Something dosnt seem right with this..... all S84 getting degraded to S80, but what happens to the stats of the weapons?
Example could be, that currently a top s84 staff (which would be S80 in GoD), got 281 m att, while a R85 staff has 270 m att according to korean official.... so wouldnt this mean that S80 would be better than R85?.. :confused:
And to get the Second SA, you need more than a normal SA Crystal, but a Special Crystal only droppable on Boss, and I don't think it's so easy to get one :)
Afaik, the Vesper/Vorpal Armors and S84 Low/Mid/Top Weapons stats didn't changed. The exemple you took for the M.Atk of the R grade staff is right too. Stats are weakier
I have been thinking about this as well, seems considering price, SA Cost, Rbsps cost etc. it has to be a R-grade +4 - 5 atleast to be worth it. Otherwise its better to stick with Vesper or top S84.
Yes, that would help, but as i understood it, the weapon needs to be blessed for that to be avalible..... and thats like 1/50 chance only :/
Weapon doesnt have to be blessed to have 2nd SA.
R weapons have special effects added to them when enchanted +4 or higher. Example: 2-hand Staffs get extra Cast Spd.,
speaking of mage weapons, i'll just remind that oe on mage weaps till now was agreat pain for pockets, which i assume will not change in GoD.
Quite frustrating isn't it! Everyone on the server is running around with duals/dual daggers two-handed swords +7-+16, and guess what? There's like 1 +7 vesper buster on the server! And I spent 4 BEWS to get +4. On PTS I failed 17 Icarus Spirits going+4 and yet people with ALL other weapon types were walking around with beautiful glowing weapons.
mage weapons have lower enchant chance than fighter weapons.
It's not writen anywhere, but yes from our experience mage weaps sux on OE, which means that mage will have to store bills for new update, not only because of new grade enchants and weaps insane prices, but also because we'll need triple or quadriple the funds a warrior or rogue will need for same weap OE
Oh but it is. There were sites that had info from the leaked retail files that confirmed it, and also I'm pretty sure the korean lab that does L2 tests had confirmed a different enchant success rate for fighter v mage wep.
Oh but it is. There were sites that had info from the leaked retail files that confirmed it, and also I'm pretty sure the korean lab that does L2 tests had confirmed a different enchant success rate for fighter v mage wep.
The test was carried out on 4 types of weapons Dynasty Sword, Dynasty Bow , Dynasty Mace and Dynasty Staff.
They enchanted untill each weapon type had FAILED a 100 times
The end result was 68.25% - 70.15% enchant success on the Dynasty Sword and Dynasty Bow.
And 39.39% - 43.82% success rate on the Dynasty Mace and Dynasty Staff.
So there is facts that back up that Mage wep are harder to enchant then melee/bow weapons.
I believe this low rate of enchant on mage weapons comes from the old times when nuker was making twice the dmg he does now and on xp all mobs were 2 hits for him.Wariors had the possibility of oe and maximize their dmg and like this there was a certain balance between them.
But nukers dmg and crit.rate was nerfed over and over and it's been nerfed again on GOD and im curious why korean players dont complain about it!There are nerfs on nukers' powers on every update,but never a fix in their weakness.Koreans dont play nukers at all so they dont care?I imagine if we were korea's forum we wouldnt stop shouting untill they fixed the issues!!!
Then they shouldnt LIE to every single player in L2, by officially writing untrue infomation in game on load screens: "All weapons have a 2/3 chance of OE success".... which means that mage weapons ALSO only SHOULD have 33% chance to blow.... AND 66% to succeed on OE.
While in reality its more like 20% success, if even. :mad:
personally I still think that enchant rates are 66.66666% for all... it's just that because there are sooooo many people enchanting fighter weapons, that 100 successfuly +7 fighter weapons, out of say 10,000 attempts, seem to be much more than 10 successful +7 mage weapons out of 1000 attempts, when in fact it's the same ratio
ofcourse it won't be exactly 66.6666%, there will always be variations in the success rate due to random streaks.
