Jump to content

Let's talk about Effective Crit


Owns
 Share

Recommended Posts

I wanted to start this conversation because I find that many people dislike these types of builds while I'm personally a fan of them, depending on the character. So I wanted to go through what I think makes a good crit build and why, and hopefully convince more people to give them a chance.

In the end crits come down to chance, and I think that's why many people avoid them because they prefer 'more reliable' builds like brave attacks with enough Mt to ORKO. While I agree that a guaranteed kill is always favorable over something that is chance-based, not many characters can actually pull that off. This leads to players doing what I call the Juggernaut approach: "who should I recruit for X route?" The answer always falls down to Sylvain, Leonie, Lysithea, Ferdinand, Felix, Shamir, Cyril etc., basically anyone who can pull off brave skills pretty well (Swift Strikes, Hunter's Volley, Point-Blank-Volley, Fierce-Iron-Fists, etc.).

Then comes, "what should I do with Ashe/Ingrid/Ignatz/Bernie" and any other poor shmuck who was designed with high speed/dex but low str and the standard answer usually lies between an evade tank incapable of anything else (poor Ingrid) or Sniper for Hunter's Volley, if not recruit someone else to use entirely. Don't get me wrong, all the above is sound and good advice because it works. Especially the Hunter's Volley approach for the bow boon characters, but I wanted to take a look at the mechanics (and numbers) behind crits and hopefully shed a light on how to maximize them. At the same time I'll look at the value of the Dex (and Lck by analogy) for such builds.

Okay so to start off, Crit = (Dex + Lck) / 2. Which as many point out, means you only add 0.5% chance to crit with each Dex point and from that many conclude that Dex sucks. The thing is, with these types of things you want to maximize your chance to crit and there's two ways to do so: 1) Increase your crit chance via Dex, Lck, and Crit from abilities/weapons. 2) Increase you chance of landing a crit by actually hitting more frequently. If you've ever played Rogue type classes in MMORPGs or other games you might recall those crit-relying units always having higher attack speed or what comes down to more hits in the same amount of time. The reason here being that the more attempts you can make "per turn", the more you're likely to crit, and this goes up in a non-linear fashion. Take a look at the graph below which shows your chance of triggering at least 1 crit out of the number of hits you make in 1 action:

chance-to-crit.jpg.9b959fae8b46ed38735f8adac4039d78.jpg

Edit: I made reference plots of effective crits from some of these cases taking into account hit as well. They are posted further down in this thread.

[Apologies for any color-blind people seeing this, higher hit goes up in the curves, sequentially].

So while the battle predictor tells you your "Crit chance per hit", the above plot is the actual "Effective Crit" for the entire attack. While yes, "a point in Dex only adds 0.5% chance to crit", that's only true for single hits. If you are getting more hits, then you can actually get more than 1% chance to crit per point of Dex (and Lck), depending on how high your crit already is. At lower values, each point of Dex (and Lck) contributes more towards your chance.

The next plot might be a little more confusing, it shows how much each added point of Dex (or Lck) adds to your chance to crit. Meaning if you're at 21 Dex for example, the y-axis value is how much Crit you'd gain going from 20 Dex to 21 Dex, etc.

Dex-value.jpg.904925c56a3aed8f321dcf552ffd5ac0.jpg

Okay so what does all this mean? Well it means that the first plot is the "effective" Crit you're getting depending on how many hits you land. Clearly, it's a lot easier to land "at least 1 crit" the more chances you can attack, and often times this is all you need for these lower str characters to get a ORKO. Not on the toughest enemies like armored units, grapplers, warriors and war masters, but most anyone else (like mages, archers, thieves, and when using effective weapons). It also means that the value of Dex can be higher than 0.5% chance to crit if you make more attempts.

For example, Dedue or Alois using One-Two Punch with only a 20 displayed crit chance, is actually a 60% of getting a crit (or roughly 35% chance of getting a crit in the first round of double punches).

This is why for example using Hunter's Volley is such a great strategy for some of these characters. It not only guarantees 2 hits, but also adds 10 to Crit. On top of that Snipers tend to have high dex/lck stats making it fairly easy to land at least 1 crit with Hunter's Volley, especially with a crit bow.

This also explains why Felix wielding gauntlets can land at least 1 crit with nearly every attack as you get close to the end game, because he has high Dex, and Spd allowing him to frequently get x4 hits with a good base crit chance.

It also means, that you can land a stupid amount of crits using Astra combined with high crit chance. This is the real reason why Astra costs so much durability, not the mediocre 1.5x base damage which would suck if that was all. Finally I wanted to look at another interesting skill and how it looks under this type of analysis: Lethality. So Lethality triggers as Dex/4, which sucks, unless? Well here's the graph:

Lethality-chance.jpg.d05d8eac8c31dddde3b9b594751a8d26.jpg

Suddenly it doesn't look too bad for an insta-kill ability if you combine it with Astra. Wait, what? I can use an unlimited Rusted Sword to get kills with 1/4 to 1/3 chance? (Okay maybe don't since they weigh 20, ruining your avoid, but a cheap Training Sword works too). At 22-23 Dex you get a 1/4 chance of Lethality using Astra, while you need a 31 Dex for a 1/3 chance (using Astra). That's sounds pretty amazing to me (and like a lot of fun). Low str characters, can typically easily evade axe wielding enemies but deal little to no damage to them, why not take that chance if they have nothing better to do on that turn? Admittedly, I don't know if Lethality can trigger if you're only doing 0 damage. The only reason I say this is because I noticed monster barriers don't get hurt if you do 0 damage, so idk.

So ultimately, should you go with str or dex? Well you want a balance of both since even if you crit you'll only be dealing 3x damage and if that damage is only 10 you're not getting a ORKO either way. That said, if you're at this junction of needing "at least 1 crit to ORKO" then you want to maximize Dex/Lck as well as "No. of Hits". Correct me if I'm wrong, but I think that usually, these low str characters are in this situation of needing at least 1 crit, unless they have Death Blow and using Hunter's Volley against a flier. And Deathblow can be quite out of the way for some of them (and unnecessary if you build crits). From my experience though, honestly anything above 70 Crit (as displayed in the battle-prediction) seems to land almost every single time. And as low as 30 Crit triggers more often than I expect it to. I don't know if the double rng is used for crits or not. If anything crits feel skewed towards triggering but that's just a feeling, and feelings are not only subjective, they also depend on expectations and mine are usually low, so don't take my word for it.

