Does Level Correction apply to everything?

[ringthree]

Right now we know that level correction applies to damage because we can see observable differences in damage based on level and because the formulas that we got a long time ago tell us that is the case. The formulas are taken on faith but they do fit the empirical evidence so there is good reason to accept them. Although the level correction functions differently, it is known to apply in some manner to both melee and magical damage.

My question is if we know that level correction applies to damage, why do we not make the same assumptions for other stats like accuracy/evasion and critical hits? Is it simply because we can see no way to test it? Should we be giving advice on gearing based on the current level of testing if it does not include level correction? Does level correction even exist for these statistics?

I have considered some hypothetical situations that may demonstrate the need to reconsider how some people currently look at the previously mentioned stats.

Scenario 1: Evasion - A long time ago it was suggested that Evasion being a melee stat may scale like a melee stat above skill level 200. Meaning that above that level each level of skill only gives .9 of a skill instead of a whole point. This would mean that flat Evasion would be better to decrease the accuracy of an opponent than Evasion skill. Even if this was not the case, then shouldn't flat Evasion be at least equal to Evasion skill? The problem is that this observation generally did not pan out, and it did indeed look like Evasion skill had a larger effect than Evasion. This could be because there is a check between a mob's "weapon" skill level and the target's Evasion skill level before the actual chance to hit is determined. I believe there are several places where a similar skill vs. skill check is made. This may not be as much of a "level correction" situation but it serves the same function.

Scenario 2: Accuracy - The current trend for many accuracy estimates is to look at currently known information for some mobs (almost always XP level mobs) and then extrapolate them out to higher level mobs. I do not know how accurately this has been tested, but I have definitely found that the accuracy drop for higher level mobs does not seem to correlate. If there was no correction then there should be a simple linear curve for accuracy against mobs, but theorycrafting here and most evidence points to a more dramatic curve against higher level mobs. This could also be a situation of skill vs. skill checks, as opposed to a direction level correction function, but the result is the same.

Scenario 3: Crit hit rate - Again, another area where the current trend seems to function off of one set of tests on Greater Colibri, while against higher level mobs the level of critical hits does not follow the assumed linear curve. It would seem this is the most obvious case where a level correction function is occurring, because a simple DEX v. AGI check would even against high level mobs should be resulting in very high crit rates that are not currently being observed. It would seem just like the melee damage formula the crit hit rate formula is based on a level correction first before the DEX v. AGI check is made.

These are all hypothetical explanations of how level correction may exist for these stats (and others like enfeebling and elemental) but may not necessarily be noticed because all of the testing is being done on XP mobs, so there is a dearth of comparable information that we can use to check if level correction is occurring or not. What we need is more parsing of higher level mobs, but the difficulties of that are obvious as there are many more factors that occur when dealing with higher level mobs than a normal XP camp. But parses to gather this kind of information is probably a good place to start.

Are some people giving poor information when they use a simple linear extrapolation for accuracy and crit hit rate? How would we even construct a model to take level correction into account?

tl;dr: If level correction applies to melee damage why doesn't it apply to everything else?

[TummieGaruda]

To be honest, I don't think there is a "level correction" for Critical Hitrate. The thing is, we already know that Accuracy vs evasion has correction to where it is the level difference x 4. We already know that 10 acc over the mobs evasion is around an 80% hitrate, but if you are fighting something 7 levels higher, you will actually need + 38 acc over the mobs evasion to hit 80% of the time. It is probably safe to assume that the reverse is true for your evasion vs mob accuracy, if you are fighting something lower level than you.

It is this level correction that leads me to believe that there isn't one for critical hit rate. Due to level correction on Acc vs Eva, your hitrate is already being scaled down, so your crit rate will depend in part on your hitrate. If you have 75% hitrate and 15% criticals, over the course of 100 melee hits, 75 will land and of those 75, around 12 will be critical hits. It wouldn't be fair in my opinion to level correct criticals because then you would end up with a much lower critical hitrate.

