GM smithing/tailoring success rates w/ Mastery 3, possible bug

[Xulan]

EDIT START
1/30/2006: It is my and others' conclusion that there is no bug and that the skill-caps are set and working, and there is no mystery bug to be found.

Here is my final success data (I stopped keeping track).

499 total combines worth of data.

Symbols + Armor
bp/leg: 114/180 = 63.33% success
sleeve: 64/89 = 71.91% success
boot/helm/glove: 127/155 = 81.94% success
wrist: 66/75 = 88% success

As for why the caps exist, see:
http://mboards.eqtraders.com/eq/show...1&postcount=36

EDIT END



The purpose of this thread is data collection to attempt to find 2 things:

1. The success rate of GM armor (smithed and tailored).
2. If there is a possible bug with 300+ skill and mastery 3. Perhaps an anomoly more than a bug (due to the combination of mastery 3, high skill, and a negative success modifier on the armor)

In the smithing forums, we have a long thread about gm armor success rates: http://mboards.eqtraders.com/eq/show...1&page=1&pp=25. There are 2 good sets of data, ~100ish attempts.

GM armor and symbols, with 300+ (modified) smithing skill and smithing mastery 3:

Set 1:
- 386 armor: 62%
- all armor: 63%
- 386 symbols: 60%
- all symbols: 65%
- all gm combines: 64% (67/105)

Set 2:
- 386 armor: 52%
- all armor: 53%
- 386 symbols: 63%
- all symbols: 65%
- all gm combines: 60% (59/99)

And a 3rd data set with no break-out, 250 to 266 smithing skill with mastery 0 to mastery 3 over the course (from http://mboards.eqtraders.com/eq/show....php?t=22887):

- 76/124 total combines: 61%

This person also reports:

"A member of my guild (270+ skill/mastery 3) did a few tailoring combines for me the other night. He informed me at the beginning of the session that he was going on 35 consecutive gm combines with no failures. He promptly failed three times in a row, and in all went 4 of 9 on combines. He was highly distressed at the end of that session."


So, we have 2 data sets of 300+ and mastery 3, and 2 sets of data sub 300 with mastery 3.

I don't have a problem if the success rates are in the low 60%'s. That actually seems pretty fair for GM armor.

What has stuck out in the first thread are more than a few blips of people with either sub 300 skill & mastery 3 or 300ish skill and no mastery 3 (2, 1 or 0) who have higher success rates (75%+) on some runs (usually runs of 20 or under). It's the amount of posts of those kinds of runs that makes one look twice. Even another friend of mine who has 260ish skill with Mastery 3, who has a good deal of bp and leg combines has 70%+ success.

Please post your data in the following manner so that it makes it easier to tally (can cut and paste):

Skill level: real / modified
mastery:
bp/leg armor:
all armor:
bp/leg symbols:
all symbols:
total:

For attempts, post as:
bp/leg armor: success/attempts = % success
ex: bp/leg armor: 5/10 = 50%

If you have data with different mastery levels or for under/over 300 modified skill, please use a seperate copy & paste of the above for each set. I'll go first as an example.

There is no way to know the exact success rate without either lots of data or knowing the actual code.

[Xulan]

Skill level: 297-300 / 311-315 modified
mastery: 3
bp/leg armor: 16/31 = 52%
all armor: 24/45 = 53%
bp/leg symbols: 21/33 = 64%
all symbols: 35/54 = 65%
total: 59/99 = 60%

[Baltazor Goldsinger]

This is what I have found on Fennin Ro. First of all, I am at the GM (300) level in Smithing and have a 15% modifier which put me at 345 equivalent skill. I also have AA Smithing Mastery level 3 and Salvage 3. Now for results...

Recently made a run making black gold inlays and doing GM symbols. This is what I have found:

Run 1 of making Augments:
10 attempts at Black Gold Inlays: 6 successes & 4 failures!

And this is for something that is a TRIVIAL combine. I was actually MUCH better at this before I got to 3oo skill level. I use to be 80% successful at making Black Gold Inlays.

Attempted 8 symbols: 6 failures and 2 successes.

In my view this is just wrong, it must be bugged or not working properly. I do not intend to make further combines until something is done to fix it, it's just too expensive.

[Ishwar]

Im a m software developper, a long time tradeskiller, and i am vell versed in statistics. Just some advice on how you should try to narrow these kind of bugs.

386 triv cobines are hard to get good data sets on. So unless you believe the bug is only for the very high end combines, you really shoudl try to think up better experimants to test theory.

I have heard problems with mastery 3, for any combines. So try to establish of this is the case. A good place to start is to ask anyone with 300+ skill and masterry 3 to make a few hundres metal bits, (or woven mandrake or whatever are easy to do combines).

Another source to look at is trying to colelct skillup data from anyone with mastery 3, and see if there are any strange jumps in succes rates, and if the succes rates are at the expected levles.

Is you conclude that the bug is only visible on high end combines, you might want to consider jewelcrafing augments as well, since that is a good source of data.

[Bupper]

Is you conclude that the bug is only visible on high end combines, you might want to consider jewelcrafing augments as well, since that is a good source of data.

Unfortunately, SOE developers seem to have capped success rates for radiant cut augments. See this JCM broken again? thread. The conclusion here is that there is no benefit from JCM AAs for augment radiant cuts when modified JC skill is near 300.