The last thing I want to say is that while any % Chance should on average yield that chance, these things are ultimately binary in this system. You roll a number, if it's less than or equal to your crit it lands. Say your crit is 30 and you rolled a 31, no crit. But had you just one more crit then it would've landed. It's not like crit damage scales with your chance, it's either x3 damage or it isn't. So that's why for me, every single point of Dex/Lck matters. Again, on average you expect to get the same % of crits as your crit value but that's only true for a very large number of trials. Anyone who's played around with rng's before knows the results can be very skewed in the short term (in either direction). As an example, take a look at the coin toss simulation on the wikipedia page of the Gambler's Fallacy (not that the gambler's fallacy is relevant here). Watch it restart say 5-10 times, and see how often one side gets favored over the other in that limited number of trials. Usually by a small amount, but this is why I say every point matters. To conclude, obviously if you have a guaranteed route to ORKO then go for it, but for characters who can't achieve that, consider a crit build that maximizes both Crit and No. of Hits.

Edit: See further discussion/analysis of Astra's effective crit chance of getting at least X crits out of the 5 hits, below.

Edit: Another analysis comparing Astra and double hits, here.

Edited by Owns
Corrected error about One-Two-punches
Link to comment
Share on other sites

  • Replies 61
  • Created
  • Last Reply

Top Posters In This Topic

I didn't know 1-2 Punch could trigger 4 hits? Though of course, other brave weapons can double for x4 as well.

Nitpicking aside, this is a really impressive post, thanks for putting the time in.

I love using crit builds as well, and it's noteworthy playing with them because they always feel dangerous, even when they shouldn't be, and for me too they seem to perform better than displayed crit would suggest (I know that's just subjective though). 

Astra + Lethality seems like a great idea that no one seems to have tried, and if anything like the best possible usage of Astra. At first, seeing 30% damage per Astra hit made me feel it was too nerfed for me to want to use it for crit builds (you'd need Astra attacks to crit three times minimum in order to exceed the damage output of a normal crit, which isn't unreasonable if you build for it, but can be annoying to rely on). But now, especially knowing how many resources there are to boost Crit, it's much more attractive. Looks like I'm finally gonna try Felix as a Swordmaster.

Link to comment
Share on other sites

Okay, on principle I like where your head's at. Crit builds can be made quite reliable in 3H(though I'm not going to lie to you, I think that's generally well known), and it's pretty fun. However, I have a few things to point out.

- It's rare to have trouble killing without crits if you quad. The issue is that usually, you can't, or you die on counter, or it requires a ton of stat boosters. Also One-two punch doesn't autoquad(man, can you imagine?). What it does is make gauntlets perform a followup at +8 damage instead of a brave at +0.

- The problem with Astra crits is that they compensate more than anything; it's like with your multiple hits example for rogues in mmos; it's nice, but 5 crits on 10 * 1000 damage is the exact same as 1 crit on 2 * 5000(assuming you don't get separate bonuses for critting, which granted, you often do... but that's not a thing in Fire Emblem, either way). Also Astra damage is rounded down for each hit, which doesn't help. The main issue really is that one or two crits do not guarantee a kill with Astra the way they would with everything else, which is what really matters at the end of the day. And assuming you need at least three to kill(which is far from impossible, after all that'd only result ~3.3 * damage, which isn't exactly crazy), suddenly what your Dex contributes to actually killing is lower than even for brave attacks.
As for Lethality, the issue is that we're talking about mastering two middling classes to perform a combo that still doesn't give that good a chance at killing(and remember, if you don't kill at 1 range, you're likely in considerable danger). It could be a fun build, I'll give you that, but yeah. And for anything else, there's the issue of lethality actually not being needed to kill in the first place.
I have to say, I don't know if lethality can trigger on 0 damage hits either.

- Death blow is actually assumed on every physical unit basically. For physical units that aren't trying to maximize speed, there's hardly a better way to spend most of your intermediate tier time. Also, Bernadetta has Vengeance, she doesn't need crit builds to kill.

- And the big one; while yes, Dex is made better by multiple hit attacks... well so is strength(or mag in some rare cases, or just mt in general). The reason crit is relevant regardless is as a result of being easy to raise in many cases(no thanks to Dex and Lck), and not having access to enough strength to compensate for typically lower base damage(which itself exists to limit multi-hit attacks); but if you actually have the choice, it's a bit of a no brainer. The only situation where this might be untrue is for Astra since it's the only one that actually nerfs strength... Outside of that, it's considerably more direct to buff a character's attack by 5 to 10 points to reach 2HKO thresholds, rather than stack dozens of points of dex and lck to raise crit chance.

Link to comment
Share on other sites

1 hour ago, Cysx said:

Also One-two punch doesn't autoquad(man, can you imagine?). What it does is make gauntlets perform a followup at +8 damage instead of a brave at +0.

While it doesn't automatically do so, One Two Punch does work with Desperation. If you don't mind mastering Cavalier, that can be really good on a Warmaster Dedue. 

Link to comment
Share on other sites

3 hours ago, haarhaarhaar said:

I didn't know 1-2 Punch could trigger 4 hits? Though of course, other brave weapons can double for x4 as well.

It doesn't, OP is mistaken. 1-2 Punch hits once, takes the counter-attack, then hits once more. So it's the same crit odds as doubling with a normal weapon. It's still a good combat art, but not "guaranteed 4 hits on player phase" good.

Neat graphs, though. And I like the topic of discussion, in general. You did miss a couple of tools to consider for such builds, though, outside of abilities, weapons, combat arts, and skills. Namely, equippables and battalions. The Critical Ring grants +5 crit, while the Accuracy Ring gives +10 hit. Small boosts, certainly, but they can provide an edge on crit-based builds. Also, the Goddess Ring grants +4 Luck, translating to a very modest +2 crit. As for Battalions, Fraldarius Soldiers stand out with an enviable +20 Crit, while fliers should strive for the +15 Hit and Crit offered by Cichol Wyvern Co. Leicester Mercenaries and Goneril Valkyries, each providing +20 Hit and +15 Crit, are also strong tools in this regard.