There is also the matter that enchanted mage weapons give better bonuses than fighter weapons. A +7 homu is about the same as a +1 wizards tear (top b), whereas a +7 dark screamer is less than a +0 demon dagger (top b), and a +7 paagrian sword (same grade as darkscreamer) is the same as a +0 great sword (1 below top b)
more fun enchant facts:
+16 homu (2nd top c grade) = +3 acrana mace (bottom s grade)
+16 dark screamer (2nd top c grade) = +2 soul sep (2nd top a grade)
+16 pa'agrian sword (2nd top c grade) = +1 dragon slayer (2nd top a grade)
+16 sword of delusion (2nd top c grade) = +0 Dark Legion Edge (2nd top a grade)
+16 storm bringer * Storm bringer (3rd top c grade duals, but it's 2 more matk than 2nd top of other fighter weapons, so it's actually higher grade already) = +2 dama*dama (2nd top a grade, same as other 2nd top a grade fighter weapons)
personally I still think that enchant rates are 66.66666% for all... it's just that because there are sooooo many people enchanting fighter weapons, that 100 successfuly +7 fighter weapons, out of say 10,000 attempts, seem to be much more than 10 successful +7 mage weapons out of 1000 attempts, when in fact it's the same ratio
No populus they arent.
Anyone who was oe freak can confirm you that.(i had at least 3-4 friends that were blowing weapons everyday,noone ever succeed to oe for me a mage weapon after certain number,the highest oe we ever made was +7 som years ago and they tried to oe later am with no results when draconic bows were +7 with 1 try)Ofc we talk about logical numbers of oe,if someone has unlimited weapons/ews/adena,yes he will some day make +16 mage weapon.
As i wrote before,old developers must kept mage weapons oe rate so low,to give advantage on wariors since back then their power was depending from their weapon,while nukers dmg was more than ok,they didnt need more boost with oed weapon.But noone fixed that when new devs started the changes and if there's noway to get fixed,then they should give nuker's spells power dmg back!
The test was carried out on 4 types of weapons Dynasty Sword, Dynasty Bow , Dynasty Mace and Dynasty Staff.
They enchanted untill each weapon type had FAILED a 100 times
The end result was 68.25% - 70.15% enchant success on the Dynasty Sword and Dynasty Bow.
And 39.39% - 43.82% success rate on the Dynasty Mace and Dynasty Staff.
So there is facts that back up that Mage wep are harder to enchant then melee/bow weapons.
I remember that thread a long time ago...
but overall, the statement that the success rate to OE a weapon is still around 66% when you consider all type of weapons (so their statement is correct), but when you break it up into specific weapons then the chance is different.
According to more than 5.000 enchant attempts aggregating data from
1) L2WH (splitting mage / fighter that was not done initially so it led to wrong results)
2) All Korean Labs testings
3) Hundreds of OE attempts done by me on Freya PTS
The figures are
40% for Mage Weapons
70% for Fighter Weapons
My experiment was posted on EB with images, charts, discussion, etc.
Only the first post is available online atm but
1) Since following posts corrected some issues
2) Since the site "hosting" the post refused to put some reference to author and EB
I really can't give the link. If anyone is interested for the numbers and the tests that present some serious hard evidence for the 40% / 70% rule you can pm me.
I remember that thread a long time ago..
3 threads actually, one with original info and pointing to EB, one with some new numbers and some complains about "stealing" only the first (original EB) post and a third I cannot remember.
On topic, the best I can do atm is to point to a single post that contains most of the numbers crunched. For some numbers and hard evidence see
I have made a long post some time ago at EB. To cut a long story short (though I could just copy / paste most of data again here) and combining data coming from 3 different sources the conclusion is that average enchant rate for fighter weapons is about 70% and for mage weapons about 40%.
The conclusion used data from old l2wh charts (c-grade and up), the korean tests (S80) and my own experiment (S84) at PTS. The % rate was so close for each data set (aka 70% vs 40%) that we could safely aggregate all available data.
A side by side comparison of % success overall ignoring possible breakpoints and / or fluctuations between different types of weapons of the same class
I also had some plots to actually show the chaotic difference with 70% vs 40% success rate. Again to cut a long story short, in theory, with 70% success rate you would need to break 100 weapons to get to +16 while with 40% you would need to break 150.000 weapons to get to +16. That happens because (0,7)x(0,7)x(0,7)x... compared to (0,4)x(0,4)x(0,4)x... start diverging very fast and keep doing even faster while consecutive OE successes happen.
If you add into equation that the +15->+16 part has 1/3 chance (flat for all weapon types, not 70% vs 40%) as shown by korean labs then the actuall numbers for fighters would be breaking about 200 weapons to get at least one safely to +16
The actuall number for mage would also increase a bit (33% less than 40%) and should be more than 150.000 but i don't think it matters anyway.
For comments, discussion etc, etc there is always search to locate the appropriate threads (I said the reasons why i don't post any more links to info)
Theoreticaly speaking, we can see above 3 experiments which draw conclusions completely leaving outside parameters such as variation and a-priori probabilities.