[Phantron]

I have been making GM bp/legs from about 250+15% mastery 3 to 280+15% mostly on just GM stuff, and I have seen no compelling reason to believe that the success rate was ever lowered by having a higher skill.

That said it does not appear success rate is ever higher for having higher skill, either, presumably due to a high inherent fail rate on GM armor.

[Xulan]

Unfortunately, SOE developers seem to have capped success rates for radiant cut augments. See this JCM broken again? thread. The conclusion here is that there is no benefit from JCM AAs for augment radiant cuts when modified JC skill is near 300.

That may be very well what we are seeing with the armor. Just trying to figure it out. The only odd part is that the GM bp/legs are about the highest trivial I know, one would think that higher skill or M3 would have a positive effect, but there doesn't seem to be any or there is very little benefit of higher skill (over 250) or having mastery 3 at all, with the posibility of a negative effect on the high end with mastery 3.

[Xulan]

I have been making GM bp/legs from about 250+15% mastery 3 to 280+15% mostly on just GM stuff, and I have seen no compelling reason to believe that the success rate was ever lowered by having a higher skill.

That said it does not appear success rate is ever higher for having higher skill, either, presumably due to a high inherent fail rate on GM armor.

Post your data if you have it please =)

[Bumkus]

2 for 2 on GM tailoring. (Is all i've been able to attempt.)

Unmodified skill: 256
Modifier: GM needle - 5% (Surely the "type" of modifier item wouldn't make a diff.)
Modified skill: 268
Tailoring 2

GM Bloodpact Gloves - Triv 334
GM Glove symbol of the Warmonger - Triv 334

Per calculator, I should have a 76.75% chance to succeed. Was pretty stunned to succeed 2 in a row.

[Baltazor Goldsinger]

I don't know if you missed my point above or were confused by what I meant as a trivial combine... but let me be clearer. I have made a lot of black gold inlays over the past months since OOW came out (more than 50). Most of these were made at a skill level of around 267 (with modifier). For the most part I was satisfied with the success rate - it seemed appropriate for the level etc.

Now I am at a skill level of 345 (with Hammer of the Ironfrost) and (I hate admitting this) but I feel like I am actually a worse blacksmith at making this item which is trivial at 327. Is it unrealistic for me to expect to fail at a rate of 10% or less on an item this is 327 when I am 18 skill points higher?

I have been tradeskilling since the game came out. I do not think this is a case of "oh well too sad, the RNG got you!". Something is just wrong imo. I sort of suspect that someone - and please this isn't meant to be a flame of any sort - but sheesh it HAS happened before with *cough* tailoring *cough* - has either done it purposefully to neutralize some high end tradeskiller's advantages, being it skill modifiers or AA Mastery OR OMG... someone slipped a bit of code in and it went live when they really meant to test it first.

If it wasn't for the fact that I am seeing this happen (smilar results) with my shaman friend making augments (who is 345 too) and some others I wouldn't be as suspicious.

The frustrating part of this is its not like you can really test this throughly like you could if you were making metal bits. These high end combines are just too rare and costly. The developers need to come forward and tell us the community if this is a change that is part of "revamping" tradeskills or if its something unintended and may be the result of changing the skill up code.

[Bobaten]

Baltazor, I think you missed the other's points. The augment combines are success capped at 70%. What this means, is after your base success chance is figured, after your mastery AA is applied, rather than the usual 95% success cap kicking in, a 70% success chance kicks in. Yes, this does mean your Hammer, and your Smithing Mastery are useless (for most current combines). This isn't a big super secret that no one knows about, it has been known since shortly after the skill caps went up.

On the other hand, I have also read many posts here by people with a lower success rate for GM smithed armor at 300 skill than they had skilling up. I can't speak to that issue, because my experience is with augment combines (JC and alchemy), not smithing.

[Mannwin Woobie]

with the posibility of a negative effect on the high end with mastery 3.

And this is what I am most concerned about. People who were posting good success rates at GM stuff seemed to be the ones with high skill no Mastery aa's. But High skill with Mastery aa's seemed to be reporting lower success rates.

[Aeght]

Actually having re-read through the JCM thread, the only comments in there are that the combines are believed to be capped. There are no references to anything from the Dev's confirming that.

In other words, it is possible that the combines are bugged, but the bugs are being accepted on the basis that some feel the bugs are rational.

Regardless of whether the observations are a bug or a feature though, it does not change the fact that JCM is rendered relatively useless by such a 'feature.' And yes, I know there is a strong faction that feels that JCM should be either open to all or if not, useless, but if that is really the intent, they should be open about it and not hide it deep in formulae that we are not given.

Aeght

[Ishwar]

It is very easy to fool yourself on success rates. Humans tend to notice things when they are different than what they expect. I am guessing that the people with max skill and AAs simply notice more failues because they are not expecting them. Statisticaly, if the succes chance is the same with at 280 skill and 300, some people will have worse luck at the high skill. These people will percieve a problem, that is just pure chance.

As for the balck gold inlays, they might also have a low max success cap. And combined with bad luck, failure streaks can happen.

[Phantron]

I've seen absolutely no personal evidence or posted information that suggests people with low skill or no mastery was making GM armor/Black Inlays more successfully than others. There are plenty of information that suggests skill/mastery do not help but there is no evidence to suggest lack of those helps, either.

Show me some logs with low skill/no mastery that manages to succeed GM bp/legs more than 75% of the time reliably in significant numbers and I'll believe there's a problem.