Link to comment
Share on other sites

Good thread!

It's uncontroversial that the best thing you can do is one-rounding enemies. Higher-str characters can achieve that fairly easily using brave effects, as has already been mentioned. Certain combat arts can also do well. Failing that, the next best thing is to reliably chip so others can kill. Mages and archers tend to be best at this, because they have incredible range (giving them both more targeting options AND letting them avoid counters more often).

Crits are mostly valuable if you want your ranged chip to sometimes (even often) kill, thus saving you actions. Which is why low-str characters opting for Hunter's Volley is such a solid choice. That said, Dex and Luck are still of questionable value, at least compared to str. 4 points of Dex mean 2 more crit, which is <4% more effective crit. By comparison, 4 more points of Str means 8 more reliable damage, which significantly increases the number of enemies you can reliably ORKO with Hunter's Volley, a much more important increase in offence than 4% crit.

Crits on range 1 moves is a build I have trouble having much respect for. If you go for an unreliable kill and fail, you almost always take a counter. Additionally, by needing to get to melee range, you have fewer targeting options, a significant disadvantage compared to other builds which don't reliably kill (i.e. low-str archers and mages). It's hard to see what a build optimized for melee crit brings to the table (besides being fun, which is subjective of course).

Link to comment
Share on other sites

7 hours ago, Armchair General said:

Critting means nothing if you can't enemy phase your units.

True enough, but like I mentioned these tend to be builds with either safe range or high avo so can remain relatively safe more often than not.

6 hours ago, haarhaarhaar said:

I didn't know 1-2 Punch could trigger 4 hits? Though of course, other brave weapons can double for x4 as well.

 

6 hours ago, Cysx said:

- It's rare to have trouble killing without crits if you quad. The issue is that usually, you can't, or you die on counter, or it requires a ton of stat boosters. Also One-two punch doesn't autoquad(man, can you imagine?). What it does is make gauntlets perform a followup at +8 damage instead of a brave at +0.

Admittedly I wasn't sure about One-Two Punch having never tried it. From its description it sounded like x4 total hits since gauntlets double by default and it provides a follow-up attack. Then are you saying One-Two Punch has a total of 2 hits with higher might and taking a counter in between? I meant to double check this point before posting and then I carelessly forgot.

 

6 hours ago, Cysx said:

As for Lethality, the issue is that we're talking about mastering two middling classes to perform a combo that still doesn't give that good a chance at killing(and remember, if you don't kill at 1 range, you're likely in considerable danger). It could be a fun build, I'll give you that, but yeah. And for anything else, there's the issue of lethality actually not being needed to kill in the first place.

Okay I just want to raise a couple of points here. 1) While yes, Swordmaster and Assassin are only Advanced classes, they provide growth rates and modifiers that are comparable to most Master Classes, albeit highly specialized into Dex and Speed for the most part, but yes the 1 range can be a high risk depending on the situation for sure.

2) About Deathblow getting you over the 2HKO threshold, how true is this really? I'm not really convinced that low str characters like Ashe, Ingrid, Ignatz as examples will start ORKO'ing everything even with Deathblow. Here's a crude estimate I did using VW Ch20 enemy stats to see who a Sniper Ignatz could potentially ORKO. Not sure what the lvl would be here, I used lvl 35 average stats for a Sniper build and compared ORKO's with/without Deathblow and with/without getting at least 1 crit:
ignatz-dblow-v-crit-HV.thumb.jpg.6950a38b41712cab61c99f0f36fec775.jpg

You can see the str/mt values I used above the list, hopefully I didn't forget any important source of damage. All cases assume I'm using Hunter's Volley to attack the enemy unit. What I'm seeing is he can only ORKO the weakest enemies, which is fine but you can do better if you land at least 1 crit with or without Deathblow. The important comparison here is that getting at least 1 crit, without deathblow, nabs you more kills than deathblow without any crits.

Obviously, the best scenario is having both Deathblow and getting at least 1 crit which I think is a fairly reasonable thing to achieve. But if I absolutely had to choose between Deathblow or successfully getting a crit, then maximizing crit is better. Ie. if going out of  my way to get deathblow comes at the expense of reaching S for crit +10, then I'd rather not get deathblow. Now I don't have enough experience/knowledge to say if this type of sacrifice has to be made in the first place, maybe someone else can tell me (assuming no excessive grinding). My guess would be that yes, if the character doesn't have a boon in Axes then you might not be able to get both Deathblow and S in bows without grinding. Whereas going straight bows you can reach S+ for both an extra Bowfaire as well as Bow Crit +10 which beats Deathblow for sure. Bear in mind that typical end game enemies on Maddening have Lck values between 17 and 25 or so, reducing that much from your crit chance. But it's quite reasonable to reach above 70 crit with bows which as I've shown above, is an effective crit of around 90 for double hits (especially since Hunter's Volley lends 10 crit as well). This is all to say that crit builds probably do best when they are highly specialized and attempt to reach at least S with the desired weapon.

6 hours ago, Cysx said:

Outside of that, it's considerably more direct to buff a character's attack by 5 to 10 points to reach 2HKO thresholds, rather than stack dozens of points of dex and lck to raise crit chance.

When you say buff by 5 to 10 points, are you implying using boosters? Also, you don't really need to "stack" dex and lck, it mostly comes from natural dex growths from your path progression. The rest is direct crit from the wapon itself and other equipment/ability options. Obviously if you have some dex/lck boosters lying around then it makes sense to pump them into a crit-built unit but I wouldn't go out of my way to farm them (maybe unless I wanted to have fun with Lethality). I do agree that given the choice, a guaranteed ORKO relying on str alone is the better path though, no arguments there.

