Dev Tracker: Discussion (Old)

[Toth]

Thank you for the translations slycer. I am not really into the mathy parts of the game but can you confirm that it appears in summary that they are going to change the system so that regular attacks and WS will cause less enmity buildup /quicker decay from certain jobs (they mention changing how much each job can generate enmity-wise, no?) and allow PLDs and other tanks jobs to have higher enmity generation and 'caps'?

[Slycer]

(they mention changing how much each job can generate enmity-wise, no?) and allow PLDs and other tanks jobs to have higher enmity generation and 'caps'?

The implication of what he wrote was that they weren't going to go that route, at least not initially. Maybe more clarity in official translation, but only the section with the four bullets at the very bottom is what's currently planned for adjustment.

[dejet]

Sounds like the decay part is where they want to start. I get the feeling this is going to be a revamp of most of the system.

I just hope whatever the plan is makes a geared pld have the power told hate vs a well geared war.

[Toth]

Okay so standard attack enmity will change but how does that help tank jobs get more enmity then?

[Slycer]

Okay so standard attack enmity will change but how does that help tank jobs get more enmity then?

The issue with the current system isn't that tanking doesn't generate enough enmity, it's that DD generates too much.

[Toth]

Okay, I think I get it. By lessening the amount of hate they create and generate from standard attacks, DDs will generate less enmity and decay hate faster, allowing PLDs and other tanks to maintain constant hate with their abilities and such. Is that in the same ballpark?

[Xanthe]

I haven't played in about a year and a half, but I like to keep up on the news. This is excellent information and I'm glad to see that there are plans to revamp enmity.

As a quick example of the ridiculousness, take a 3000 damage weapon skill. Assuming standard damage of 65, a 3000 damage weapon skill generates 11077 time-volatile enmity (which is probably over the cap anyway, but he didn't say exactly what the cap was). For that time-volatile enmity to completely decay, you'd have to stand there idle for over 3 minutes without doing anything. You'd also have to take twice your maximum HP in damage for the damage-volatile enmity that same WS generated to decay.

I think this captures the essence of the problem with enmity as it stands. Tanks do an excellent job of building and maintaining enmity, but DDs can build and hold it just as well through ridiculous damage output, hence hate bounces from one target to the next with every enemy auto-attack and TP move. If the rate of buildup and decay of enmity via damage can be adjusted (Matsui's first planned move) I agree that many issues with tanks maintaining their spot on the top of the enmity list will correct themselves. The proof of course is in the implementation results, but I think that this is an excellent first step.

I also agree that it's awesome to see so many of the player-derived formulas being indirectly acknowledged as correct.

[Byrthnoth]

Yeah, sounds like they're decreasing both"Level-related modifiers" from Kaeko's work, which is sensible considering they have been essentially constant since 75 while our damage potential has skyrocketed. Then they say they're going to adjust (probably downwards) other JAs and such.

I can think of various JAs and enmity sources they might forget to adjust during this process (resulting in unintended tanks), but by and large this is the simplest solution that would have the intended effect. I also believe that I have seen it advocated across various forums since about the 85 cap.

[Darkmagi]

I just hope this doesn't make curing take too much enmity to do long-ish fights with DDs that take heavy damage.

[Byrthnoth]

I just hope this doesn't make curing take too much enmity to do long-ish fights with DDs that take heavy damage.

I believe that this would be the sole purpose of the enmity update.

[Slycer]

One thing confuses me. I had to check two or three times to make sure I didn't miss anything:

Calculation of decay from damage taken
When a PC receives "d" damage from a monster, the damage-volatile enmity of the PC is reduced.

Damage-volatile enmity reduction = (1800 * d)/PC's max HP

In other words, if the PC takes the same amount of damage as their maximum HP, 1800 damage-volatile enmity is decayed.

*This calculation is subject to correction by Sentinel's effect.
*From the point of view of back-line jobs, this may be too low. By increasing the 1800 value, more enmity will be decayed when damage is taken. However, this would also have the reverse effect of making it difficult to maintain enmity when receiving damage as a tank. This all needs to be taken into consideration when making adjustments.
*For healing magic, the value is half of the above calculation.

The last note, about healing magic - what is he getting at? I really have no idea what that reference is. Do you lose enmity explicitly when someone heals you? That's how I'm reading it, and I didn't know that. If all original enmity tests were only done with two players, this would have been an invisible variable that the tester would not have realized, which would have led to the assumption that cures are actually generating more enmity than they really do if they also reduce the target's enmity proportional to the amount healed.

