would have been interesting to see the results for gacha eidos too.

Ok, maybe when I have time I'll take a look.

I'd always had the impression that event eidos with high hp got shafted on attack, but seems like that isn't necessarily the case.

Actually they kinda are - the slope of the line suggests that you have to give up about 3.5 atk to gain 1 hp, which is why high hp event eidolons always look so awful in total power. This is also why I caveat everything with if you need the hp. I actually often find that trade to be a bargain for AAB builds where the AI does a lot of stupid things so I need a bigger hp pool to offset its stupidity (I AAB AQ5 and most ults - not dark ult, actually I AAB that, too and rely on friends to clean up my mess ) If you don't find the hp useful, do what Slashley does and only look at atk (in which case Tiamat rules, by a pretty big margin.)

Originally Posted by dreamlitz

Actually they kinda are - the slope of the line suggests that you have to give up about 3.5 atk to gain 1 hp, which is why high hp event eidolons always look so awful in total power. This is also why I caveat everything with if you need the hp. I actually often find that trade to be a bargain for AAB builds where the AI does a lot of stupid things so I need a bigger hp pool to offset its stupidity (I AAB AQ5 and most ults - not dark ult, actually I AAB that, too and rely on friends to clean up my mess ) If you don't find the hp useful, do what Slashley does and only look at atk (in which case Tiamat rules, by a pretty big margin.)

A similar slope also holds for weapons within their quality type. I did something similar awhile ago and noticed that all weapons released have a similar curve depending on their accepted power level. Different weapon grades have different slopes (SRs are below what you're quoting, while SSRs are a bit above).

All else being equal, more hit points help period. This is actually the really big appeal of phantom weapons and their grid capabilities. Even with a couple of off-element weapons (provided their maxed) you can easily match the expected damage of an on-element non-phantom grid while upping your party's hitpoints by about 25-35% (depending on you amount of defender). You lose no damage output, but gain high resiliency. If you have the ability to go all on-element then you also benefit from the damage increase, but it's less because everyone has a decent base of assault to begin with. The real power advancement is left by base assault at that point.


Dejnov.

Originally Posted by Dejnov

-- This is actually the really big appeal of phantom weapons and their grid capabilities. Even with a couple of off-element weapons (provided their maxed) you can easily match the expected damage of an on-element non-phantom grid while upping your party's hitpoints by about 25-35% (depending on you amount of defender). --

Eerrrhhh? This sounds like massive bullshit, and _might_ only hold water with Lance Grids. Even then only for one off-element Lance. Maybe.

There is no way that adding off-element weapons will do you ANY good past extremely early game when you don't have on-element weapons yet. I'd like to link my damage calcs again, but sadly, neither of them have support for such details when it comes to survival like HP increase from Phantom weapons.

In any event, don't underestimate just how much 10% constant Assault or Defender is for your team. It easily surpasses Base Atk or HP gains. Also, here's some stats from my teams:
Fire: W: 1742 HP, 24727 Atk (Phantom), E: 3549 HP, 13703 Atk
Water: W: 1937 HP, 30084 Atk (Phantom), E: 3641 HP, 14496 Atk
Wind: W: 1393 HP, 24141 Atk, E: 3924 HP, 14457 Atk
Thunder: W: 2764 HP, 34490 Atk (Phantom), E: 3894 HP, 15255 Atk
The point here is, the majority of your HP is going to come from Eidolons. Even if you focus solely on Atk and thus get use the little HP ones, and even if you run a Phantom Grid with full +99 weapons (like my Thunder).

So, I'll say it again: As soon as you get a full MLB set of Eidolons, HP shouldn't be an issue. If it is, you're going to need to wait for Assault/HP double-skill weapon to really make a difference.

Originally Posted by Slashley

Eerrrhhh? This sounds like massive bullshit, and _might_ only hold water with Lance Grids. Even then only for one off-element Lance. Maybe.

There is no way that adding off-element weapons will do you ANY good past extremely early game when you don't have on-element weapons yet. I'd like to link my damage calcs again, but sadly, neither of them have support for such details when it comes to survival like HP increase from Phantom weapons.

In any event, don't underestimate just how much 10% constant Assault or Defender is for your team. It easily surpasses Base Atk or HP gains. The point here is, the majority of your HP is going to come from Eidolons. Even if you focus solely on Atk and thus get use the little HP ones, and even if you run a Phantom Grid with full +99 weapons (like my Thunder).

Warning: Wall of Text!!