 

Edited by Owns
Link to comment
Share on other sites

13 hours ago, LoneRecon400 said:

While it doesn't automatically do so, One Two Punch does work with Desperation. If you don't mind mastering Cavalier, that can be really good on a Warmaster Dedue. 

That's an interesting combination... I think I'd still rather just go for FiF, though it's admittedly not hard to justify spending some time in cavalier for Dedue.

10 hours ago, Dark Holy Elf said:

Crits are mostly valuable if you want your ranged chip to sometimes (even often) kill, thus saving you actions.

Man, it's unfortunate that we still don't seem to see eye to eye on that...

8 hours ago, Owns said:

Then are you saying One-Two Punch has a total of 2 hits with higher might and taking a counter in between? I meant to double check this point before posting and then I carelessly forgot.

Yup. That's what it does.

I wasn't saying those two classes were "middling" in that sense, I am a big advocate for Grappler/Sniper over WM/BK on any character that cannot double reliably after all; what I meant is that they aren't the greatest. Swordmaster is unfortunately a lower movement class specializing in one of the more limited weapon types in the game. It's true that statistically it isn't bad per se, but yeah, let's just say that if mastering made Astra useable on every class, there's be limited reason to stay in SM. Assassin on the other hand absolutely is better, with 6 move ignoring most terrain +shade for survivability, but any Astra build couldn't ultimately take advantage of those. It's also ultimately hurt by also specializing in a low damage weapon, combined with having no Str mod or for what it's worth, growth.

I also need to rephrase about Death blow; I wasn't saying that it was enough to kill everything, just going against your idea that characters might not have it, because there's almost no reason for them not to. It's the most valuable thing they can get from intermediate, especially if they're male, and brigand is a good class to be in. It also doesn't go against crit builds at all since intermediate classes provide next to nothing in that regard beyond +10% Dex growth from archer. At best it competes with hit +20 to hit reliably from farther.
To be clear, in a theoretical context where those characters would need to deal 4 times their damage to start ORKOing, dex would look a bit better, it's true. But even they aren't that weak, and 3H brings up anyone with class bases and all the direct stat boosts it provides.

You did forget str +2 and bowfaire in that table.

8 hours ago, Owns said:

if going out of  my way to get deathblow comes at the expense of reaching S for crit +10, then I'd rather not get deathblow. (...) My guess would be that yes, if the character doesn't have a boon in Axes then you might not be able to get both Deathblow and S in bows without grinding. Whereas going straight bows you can reach S+ for both an extra Bowfaire as well as Bow Crit +10 which beats Deathblow for sure.

For better or for worse, that's not how the game is laid down for us. S to S+ is 1080 Wexp, E to D+ is 180, boons don't come close to compensating for such a difference.
Getting S rank and Death blow isn't unreasonable at all either(it typically means not having it for a month, at worse two), and even if it was, I'd definitely argue that going for +10 crit is the wrong choice. Nnot taking into account that you'd have Death blow for a good 10 more chapters, 10 isn't the difference between completely reliable and dangerously luck based, not on HV anyway. 60 crit(so 84% with the brave effect) falls into luck based, 70(91% as you said) into fairly reliable, 80(96%) basically into completely reliable, though you'd probably want a bit more still.

8 hours ago, Owns said:

When you say buff by 5 to 10 points, are you implying using boosters?

Well uh, yeah. I thought that's what you were doing yourself to be honest, since as far as class progression goes, going through archer as opposed to brigand amount to +1 dex in 10 levels and all.

Not trying to be a stick in the mud btw. I mean it, I do appreciate the analysis. I just... disagree with your conclusions. Also, I run crit builds on these characters too, I can attest that they work, many others can as well. So I'm generally on your side of the bigger argument.
... well, one of the two bigger arguments, in any case, because when it comes to the value of Dex... yeah.

 

Edited by Cysx
Link to comment
Share on other sites

First of all, this was a great post, with a lot of detail and thought. Thanks for taking the time to make it.

Now, for myself, one of the reasons I tend to avoid crit builds is that they increase the complexity of the decision tree I build in my mind as I play, which increases the likelihood that I make mistakes. To be clear, this is very much a personal preference thing and not something that I'm presenting as objectively optimal, but let me explain what I mean. I think we all agree that the ideal situation is to have an attack with a 100% hit rate that is guaranteed to kill an enemy, but that we also agree that there are a lot of circumstances where this isn't possible. In any situation where it isn't possible -- less than 100% hit, needing to get a crit, having a possibility of an enemy skill proc, etc. -- then we generally need to make contingency plans. "I'll make this attack and if it kills the enemy then this will be my next move, but if it doesn't then I'll have to do that instead", that sort of thing.

The ideal circumstance for any contingency planning is to have a full decision tree for all possible outcomes, that will make the most of good luck if we get it (eg, by pushing forward further, killing more enemies, claiming secondary objectives, etc.) but that can react to and mitigate bad luck if we get that (eg, not overextending, having backup strategies, keeping units in reserve in case they are needed, etc.). It isn't typically possible to have a full decision tree for every possible outcome for an entire turn, but it is possible to plan for a few different outcomes, and I like to consider as many contingencies as possible before making a move.

Crit builds make it more difficult to effectively plan because there are more possible outcomes that need to be accounted for. Instead of having to plan for what to do if I miss and what if I hit, I have to plan for missing, hitting without a crit, and hitting with a crit. If I'm double attacking, then I may need to consider the difference between critting on the first attack and critting on the second attack, since that may be the difference between taking a counterattack and not. It's just more things to keep in mind. I know from past experience that the more variables I am trying to keep track of in my head, the more likely I am to miscalculate or misremember something and make a mistake. I like to keep things simple not because I think it's necessarily best, but to try to play around my own human fallibility.

Link to comment
Share on other sites

2 hours ago, Cysx said:

60 crit(84% real) falls into luck based, 70(91% as you said) into fairly reliable, 80(96%)

Wait, you’re saying that Crits use a 2 RN system in this game? I thought Crits always used 1 RN and Hit was what used 2 RN?

Link to comment
Share on other sites

4 hours ago, Cysx said:

Man, it's unfortunate that we still don't seem to see eye to eye on that...