[Spider-Dan]

The thing is, we already know that Accuracy vs evasion has correction to where it is the level difference x 4. We already know that 10 acc over the mobs evasion is around an 80% hitrate, but if you are fighting something 7 levels higher, you will actually need + 38 acc over the mobs evasion to hit 80% of the time. It is probably safe to assume that the reverse is true for your evasion vs mob accuracy, if you are fighting something lower level than you.

I'm fairly certain that this is not true if you don't have Signet. With Signet (in applicable areas), it might be the case, but not against NMs (regardless of level).

[ringthree]

To be honest, I don't think there is a "level correction" for Critical Hitrate. The thing is, we already know that Accuracy vs evasion has correction to where it is the level difference x 4. We already know that 10 acc over the mobs evasion is around an 80% hitrate, but if you are fighting something 7 levels higher, you will actually need + 38 acc over the mobs evasion to hit 80% of the time. It is probably safe to assume that the reverse is true for your evasion vs mob accuracy, if you are fighting something lower level than you.

It is this level correction that leads me to believe that there isn't one for critical hit rate. Due to level correction on Acc vs Eva, your hitrate is already being scaled down, so your crit rate will depend in part on your hitrate. If you have 75% hitrate and 15% criticals, over the course of 100 melee hits, 75 will land and of those 75, around 12 will be critical hits. It wouldn't be fair in my opinion to level correct criticals because then you would end up with a much lower critical hitrate.

1) For accuracy, is that really the level correction, or is that just the accuracy formula applied to higher level mobs? Meaning, is that information from testing over a range of mob levels or is it just /checking for accuracy at different character levels? If it is the latter then it doesn't prove or disprove level correction really.

Does level correction effect /check for low/high defense mobs? Or is it just a straight attack vs. defense comparison? If it is just a straight comparison without considering level correction why would accuracy be any different? I don't know which is the case really, could you provide the source of the information on the accuracy level correction?

2) I understand that it might not necessarily be "fair" to level correct criticals, but it still might be the case. Crit rates can get pretty low on high level mobs.

[Kirschy]

I wasn't a big fan of the accuracy formulas we currently had, but after some of my own testing they do seem to work out in the end. I haven't had a chance to test out the level correction on accuracy, but I'm guessing it'll hold up as true in the end. A difference of 4 accuracy per level is fairly easy to see if you compare tests from mobs 3 levels apart~

[TummieGaruda]

1) For accuracy, is that really the level correction, or is that just the accuracy formula applied to higher level mobs? Meaning, is that information from testing over a range of mob levels or is it just /checking for accuracy at different character levels? If it is the latter then it doesn't prove or disprove level correction really.

Does level correction effect /check for low/high defense mobs? Or is it just a straight attack vs. defense comparison? If it is just a straight comparison without considering level correction why would accuracy be any different? I don't know which is the case really, could you provide the source of the information on the accuracy level correction?

2) I understand that it might not necessarily be "fair" to level correct criticals, but it still might be the case. Crit rates can get pretty low on high level mobs.

Can't use /check function because it does not take level correction into play on anything above EM. If you do fight an EM, the /check function displays the correct information, in which +10 acc over the mobs evasion will make it check Low Eva.

On high level mobs I felt that the low critical hitrate was due to mobs having much higher AGI and Evasion which skews the hitrate and lowers criticals. Should be pretty easy to test though.

[Nameless]

I do experience some very low critical hit rate against high lvl mobs on some very large sample. With a 5% base and 4% from merit, I should always be looking up for at least 9% critical hit rate. Yet, against harder mobs, I see myself at 7-8% alot of time with capped accuracy.

I have drg and drk and both have around 85-90ish dex during TP, no b.pants jobs. From the same parses I usually see b.pants jobs able to achieve 9-11% under the same situation. I am not sure but perhaps you do get negative critical hit rate if your dDex is negative compare to mob's AGI?

This is likely worth finding out as if the penalty is very harsh (I don't believe, even high lvl mobs, will have that much AGI over me, 1-5 at the most. So for me to lose about 1-2 critical hit rate, the penalty has to be something like every -1 AGI is 0.5% or something), and if it is the case then askar helm and c.mantle would both turn out to be pretty sweet to avoid such penalty.