When we start an experiment of sucess/fail we have 2 ways to proceed:
either we take a safe number of weapons +3 and start enchant to +4, and taking in mind variation of sample and CV of sample we draw the sucess/fail probability (bayes)
or we take in mind the variation of enchanting distribution and a-priori probabilities to draw a conclusion on random consecutive enchants
more fun enchant facts:
+16 homu (2nd top c grade) = +3 acrana mace (bottom s grade)
+16 dark screamer (2nd top c grade) = +2 soul sep (2nd top a grade)
+16 pa'agrian sword (2nd top c grade) = +1 dragon slayer (2nd top a grade)
+16 sword of delusion (2nd top c grade) = +0 Dark Legion Edge (2nd top a grade)
+16 storm bringer * Storm bringer (3rd top c grade duals, but it's 2 more matk than 2nd top of other fighter weapons, so it's actually higher grade already) = +2 dama*dama (2nd top a grade, same as other 2nd top a grade fighter weapons)
with a +16 Homonkulus you make less damages than with a +3 arcana mace because of added elements (and PVP stuff is you PVP)
with a +16 Homonkulus you make less damages than with a +3 arcana mace because of added elements (and PVP stuff is you PVP)
I'm just comparing the m/patk increases for each weapon-type when OEed, mage weapons are better off in that respect
Theoreticaly speaking, we can see above 3 experiments which draw conclusions completely leaving outside parameters such as variation and a-priori probabilities.
When we start an experiment of sucess/fail we have 2 ways to proceed:
either we take a safe number of weapons +3 and start enchant to +4, and taking in mind variation of sample and CV of sample we draw the sucess/fail probability (bayes)
or we take in mind the variation of enchanting distribution and a-priori probabilities to draw a conclusion on random consecutive enchants
If you exclude l2wh numbers the other 2 experiments had that logic embedded. A safe number of weapons were made +3 and then +4 and then +5 and so on.
At all steps no significant statistical variation was observed except +15->+16 which was shown to be 33% for all type of weapons.
Seeing a constant 40% vs 70% at all experiments and at all OE stages with a fairly large number of samples (to prove a 40% or 70% chance) can lead to a safe conclusion.
Your observations are right (though confusing to most of forum members) in general. But some thousand of tries to prove a 70% are more than enough. And when different experiments, at different OE levels, at different game versions all seem to end up to the same 40% vs 70% .. There is also one more generic rule .. Ochams razor ..
I really have no reason to believe that there is a complexity / game mechanic / statistical "abnormality" that we missed .. Simply because there would be no reason to implement it ..
Case 1) A random number generator for 40% - 70%. Some hundreds of tries would be more than enough to prove it and they really do
Case 2) Some weird distribution depending on (something?) that would generate different rates (which we cannot observe ?) because (reason?)
LF confirmation. In patch notes thread, they wrote that only Vesper equipment will be degraded to S80, so I understand that Vorpal/Elegia will be S84 as at the moment.
LF confirmation. In patch notes thread, they wrote that only Vesper equipment will be degraded to S80, so I understand that Vorpal/Elegia will be S84 as at the moment.
all current s84 gear will be s80 in GoD... as laby says use search and you would have known this already..
Will there be EWR and EAR and what will be their bonuses?
After all, for weapons, D grade was 2/4. C, B, and A grade were 3/6. S, S-80, and S-84 were 4/8. Where the first number is safe enchant pluses to matk and the second number are over-enchant pluses to matk. Matk chosen as example.
Will EAR finally break out of the 1/3 mold to which all previous armors hold?
Theoreticaly +4 has different sucess probability than +5
Indeed, I did not state it correctly.
+3 -> +4
+4 -> +5
+5 -> +6
were examined individually as different enchant cases. That was the % rate calculated first, then the overall rate was calculated by using ALL enchant steps as equal, not the other way round.
i.e. I did not suppose that the chance is 70% because I saw an overall rate of 49% from +3 -> +5
I observed ~70% / ~40% for +3 -> +4
I observed ~70% / ~40% for +4 -> +5
.
.
.
I observed ~70% / ~40% for +14 -> +15 (much less weapons to test with)
Korean Labs did also that, they had individual numbers for each OE step
l2wh also had individual steps (though wrongly grouped I fixed that)
1) 3 different sources calculating ~70% for all intermediate steps
2) 3 different Patch Levels (C1->C4 NA from l2wh & Freya / Hi5 Korea from Labs & Freya PTS NA from me)
3) Averaging the same OE step from different sources so close to the number calculated IF all steps were considered equal
Led to the hypothesis that all steps are the same. Not the other way around. I did not take the overall number of success vs fail and assumed the same %rate for same step to end up to a number. I calculated the % rate of EACH step FIRST. Then I calculated the overall % rate. Then I saw that it would be pretty accurate to consider each OE step same with any other since the figure were so close to that.