Well, feel free to elaborate! I think you make very well-thought out arguments and if you feel differently than me I'm definitely gonna listen.

I come from the perspective that it's not that hard to reach ORKO levels without crits in the second half of this game (and I say this as someone who doesn't garden for str boosters, which seems like it'd make it even easier?), and with all the tools you have to boost accuracy those get very reliable even using nominally low-hit (e.g. brave) weapons, while crit builds are often very high investment to reach even semi-reliable numbers (requiring an S rank is a huge negative to me), and even then are inferior because "semi-reliable" tends to mean only 90-95%.

Link to comment
Share on other sites

Thanks to everyone who clarified about 1-2punches, corrected in the OP.

@Cysx Okay, good catch, I've added str +2 and Bowfaire to the Ignatz analysis I did above and well it just supports my argument further. He gets practically the same kills with or without Deathblow (assuming 0 crits), but getting at least 1 crit opens up far more ORKO options. So I still think raising crit is superior to going out of my way for Deathblow depending on the build I want. This shouldn't be too surprising tbh, you're comparing +6 or +12 damage to x3 or x6 damage. The only downside is that of course the crit is chance-based while deathblow is guaranteed and I guess here it simply comes down to the player's preference. The way I see it, the odds of killing tougher enemies with crit are better than the odds of failing to ORKO without Deathblow.

ignatz-dblow-v-crit-HV.thumb.jpg.cabcd2d15d4ff815bec596e859b92281.jpg

You make an important point about the gap between going from S to S+ vs reaching D+ for Brigand but I don't think it's that simple. You're not factoring in all the extra Wexp you could be getting from class weapon proficiency being in the right class as opposed to sidetracking to Brigand for Deathblow. I don't think it'll make up the gap, but I wouldn't think it insignificant either, I don't have those numbers to crunch either way. One thing I can say though, and this goes to @Dark Holy Elf as well, is that it isn't difficult to reach S or even S+ if you focus. In my current run my Leonie reached A in bows by the end of Ch. 9's mission (ie. starting Ch. 10) while she also had a C in Auth, C+ in Lances and a C in Riding having gone Soldier > Archer > Sniper > Cavalier. As soon as she reached A in Bows I stopped training her Bow and switched to cavalier to focus on riding to make her a Bow Knight. She's currently sitting at A in Bows, A in Riding, B in Lances, C+ in Auth as a Lv 28 Paladin on Ch. 14's Mission Day. Had I continued focusing on Bows (and stayed in a Bow proficiency class), I expect she'd be at S rank already.

Now, I wanted to take a a closer look at Astra since it's been discussed a bit but not its mechanics. As has already been pointed out before, dealing 0.3x5 damage means you need to get at least 3 crits (out of the 5 hits) to deal more damage than you would with a normal Critical Hit (ie. you'd get x3.3 vs x3 damage). That said, you're vastly more likely to crit with Astra than without, and if your crit is already high, then your'e much more likely to get 3 or more crits with Astra. Let's look at the numbers to put this into context. Below I've plotted the likelihood of getting at least X number of crits using Astra as a function of your battle Crit:

Astra-analysis.jpg.e8feebd7be03f165b535574591e71515.jpg

[If you're color-blind, in this case the curves go down sequentially from P(at least 1) to P(all 5)]

So you can see that all you need is 50 Crit (battle) for Astra to (consistently) deal more damage than a normal Critical Hit. For swords, 50 Crit is nothing and very easy to achieve. Below the plot I've listed a typical Crit build to look at some Battle crit values one could achieve using endgame enemy Luck stats. This is assuming no dex/lck boosters and based on Ignatz growths. I actually shortchanged his dex+lck for a conservative measure. I'm also using the higher end of the Lck spectrum which very few typical enemies actually have (many of them are around 17 Lck). Basically, as a Swordmaster, you can reach >90 battle Crit for the vast majority of enemies if optimizing, or >80 with a sub-optimal setup. This practically guarantees at least 3 Crits using Astra with a high likelihood of getting 4 or even all 5 hits to crit. Even with a Silver Sword+ you can get >50 Crit depending on the enemy.

The last bit of info in the above image is comparing damage from different weapons, basically I'm asking: Is it better to get more crits or have higher mt? The base case is assuming n crits hit so long as they have a higher chance than a single normal attack Critical Hit. What this shows is that higher str deals more damage than more crits, not suprising since increasing attempts does much more for your effective crit chance than increasing the value of Crit. But going for Mt is also less reliable with the best option obviously being Cursed Ashiya Sword+. While these weapons do have a low durability of only 20-25 (meaning only two uses of Astra, pitiful), they're also readily available using only Black-Sand Steel to forge/repair (apart from Sword of Zoltan which needs Wootz Steel). You don't need to use Astra all the time anyway. Not in the image, but even a Killing Edge+ is fine as it has the same Mt as Wo Dao+ and only 5 less Crit. So I'd probably have an inventory of 1 Cursed Ashiya Sword+, the rest of the slots with either Wo Dao+ or Killing Edge+, a Crit Ring, maybe an Evasion Ring if the SM was doubling as an evade tank, maybe a Training Sword+ if I also had Lethality and just want to try my luck.

Major EDITs from here:
Okay all this is good and interesting, but how much damage are we actually doing, am I getting ORKO's with this? I wanted to do a similar analysis as the one above with Hunter's Volley but as I mentioned before, I'm not entirely sure how Astra's damage is calculated. EDIT: Thanks @Shanty Pete's 1st Mate and @Whisky for explaining. So the formula is:

Dmg = RoundDown{ [total.mt. - target.prot]* 0.3 * 3 } * n.crit + RoundDown{ [total.mt. - target.prot]* 0.3 } * (5 - n.crit)

 