So I understand what you're saying and what I'm saying really does sound like bullshit. I've been really hesitant to share what I've experienced for two reasons. One, it sounds wrong or at least counter to what we know about how the game works. Two, I initially couldn't explain what I'd seen. I think I might be able to now. I only know when I've been building these off-element lance grids (as that is my only maxed phantom weapon atm) in weaker element grids, my damage can go up. It doesn't always go up (it has gone down by 3-5% also in some situations) and it's usually not by much, usually 3-5%, but because of the hitpoint and the defense bonus in a phantom grid my hitpoints for my teams have gone up a lot (like I mentioned sometimes as high as 30%, but easily 20%). This has made my teams a lot more durable, while sacrificing very little in damage output ability (and, as mentioned, sometimes with a net gain). I've done a lot of experimenting with the grids (all on-element no phantom, defense SR placeholders to stay phantom, and 1 or more off-element in phantom layout, etc.) and I think my results should be easily replicated and I might be able to explain them now.

One caveat: I'm not a true long term player. I played when the game first came out, but due to work/life balance, I had to drop the game for almost a year. I've been back as a stable player for the last 9 months or so. I only have two really veteran level teams (with himes, weapons, and eidolons) - Light and Dark. My other teams have been a constant race to upgrade prior to their corresponding tower event and are still a work in progress. They are all assault and they are fairly well skill leveled and full assault, but even that tops out at 13-15 avg. skill level across all ten weapons in the grid (usually it's 1 or 2 FLBed full skill leveled weapon) and one or two zero star SSR at 9-12 skill level at the bottom end slowly being raised.

All comparisons are against a non-phantom full assault on-element grid versus a phantom grid which might use defense SRs or off-element SSRs in place of the missing 1 or 2 weapons to complete the grid. All grids are usually tested against a normal raid run (no element advantage, no skills used, exact same support Eidolons) to analyze raw damage and hitpoints seen in battle to quantify the change I expect. When I did the comparison, I found that the on-element defense SR was actually giving me less damage output than the off-element substitute and it was sometimes giving me less damage than the full on-element grid. Since the off-element weapon can only supply base attack (prior to any multiplicative effects), this has led me to presume what I'm seeing is an effect of base attack instead of any assault related multiplier effect.

My best example is my water grid. I have 3 water lances and, as mentioned earlier, the old configuration was a on-element all assault full SSR grid, avg. skill level on all weapons is 15, as the baseline comparison. The weapon I've been substituting in all grids is my FLBed +99 Michael Lance with a base atk of 3195 and phantom atk of 4153 (in grid). The standard Hime reported atk in the non-phantom grid is around 50000 atk. What I have in my grids is usually a couple of zero or one star assault/skill on-element SSR weapons that I've been substituting with the above mentioned Michael lance (to complete the phantom effect). The base zero or 1 star SSR usually provides around 10 to 13% additional character assault, while also having a base assault of around 1400 atk. For easy math purposes, I'll start generalizing now. This means that my chosen substitute usually brings about 2500 base atk to the grid for the loss of 12.5% character assault. I'll assume that I have around 100% character assault and 100% eidolon assault for the grid so that my on-element weapon is giving me an additional 6.25% damage output. When I substitute the Michael lance I lose that. I do gain 2500 base atk, which works out to about a 5% increase in base assault (based on the Hime reported atk of 50000). After the weapon is in the grid, all base attack gets multiplied by current character assault and eidolon assault which means that the additional 2500 base atk gets multiplied by 4 (x2 character attack, x2 eidolon attack) to give me a full gain in combat of around 10000 atk. This works out to a damage gain of 18.7% (5% x 1.87 character attack x 2 eidolon attack). Please note: I haven't seen this as the water lances I have used are not skill leveled or even weapon leveled to the base on-element grid, but I have been seeing a damage increase. I also may be assuming incorrectly how the effect is occurring (and if you have insight, I'd like to know). My current damage gain is currently around 3-6% (in both water and wind grids) and as I'm raising my Shingen and other on-element lances, it's been going up. The effect is dramatically less pronounced if you have to substitute two off-element weapons and is usually negative when I substitute three off-element lances. This leads me to believe once you hit the character attack bonus enough it starts to penalize the increase in base attack you get from an off-element weapon and the effect I'm seeing is lost fairly quickly.


Dejnov.

Originally Posted by Tanukimo

Would have been interesting to see the results for gacha eidos too.

Kinda slow at work today, so I decided to do this. Gacha eidos actually turned out to be pretty interesting, though I'm not sure if there's much practical applications to this:



I distinguished between eidolons released before 100% (green squares) and after (blue circles). The dotted purple line is my best shot of a best fit frontier through only the green squares excluding kaisers and red line is the best fit frontier for everything excluding kaisers and mythical beasts (won't be out until next year for us.) There are several interesting observations here:

1. The addition of 100%+ eidolons actually pushed out the frontier noticeably as seen by the difference between the purple and red line.
1a. Kirin and Belial are the only 100% eidolons that are not on or close to the new frontier (red line.)
1b. Huanglong and Hecatonchires are the only non-p2w eidolons that are actually closer to the new than old frontier (looks more like it is part of the red rather than purple line.)
2. The 120% (140% eventually) eidolons are all between the old and new frontier (i.e. stats are better than old gacha eidolons but worst than the 100%.)
3. With the exception of fluffy and Anubis, all blue dots have comparatively low hp (all below 700.)
4. Both frontiers have more favorable atk/hp trade ratio than the one I estimated for event eidolons (purple is about 2 atk per hp, red is about 3 atk per hp, event eidolon is about 3.5 atk per hp.)