[Darkmagi]

I just assumed it would make PLD viable for fights that it would actually be difficult to DD tank. Plenty of things are quite harmless but still require a lot of curing from your mages endless mp pool.

[Slycer]

We know separately that curing someone generates hate, depending on the amount cured for most cure spells (except V/VI). However, I never heard or read anyone say that BEING cured actually REDUCES your own enmity.

The implication of the quoted section above is when you are healed for "h" HP, your own enmity is reduced by:

Damage-volatile enmity reduction = (900 * h)/your max HP

My question is was this something that was ever tested for or were all cure enmity tests only ever done with one player targeting another?

[Stromgarde]

I think it's more likely that this was just poorly placed in the context of the post. Healing magic uses a curve that's half the damage dealt calculation, and deliberately knocking down your hp using astral rings/etc. was a viable tool for building hate when you had nothing else to do with your mp.

[Slycer]

I think it's more likely that this was just poorly placed in the context of the post. Healing magic uses a curve that's half the damage dealt calculation, and deliberately knocking down your hp using astral rings/etc. was a viable tool for building hate when you had nothing else to do with your mp.

That would make sense. I have no idea how he could have done that, though, considering it's right along with two other notes that are placed correctly.

Edit: Actually it doesn't make a whole lot of sense considering "standard damage" would be nonsense terminology for cure spells, but we do know they are level correlated to target, the same way dealing damage is. Hopefully the official translation clears it up, but I'm sure the comment will be located in the same place.

[Stromgarde]

'Standard damage' is somewhat nonsense in the first place. There's zero correlation between the damage of an average sword and the actual attenuation curve. On its face it sounds an analogy pulled out of thin air because it sounded nice. To be fair though, the actual level-dependent function does 'work' in the sense that the higher level you are, the less enmity you generate via damage but like Byrth mentioned the slope flattens out when it should take a sharp dip.

As for the curing part I have to say it's certainly not impossible, just unintuitive and counterproductive to giving Paladins the ability to cure themselves. It should be fairly easily tested by having two players gain equal levels of hate and then having a third party cure one of them to see if it causes the mob to change targets to the other of the first two.

Graphs

Spoiler: show

[Therin]

I can't wait to gear my PLD back up. Old school baby.

[Byrthnoth]

Yeah, compared to the graph of damage vs. level for most 240 delay swords in the game:

I excluded some weird ones (like Trainee Sword).

Anyway, I don't think this "Standard Damage with a 240 delay weapon" is necessarily meant to be taken the way you guys seem to be interpreting it. He probably meant it more like:
1) Let us create a quantity called "Standard Damage", for instance, the damage that you do with a 240 delay sword under the right conditions.
2) Let us define the VE of this quantity as (240*d)/"standard damage"
3) Let us define the CE of this quantity as (60*d)/"standard damage"
4) Observe, VE/4 = CE
5) Also observe, if it's 240 delay and 60 delay = 1 second, then despite being higher, VE totally decays in the time it takes to swing the the sword.

I think he was just setting up a convenient test case. We know for sure that VE and CE are based on damage dealt, which is affected by things like attack, fSTR, etc. and furthermore it's not constant due to pDIF. The comparisons he makes don't make any sense unless you assume that he's defining standard damage as #2/3 above.

[Slycer]

1) Let us create a quantity called "Standard Damage", for instance, the damage that you do with a 240 delay sword under the right conditions.

He explicitly said it's the D rating (DMG rating in English) of a hypothetical baseline 240 delay sword for that level. Not the damage you deal under certain conditions, but the DMG rating of the weapon itself.

Meaningless moonspeak below, but you can see the letter D at least. Quoted text says that the "standard damage" value is about the same as the "D value" (DMG value) of the reference data used when creating a 240 delay sword for that level.

※この値は、武器データを作成するときの基準値（攻撃間隔240の片手剣）のD値とほぼ同じにしました。

[Xanthe]

1) Let us create a quantity called "Standard Damage", for instance, the damage that you do with a 240 delay sword under the right conditions.
2) Let us define the VE of this quantity as (240*d)/"standard damage"
3) Let us define the CE of this quantity as (80*d)/"standard damage"
4) Observe, VE/3 = CE
5) Also observe, if it's 240 delay and 80 delay = 1 second, then despite being higher, VE totally decays in the time it takes to swing the the sword.

Fixed your numbers slightly. VE and CE are a 3:1 ratio.