Okay so as already established, str is going to be more valuable than Dex. So i'm thinking a good progression path might be: Fighter* > Brigand* > Mercenary* > Hero* > Swordmaster. This way you have final abilities to choose from of: Deathblow, Str +2, Defiant Str, possibly Vantage, Sword Crit +10, Sword Prowess, Axe Breaker, possibly Battalion Desperation. I compared this route with going Fighter > Thief > Archer > Assassin > Swordmaster and the difference in final stats is minuscule, whereas the first one yields much more useful abilities (unless you wanted to play around with Lethality+Astra). Using Cursed Ashiya Sword can help you get down to Vantage and Defiant Str thresholds. Or the other strategy many use is to rely on Blessing. Just speaking hypothetically here. Anyway, let's look at who we're ORKO'ing. The results can vary quite a bit depending on weapon might or character str (say Felix vs Ignatz). So i'll put a couple of examples hidden in spoilers, this is based on lv 35 Ignatz with 24 str as well as Deathblow and Str +2. Anything highlighted in green is a kill:

1. Cursed Ashiya Sword+

Spoiler

Astra-ignatz-cursed-ashiya.thumb.jpg.00f54d63898dbe39151f61fba0a60a7e.jpg

 

2. Wo Dao+ / Killing Edge+

Spoiler

Astra-ignatz-wo-dao.thumb.jpg.2e62856c29b61501e543b99fbb69f151.jpg

 

3. Sword of Zoltan+ / Devil Sword+

Spoiler

Astra-ignatz-zoltan.thumb.jpg.c6d5e09aad5d4813ff55d96f334fca3d.jpg

Honestly the potential damage is pretty impressive and can be improved significantly with str boosts whether that is simply a higher str character or using boosters. The question that remains is how reliable is it? If you're fully optimized using Cursed Ashiya Sword+, I'd say it's reliable since you can get >90 Crit against most enemies, meaning you can expect to get all 5 hits as crits. But for the most part I'd probably use Wo Dao+ or Killing Edge+ and that would still be pretty good since they can reliably get at least 3 crits, and frequently 4 or even 5 crits (and Wo Dao only weighs 5 allowing maximum AS/Evade which will be important on the front lines). By the way, even Silver Sword+ is viable here, you can still get >40 crit chance easily, putting you at >92 for at least 1 crit and >66 for at least 2 crits opening up more or less the same ORKO's as you see for Cursed Ashiya Sword under those categories. At that point though, you might as well normal attack with Wo Dao+ (especially if you can double) but it's a way to spread the repair costs.

Is this better than Hunter's Volley? Not if you're reliably getting at least 1 crit there, which won't be as reliable but not uncommon either. The advantage of Sniper of course is safety from range. Swordmaster can have pretty good evade but the class inevitably still remains in some risk of danger. So if a character can do either, I'd probably default to Sniper as the safer bet, unless I wanted to diversify or try something different. But I think this shows that SM can at least be pretty good, probably more than it's given credit for.

Edited by Owns
edited with correct Astra damage formula
Link to comment
Share on other sites

6 minutes ago, Owns said:

Now, let's take a closer look at Astra. Dealing 0.3x5 damage, looks like a total of 1.5x damage on the surface as a minimum (ie. with 0 crits). But actually, I think if you get 0 crits then you're really looking at 0 damage if my damage calculation is correct. Admittedly I'm not certain about it, but I'm assuming each hit of the 5 is calculated the same way as a normal hit, in which case for each hit you're looking at


0.3 x Mt - Target's Prot

per hit, which on Maddening means you'll be getting 0 dmg on even squishy enemies (maybe unless you're a high str character). If I'm right about this calculation then the whole point of using Astra is to help increase crit chance. As has been pointed out, you need to get at least 3 crits (out of the 5 hits) to deal more damage than you would with a normal single attack that also crits (ie. you'd get x3.3 vs x3 damage). That said, you're vastly more likely to crit with Astra than without, and if your crit is already high, then your much more likely to get 3 or more crits with Astra. Let's look at the numbers to put this into context. Below I've plotted the likelihood of getting X number of crits using Astra as a function of your battle Crit:

I'm pretty sure Astra damage, per hit, is calculated as floor(0.3*(Atk - Prt)). If it multiplied Mt by 0.3 before subtracting enemy Prt, then it would basically never deal damage, even with a crit (0 times 3 is still 0).

Link to comment
Share on other sites

6 minutes ago, Shanty Pete's 1st Mate said:

I'm pretty sure Astra damage, per hit, is calculated as floor(0.3*(Atk - Prt)). If it multiplied Mt by 0.3 before subtracting enemy Prt, then it would basically never deal damage, even with a crit (0 times 3 is still 0).

Yeah, your damage per Hit is 30% of the damage you would deal with a normal attack, however the damage is rounded down so the total damage is usually less than 150% of a normal hit (excluding Crits). Astra does not add any Crit and actually has a Hit penalty, making it pretty bad overall. Certainly much worse than Hunters Volley before even factoring in the 9 durability cost. and if you can double, then doubling will almost always be more effective than Astra. It can be useful at times for a Swordmaster if you don’t mind using nearly half of a weapon’s durability in some cases, but I wouldn’t go out of my way for it. Could be fun to use though, that’s a different discussion. I believe the most effective use of Astra is with a Rapier against mounted enemies.

Link to comment
Share on other sites

If anyone wants to see an example of how Astra works in practice, see here. (Though kind in mind with similar amounts of investment, things like This are possible.)

Also TC, i gotta ask, how much Aux battles do you do? Because if you participate in a lot of them, it can really skew a unit's Wexp. Like fighting an extra 20 battles as a Swordmaster with a grant an additional 100 Sword Exp without even a boon.

Edited by LoneRecon400
Link to comment
Share on other sites

2 hours ago, Shanty Pete's 1st Mate said:

I'm pretty sure Astra damage, per hit, is calculated as floor(0.3*(Atk - Prt)). If it multiplied Mt by 0.3 before subtracting enemy Prt, then it would basically never deal damage, even with a crit (0 times 3 is still 0).

 

1 hour ago, Whisky said:

Yeah, your damage per Hit is 30% of the damage you would deal with a normal attack,

Oh! of course, that makes much more sense.  Okay, I've edited the earlier post with further damage analysis and comments.

13 minutes ago, LoneRecon400 said:

If anyone wants to see an example of how Astra works in practice, see here. (Though kind in mind with this amount of investment, things like This are possible.)