On point 4, it is important to note that ALL non-p2w gacha eidolons are above the event eidolon frontier:



In above chart, green squares are non-p2w gacha eidos with purple dotted frontier, blue circles are event eidos with red frontier.

Corollary

1. If you're f2p or dolphin, none of this is relevant to you, since it's not like you'll have a bunch of MLB gacha ediolons sitting around to play with.
2. If you're a whale, none of this is relevant to you, since you should go for kaisers (or mythical beasts when they're out) if you want stats sticks (they are literally off-the-chart - I had to change the axes specifically to accommodate them.)
3. If you're a chubby dolphin/some other large-but-not-too-large marine animal, this MIGHT be relevant to you:
3a. Rule-of-thumb - use as many MLB gacha eidolons as you can, as they're all better value than even the best event eidolons.
3b. If you can't use all MLB gacha eidolons 'cos you need certain actives, certain elements or what not, your best bet is to get as much hp from gacha eidolons and get atk from event eidolons - it is 'cheaper' to get hp from gacha eidos (roughly 2:1 trade vs 3.5:1 for event eidolons.) For example, if you're using Behemoth and Azathoth and you really need to add barrier as an active, use Tiamat instead of Python and kick Azathoth instead of Behemoth.
3c. Huanglong and Hecatonchires are the best value for you if you want high hp stats sticks. If you want even higher hp, Behemoth is the best gacha non-p2w.

Originally Posted by Dejnov

A similar slope also holds for weapons within their quality type. I did something similar awhile ago and noticed that all weapons released have a similar curve depending on their accepted power level. Different weapon grades have different slopes (SRs are below what you're quoting, while SSRs are a bit above).

That is interesting. Aren't most weapons between like 100-200 hp-ish though?

Originally Posted by Slashley

So, I'll say it again: As soon as you get a full MLB set of Eidolons, HP shouldn't be an issue. If it is, you're going to need to wait for Assault/HP double-skill weapon to really make a difference.

Sometimes there's a drought of good weapons though. For light, the next assault/defender weapon is Sphinx reprint like 4 months from now... I don't want to wait that long, lol, so I try to tinker with what I have. I managed to boost my base hp from 8k to 9k-ish and base atk only dropped from around 51k to 49k-ish once I realized I was using sub-optimal eidos. It's not earth-shattering, but enough to make a lot of my battles (especially AAB ones) clear with some cushion as opposed to a nail-biter every time/outright wiped occasionally.

Originally Posted by Dejnov

I found that the on-element defense SR was actually giving me less damage output than the off-element substitute and it was sometimes giving me less damage than the full on-element grid.

Actually, you don't even need phantom weapons for this to happen. I used rainbow grids a lot for my f2p accounts during their infancy for this reason (put in whatever advent and UE weapons I had regardless of element instead of using SRs.) Defender weapons give zero assault value, so replacing it with higher stats weapon will automatically boosts its dmg. On the hp side, if your base hp is low, 13% defender doesn't do you much good, and having a +99 off-element weapon can often offset the defender value effect (e.g. at base 5k hp, dropping 1 M defender weapon loses you 650 hp, and a +99 FLB weapon can easily have 200-300 more hp than SR; if you still have other defender weapons left on the grid, that 200-300 translates to even more, so you may only end up losing a couple hundred hp or so.)

What I have in my grids is usually a couple of zero or one star assault/skill on-element SSR weapons that I've been substituting with the above mentioned Michael lance (to complete the phantom effect). The base zero or 1 star SSR usually provides around 10 to 13% additional character assault, while also having a base assault of around 1400 atk.

I think what's happening here is that you're kicking off an on-element low-stat assault weapon with a high-stat off-element, but also activating the phantom effect - the phantom assault from lance equals or exceeds the assault lost from replacing the on-element, and the stats from the off-element exceeds the on-element, so every component of the dmg calc is actually going up.

However, when you do this with a second replacement, the assault value goes down with no offsetting increases (since phantom lance skill has already been activated by the previous replacement.) The raw atk is still going up, so the question is whether the % increase in raw atk exceeds the % decrease in assault or not. The calculation is actually not hard to do, just pulling all the data together is a little tedious (need the exact assault value of the grid and base atk of all the himes accounting for weapon preferences as well - 20% boost from favored weapons.) Generally though, this is not a favorable trade at all unless your on-element weapons are just really bad (which they sound like they are in your case, lol, no offense meant to your water grid... My water grid sucks, too, so...)