I can't say that it is theoretically true, there is still a chance of some game mechanic dictating +3->+4 72% and +4 -> +5 68%, but there is not an experiment that could be executed on real servers to figure out that, mainly due to physical constrains.
The point of the experiment was figure out vast differences making +16 fighter vs + 16 Mage. Calculating an average of 70% vs 40% with 5.00 individual OE attempts was more than enough.
To be honest, at start, my only point was to calculate an approximation of weapons needed just to find out how _harder_ it would be for +16 Mage weapon, starting out from the fact that in Freya PTS I saw so many +14 / +15 / +16 melee and i did not saw a single Archangel / Sacredium.
Doing some fast calculations at start I discovered that a +16 melee would need about 150 weapons while a +16 mage would need about 150.000 weapons.
That triggered a discussion in EB.net and then I started using numbers from Korean Labs and l2wh (after fixing the hypothesis of same OE rate) trying to produce some really accurate numbers. What I copy/paste/quote is the conclusion.
I did not copy / paste the methodology, so you are correct 100% on what you say. Even with the methodology explanation there still is some room for discussion about distributions / randomness / etc.
The fact (and what I was trying to prove) is that a player will observe a success rate of about 70% when trying to OE a melee weapon and of about 40% when trying to OE a mage weapon. The second point was to show people that the 70% and 40% have a big difference even at lower OE levels, a huge difference at mid OE levels and a vast (by 3 orders of magnitude) difference for high OE levels
no significant statistical V is always a matter of discuss :D
indeed ;)
case 1) it depends on how deep on statistical research nc soft went when created l2
I try to keep my thought for me regarding that :)
case 2) observation experiments are an important tool for statistics but still there are a lot more. If interested check pms
You are 100% correct again, i read the pm.
But, does it really matter for casual players that that try to OE a weapon ?
Better, is there enough reasoning to keep the 70% vs 40% info contained and letting lot's of people to believe that OE melee is the same with OE mage ? Ending up to people that were blowing hundreds of millions for no reason ? And some of them still not convinced for the difference ? I have witnessed people quitting the game for months after blowing up everything trying to get an Arcana Mace +6 trying to compete with buddies that already had +9/+10 Draconic Bows.
There always were 3 theories about that fact
1) Mages are "chickens", they are not "brave" enough to OE, they are afraid
2) Mage weapons are less, that's why we can not observe highly OE'd weapons
3) Success rates are different, go OE a DS and exchange it with Homu
Since Game Client states (starting up tip) that there is a 1/3 chance of failing backed up by widely (back then) accepted (though wrong) analysis of l2wh the ONLY possible explanation for an educated player believing NCSOFT would be number 2. Which eventually would lead up to the dangerous hypothesis number 1, that effectively destroyed lot's of weapons / players until some popular wisdom / word of mouth led only the "burned" and some smart Giran Market players turn to number 3
So, though you being correct in theory, that there is still a chance for some statistical errors coming from interpretation of numbers, that would make no difference to any player .. Maybe only to some (legit ?) that made a fortune (IG?) being in OE (true?) business in a "grey" area..
To sum up, though nobody can prove 100% unless he sees the actual code, the 70% / 40% "rule" is a huge improvement compared to uniform 66,6% / 66,6%.
The main issue here is that
If you believe NCSOFT you need about 200 weapons to make a +16 ANY weapon with normal scrolls
If you believe experiments you need about 200.000 weapons to make a +16 mage weapon with normal scrolls
Actual experiments methodologies / hypotheses may lead to some errors.
Those errors can be safely ignored, at start, in any science, when you fix a 3 orders of magnitude error.
I believe that you will agree with the following ..
Some errors that MAY exist IF there is something more complicated than a simple random number generator are nothing compared to the 3 orders of magnitude between expected vs observed. And that the number of tests, regardless to the exact methodology used rules out any kind of THAT great error.
To sum it up ..
The experiments / methodologies / sample / number of tries / distribution MAY lead experiencing 1%-3% variations IF some conditions are met that do not change the general picture / game experience
The officially announced and widely accepted number allows clearly / instantly experiencing 55% - 59% variations that dramatically change the general picture / game experience