Also TC, i gotta ask, how much Aux battles do you? Because if you participate in a lot of them, it can really skew a unit's Wexp. Like fighting an extra 20 battles as a Swordmaster with a grant an additional 100 Sword Exp without even a boon.

Thanks for that, confirms what the previous 2 said about how its damage works. About Aux battles, for my current run i only do 1 day of battles each month, using all my battle points each time. Meaning starting from Ch 10 I was doing 3 battles (whether aux/paralogues) per month, which I think is reasonable? Oh and don't really use grinding strategies during those battles. In one of my earlier months in this run I did 2 days of battle but that was the only time (as a result of poor planning more than anything). With the help of some of you fine SF members, I've been trying to improve my Professor Level efficiency and adopted this approach for my current run. You can see this post about specifics if you're interested. Sadly I actually didn't have a chest Key with me when I did Sothis' paralogue which means I didn't even get a Knowledge gem until Ch. 12, didn't feel like repeating it.

Edited by Owns
Link to comment
Share on other sites

22 hours ago, Dark Holy Elf said:

Well, feel free to elaborate! I think you make very well-thought out arguments and if you feel differently than me I'm definitely gonna listen.

I come from the perspective that it's not that hard to reach ORKO levels without crits in the second half of this game (and I say this as someone who doesn't garden for str boosters, which seems like it'd make it even easier?), and with all the tools you have to boost accuracy those get very reliable even using nominally low-hit (e.g. brave) weapons, while crit builds are often very high investment to reach even semi-reliable numbers (requiring an S rank is a huge negative to me), and even then are inferior because "semi-reliable" tends to mean only 90-95%.

Thanks for the kind words! But frankly I think there's ample merit to both stances on this one, especially in a context where indeed, nothing stops you from just giving everyone strength/magic boosters, and indeed, you'll want S rank in your weapon. They also have limited enemy phase potential of course.
Here's a list of the advantages the way I see it:

- Killer weaponry durability is entirely sustained by money from very early on
- There's an ample amount of 15 crit battalions; similarly, crit rings are in very low demand
- Those builds don't really have weaknesses beyond very high def enemies, which are rarely dangerous or difficult to take down; they tend to deal with other generics fine with or without crits until lategame where they should have all of their tools
- They shine brighter than any other when it comes to dealing with demonic beasts(since most have low luck and no proficiency, leading to easy 90+ displayed rates), and are also competent boss killers
- They do not struggle with the final chapters the way many other builds do, since their ideal damage is consistently overkill
- Just like in Echoes, you can abuse divine pulse to rig crits if needed, which gives them a reliability push as you have ample charges and there are few ways to use those offensively.

Posting numbers for the sake of it. This is at level 30, enemies have around 25 crit avo at that point on Maddening. No S rank, it's possible but let's just not.

Sniper: ~25 crit from (Dex+Lck)/2, + 35 from Killer+, + 5 from crit ring, +15 from battalion, +10 from HV = 90. 65 displayed, aka 88% chance.
Grappler: ~20 crit from(Dex+Lck)/2, + 30 from Killer +, +5 from crit ring, +15 from battalion, +10 from FiF = 80. 55 displayed, aka 91% chance.
Warmaster. ~20 crit from (Dex+Lck)/2, +30 from Killer+, +5 from crit ring, +15 from battalion, + Crit +20 = 90. 65 displayed, aka 88% chance.

So yeah, them's the breaks. In practice, they're regular damage builds unless facing beasts or crit rigging for bosses, until the lategame where they've hopefully acquired their S rank and need to rely on crits to kill. As said, the good thing is that they're straightforward, and deal with problematic things well. But they do walk the reliability line a bit, and all of these classes are on foot as well.

-------

Just as a disclaimer, haarhaarhaar did not repeat what I wrote, if anything it's the opposite(happened unwillingly, though) since everything under this is an edit.

13 hours ago, Owns said:

I've added str +2 and Bowfaire to the Ignatz analysis I did above and well it just supports my argument further. (...) The only downside is that of course the crit is chance-based while deathblow is guaranteed and I guess here it simply comes down to the player's preference. The way I see it, the odds of killing tougher enemies with crit are better than the odds of failing to ORKO without Deathblow.

Wait, no, my argument never was that deathblow is indispensable, I was asking for a good reason to skip it. Though to be frank, guaranteeing kills on roughly half the maps, that doesn't sound awful, but then again, ch 20 has a lot of low def enemies.
 

13 hours ago, Owns said:

You're not factoring in all the extra Wexp you could be getting from class weapon proficiency being in the right class as opposed to sidetracking to Brigand for Deathblow. I don't think it'll make up the gap, but I wouldn't think it insignificant either, I don't have those numbers to crunch either way.

That's fair, though this is easy to calculate. It takes 50 rounds of combat to master Brigand, and Archer gets +2 bow exp per fight, so that amounts to a clean 100 Wexp loss. Added to the 180 *1.5(because boon) from earlier, that's 370 bow Wexp. Which isn't nothing, but still not that significant.
Also, crit builds usually don't need S+ rank in the first place, as they deal enough damage already. Well, if they have Deathblow, anyway, but foregoing a +6 damage boost for most of the game in order to get a +5 one at the end... that just doesn't make much sense. Which I wish I brought up earlier.
 

13 hours ago, Owns said:

One thing I can say though, is that it isn't difficult to reach S or even S+ if you focus.

I agree, believe it or not. But wait, what was that about Deathblow getting in the way of getting S rank, then?


Aaaaasutora

Spoiler

Okay so, Imma need your math here. That may be on me(like genuinely, ironically I suck at math, I'm just thorough enough that it compensates... sometimes), but in my head a 50% chance to crit 5 times amounts to 2.5 crits, not 3.
... well, that might not be worth getting into, actually. The table gives a good idea, which is the most important.

- Lategame enemies do have 25 Lck, but they also have +10 from proficiency. And while earlier they get lower luck for sure, the further you go back, the less likely you are to have +10 crit as a skill; overall, I think your numbers are too high for most chapters.
- The Cursed Ashiya Sword+ has 70 Hit(and the Devil Sword + has 65). As Astra lowers hit by 10, and the most accurate +15 crit battalion gives +20, you're looking at a 120 accuracy rate. Considering lategame maddening enemies have ~40 to 70 avoid before terrain, even with true hit it's highly likely that you'll miss at least once in five hits, lowering the damage output a good bunch. Since this is by far the best crit weapon swords get, that's a problem. And yeah, there's the durability thing.