Edit: And obviously if that's the case, decent amount of dex is needed for WS like drakesbane even when you know you have no way to cap it (like you hear alot of ppl say, don't bother with dex, if you don't have enough, don't stack it).

[pchan]

Direct magic damage isn't affected by level difference afaik, only magic accuracy (evasion) is, but that's not clearly tested either. Spell interruption isn't affected either.

[ringthree]

I wasn't a big fan of the accuracy formulas we currently had, but after some of my own testing they do seem to work out in the end. I haven't had a chance to test out the level correction on accuracy, but I'm guessing it'll hold up as true in the end. A difference of 4 accuracy per level is fairly easy to see if you compare tests from mobs 3 levels apart~

The problem I was having is that testing mobs three levels apart might not tell you anything important about level correction if the formula doesn't change over that span.

Can you suggest a good method for testing on higher level mobs other than just parsing different mobs? The problem with parses is that they are just not reliable all the time, and over the course of a parse conditions can change through buffs and recasts.

[TummieGaruda]

I do experience some very low critical hit rate against high lvl mobs on some very large sample. With a 5% base and 4% from merit, I should always be looking up for at least 9% critical hit rate. Yet, against harder mobs, I see myself at 7-8% alot of time with capped accuracy.

I have drg and drk and both have around 85-90ish dex during TP, no b.pants jobs. From the same parses I usually see b.pants jobs able to achieve 9-11% under the same situation. I am not sure but perhaps you do get negative critical hit rate if your dDex is negative compare to mob's AGI?

This is likely worth finding out as if the penalty is very harsh (I don't believe, even high lvl mobs, will have that much AGI over me, 1-5 at the most. So for me to lose about 1-2 critical hit rate, the penalty has to be something like every -1 AGI is 0.5% or something), and if it is the case then askar helm and c.mantle would both turn out to be pretty sweet to avoid such penalty.

Edit: And obviously if that's the case, decent amount of dex is needed for WS like drakesbane even when you know you have no way to cap it (like you hear alot of ppl say, don't bother with dex, if you don't have enough, don't stack it).

Take from this what you want but it does show some interesting information. Took me a while to dig this up:

seaseaseaseasea - January 6th, 2008

[ringthree]

I do experience some very low critical hit rate against high lvl mobs on some very large sample. With a 5% base and 4% from merit, I should always be looking up for at least 9% critical hit rate. Yet, against harder mobs, I see myself at 7-8% alot of time with capped accuracy.

I have drg and drk and both have around 85-90ish dex during TP, no b.pants jobs. From the same parses I usually see b.pants jobs able to achieve 9-11% under the same situation. I am not sure but perhaps you do get negative critical hit rate if your dDex is negative compare to mob's AGI?

This is likely worth finding out as if the penalty is very harsh (I don't believe, even high lvl mobs, will have that much AGI over me, 1-5 at the most. So for me to lose about 1-2 critical hit rate, the penalty has to be something like every -1 AGI is 0.5% or something), and if it is the case then askar helm and c.mantle would both turn out to be pretty sweet to avoid such penalty.

Edit: And obviously if that's the case, decent amount of dex is needed for WS like drakesbane even when you know you have no way to cap it (like you hear alot of ppl say, don't bother with dex, if you don't have enough, don't stack it).

Many times people report lower actual crit rates than what should appear based on their gear bonuses. I wonder if that is because the crit hit rate bonuses are added before DEX v. AGI (or whatever the way crit rate is calculated) is checked.

[ringthree]

Take from this what you want but it does show some interesting information. Took me a while to dig this up:

seaseaseaseasea - January 6th, 2008

Additionally, if you search on BG you will find several threads about this that I have specifically discussed these results including graphs speculating the way that crit hit rate is calculated.

[Kirschy]

The problem I was having is that testing mobs three levels apart might not tell you anything important about level correction if the formula doesn't change over that span.