As for other individual weapons, using your table, one has roughly a 90% chance to crit thrice with a Wo dao+ and a crit rate of 80. That's frankly not that good and leaves a lot of enemies alive, on top of the possibility of misses on sword users and lance users with swordbreaker.The same general logic applies a lot to Zoltan(which outside of Azure moon is pretty much a last chapter only weapon) and anything with low crit; they're too inconsistent in a context where the user could die on counter.

It's nice to have expected damage tables, though, major props for that. But to me clearly Astra doesn't work here, there really are limited ways to make it consistent due to hit issues and the randomness, on top of the rest of what's wrong with Astra.

 

11 hours ago, LoneRecon400 said:

If anyone wants to see an example of how Astra works in practice, see here. (Though kind in mind with similar amounts of investment, things like This are possible.)

And now I regret not looking lower earlier, because that's way more comprehensive than my walls of text... even though GD 22 isn't the kindest playground and she could easily have swordfaire² I mean she's goddamn Catherine too
Out of curiosity, on what occasion did you screencap all of that? Just for research?

Edited by Cysx
Link to comment
Share on other sites

13 hours ago, Owns said:

1-2 Cursed Ashiya Sword+

Isn't there only one Cursed Ashiya Sword available per run? It's important, because that only gives 2 Astra uses per battle, and especially in the late game with HP inflation, Astra from weaker swords like Wo Dao aren't necessarily going to kill even if they crit 5 times (and stronger swords obviously have significantly lower crit than CAS).

I also just realised that your graphs of crit activation only hold when the hit rate is at 100 (because, correct me if I'm wrong, the crit chance doesn't actually function independently of hit). Given that Cursed Ashiya Sword has base 70 Hit, and Astra brings -10, anyone with 45-50 Avo or more is probably an issue (lots of enemies have less Avo, so this isn't a fatal concern though). 

That endgame Catherine (from @LoneRecon400's post) had 21 base crit (only 17 Luck, and her bases and growths are fairly average for the roster), which means between 75-90 crit with CAS+ (excluding against monsters). Her displayed hit rates for CAS+ Astra aren't likely to be above 78 (90 True Hit) against anyone except axe infantry. Assuming 90 true hit and 85 crit, that's 94.5% chance of 3 crits, and 26% chance of 5 crits. But CAS+ is the only sword that is likely to achieve crit above 80, so there are only two Astra attacks per map with odds that could be that good. Effective damage Rapier, Sword of Zoltan+ and Mercurius can beat out CAS+ Mt and have more uses, but all have lower crit, the former is only sometimes available and the latter two are harder to maintain. Wo Dao+ is far more available, which'd easily get 60 Crit, and could get to 80 Crit, but -5 Mt per Astra attack is likely to be a dealbreaker for KOs, particularly against physical units in the lategame (even 5 crits aren't guaranteed to kill endgame enemy, and there's only a 24% chance of that with 100 TH and 75 crit).

EDIT: Having had more of a look into enemy stats, things aren't nearly as bad with Wo Dao+ as I had originally thought, so the so-called 'solutions' below might not be necessary at all. I'd still want Death Blow and a higher-strength character than Ignatz, but I'd be fairly confident if I had those.

The best solution is probably adding in S+ Swordfaire/Defiant Strength. That should save Wo Dao+, and it'll probably let CAS+ get kills it couldn't before. Defiant Strength is very achievable, but it's fairly finicky, and you're gonna have to protect your sword user every turn (a problem, but not a fatal one). As for Swordfaire, I'll just say that some of us are unfortunate enough not to get S+ regularly/ever on anyone, so for players like that it's very difficult to rely upon. If you are getting it though, then it's probably around Ch. 20-1, in time for the late chapters and the stat escalation.

Another potential solution to the lack of late-game killing power is to try and find a better unit for this build - Felix is an obvious candidate, and will have better Dex/Luck than Catherine, but I doubt it would be enough to make a real difference. The only people I can think of who might really make a difference would be Ingrid and Sylvain, whose Crests have 40% percent activation rate to raise CA Mt by 5 (which means you can probably rely on them triggering twice per Astra). Ingrid in particular will need Strength help to be of real use, but it's not impossible to achieve.

Another is to throw Lethality into the mix. Catherine was also running Prowess, Crit+10, Death Blow, Axebreaker, and Wt -3. Wt -3 is a good skill, but if she got a single Rocky Burdock, Catherine's AS/Avo would only go down by 1 if she were to unequip Wt -3. In other words, Lethality could replace it without much loss. Which means that Lethality could be fit into an Astra crit build. I'm expecting on average my units to get roughly 25 Dex, and my best Dex units to be at most 35-6 Dex (from growths/class boosts). Anyway, assuming I'm not trying to farm Dex boosters, I won't be able to raise any Swordmaster's Dex beyond 40. And with Astra, that's roughly a 34% chance of killing non-boss enemies. Catherine had 25 Dex (her bases and growths are again basically the average for 3H characters, so that works out), so her chance of triggering Lethality with Astra would have been just under 23%. Adding in your odds of Lethality can bring your total KO chance up to double what it was with just the probability of KOing the enemy with 4-5 crits, although it still doesn't make this build anywhere near certain. 

Just some thoughts from me (I'm still definitely gonna try some version of this build anyway).

Edited by haarhaarhaar
Link to comment
Share on other sites

2 hours ago, Cysx said:

Out of curiosity, on what occasion did you screencap all of that? Just for research?

Just trying to contribute to the thread. It was just on a casual playthrough I had that laying where I made Catherine a Wyvern Lord. 

Actually had to grind her out in Swordmaster like this since I didn't bother to raise swords or master Swordmaster.

Link to comment
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

 Share

  • Recently Browsing   0 members

    • No registered users viewing this page.
×
×
  • Create New...