Can you suggest a good method for testing on higher level mobs other than just parsing different mobs? The problem with parses is that they are just not reliable all the time, and over the course of a parse conditions can change through buffs and recasts.

I probably should have elaborated. I started doing my own accuracy tests with Baha Scythe on WAR/SAM. Using defender, you end up getting around ~450 hits in on a Robber Crab. Since there's no accuracy fluctuations throughout the battle, the parsed hit rate becomes very precise.

I have some free time right now, so I'll design a quick test I can run. Wiki has steelshells at 73-76. Using /checks we should be able find the evasion on the Lv 73 crabs and Lv 76 Crabs. If our current accuracy formulas are correct, ~1k hits on each crab should show us the predicted hit rate% and level correction. (This test assumes that the level correction is applied to lower level mobs as well. If this test does not confirm this, we'll need to repeat the test on Lv76/Lv79 set, or something close.)

As for your earlier question, /check on att/def, acc/eva does not include level correction.

[Nightfyre]

Direct magic damage isn't affected by level difference afaik,

I believe this is true.

only magic accuracy (evasion) is, but that's not clearly tested either.

Don't know about this either way.

Spell interruption isn't affected either.

I'm inclined to disagree. Admittedly my evidence is anecdotal, but it's pretty annoying when a mob 8 levels higher than you breaks Stoneskin and deals merely 16 damage, yet inevitably causes an interruption. On the other hand, a level 70 mob could break Stoneskin for 50 damage and I still have a strong chance of the spell going off. Might just be horrible luck on my part though.

[ringthree]

I probably should have elaborated. I started doing my own accuracy tests with Baha Scythe on WAR/SAM. Using defender, you end up getting around ~450 hits in on a Robber Crab. Since there's no accuracy fluctuations throughout the battle, the parsed hit rate becomes very precise.

I have some free time right now, so I'll design a quick test I can run. Wiki has steelshells at 73-76. Using /checks we should be able find the evasion on the Lv 73 crabs and Lv 76 Crabs. If our current accuracy formulas are correct, ~1k hits on each crab should show us the predicted hit rate% and level correction. (This test assumes that the level correction is applied to lower level mobs as well. If this test does not confirm this, we'll need to repeat the test on Lv76/Lv79 set, or something close.)

As for your earlier question, /check on att/def, acc/eva does not include level correction.

Do you determine the level of the mob post-hoc?

[BerenTebogo]

I've always advocated that 1 evasion skill = 1 evasion+ and in situations where equal on NIN I went for EVA+ not skill+ because of this theory. I never parsed the difference since parsing evasion is retarded but I like the results I got.

[pchan]

I believe this is true.


Don't know about this either way.


I'm inclined to disagree. Admittedly my evidence is anecdotal, but it's pretty annoying when a mob 8 levels higher than you breaks Stoneskin and deals merely 16 damage, yet inevitably causes an interruption. On the other hand, a level 70 mob could break Stoneskin for 50 damage and I still have a strong chance of the spell going off. Might just be horrible luck on my part though.

100% spell interruption gives no interrupt on mobs from any level even level 95 HNMs

[TummieGaruda]

Do you determine the level of the mob post-hoc?

Always best to do that via the exp given so you know what level the mob is.

[ringthree]

Always best to do that via the exp given so you know what level the mob is.

Yep, that is what I figured. I just didn't know it they were guesstimating just from /check.

[Malacite]

Well, from what I've noticed in regards to accuracy there has got be a level correlation because it seems to work both ways;

Even as a THF 75 with my evasion gear on, high lv mobs will hit me much more frequently than I care for. Even low level mobs tbh which to this day boggles my mind a how a lv 50-ish or lower enemy can strike me with any kind of frequency when I have well over 300 evasion but whatever.

On the flipside, Even with horribly under leveled skills (100 or less) I can still hit EP/DC crabs in Kuftal Tunnel just fine at 75 where if I'd been around the usual 50's that people EXP on them I would never land a hit even with sushi. So going on this, I'd say there has to be some form of level correction for accuracy & evasion bonuses/penalties.