Tradeskill success not matching wiki formula

[Wisteso]

According to the wiki, doing a combine of something after being the exact 'trivial' skill amount should succeed 66% of the time, but this was definitely not even close to the success rate I had when doing quite a lot of Heady Kiola combines. Other combines seemed to mirror this as well but I don't have as much data on those.

Over the course of probably 100 combines I observed about a 90-95% success rate at 46 skill, not 66%. I suppose I could have 1 in a million luck, but seems more likely that the wiki is just not correct. I never recall having only 66% chance when trivial in retail either.

Anyone know what's going on here?

[chaos1990]

tradeskills now adjusted to be timeline appropriate and all leveling exorbitantly expensive leaving only those who got in early profitable

[eqravenprince]

I can't speak for the wiki, but I only remember it being 95% success rate once trivial.

[Hoppkins_Wytchfinder]

95% is the correct figure for trivial combines

[loramin]

If the wiki is wrong, fix it! :)

[Jungleberry]

If the wiki is wrong, fix it! :)

[Swish]

If the wiki is wrong, fix it! :)

[ewjax]

Not sure where the 66% value is coming from? The equations I got from the wiki predict 95% success when skill >= trivial.

[larper99]

Not sure where the 66% value is coming from? The equations I got from the wiki predict 95% success when skill >= trivial.

The wiki states:
MIN(SKILL - (.75*trivial) + 51.5, 95) -- this is for things w/ trivial higher than 68.

So, for Skill = 68, trivial = 68,
SKILL-(.75*trivial) + 51.5 = 68 - 51 + 51.5 = 68.5.
MIN(68.5,95) = 68.5 for a skill with a trvial of exactly 68. For skills with higher trivial values, the "trivial" percentage would be higher until it reaches 95.
It would actually take a skill with a trivial value of 174 to have a 95% success rate at the trivial value, according to this formula (x=(95-51.5)/.25 => x=174).

For Heady Kiola, the trivial is 46. Wiki states the formula for skill with trivials less than 68 as:
MIN(Skill - Trivial + 66, 95)

For Skill = 46, trivial = 46,
SKILL-Trivial + 66 would be 0+66 = 66.
MIN(66,95) = 66.

That all being said, I don't believe the given formulas are correct. I get about a 95% success on crafting when above trivial, although I have never analyzed the results systematically.

[Wisteso]

Roughly same conclusion I was having larper99. The formula was saying I should have 66% chance but I was having significantly higher with a fairly high sample size that should have ruled out random luck.

As for updating the wiki, that would assume I know the correct calculation. I don't. I just know that what's on there presently looks to be wrong.

[ewjax]

The wiki states:
MIN(SKILL - (.75*trivial) + 51.5, 95) -- this is for things w/ trivial higher than 68.

So, for Skill = 68, trivial = 68,
SKILL-(.75*trivial) + 51.5 = 68 - 51 + 51.5 = 68.5.
MIN(68.5,95) = 68.5 for a skill with a trvial of exactly 68. For skills with higher trivial values, the "trivial" percentage would be higher until it reaches 95.
It would actually take a skill with a trivial value of 174 to have a 95% success rate at the trivial value, according to this formula (x=(95-51.5)/.25 => x=174).

For Heady Kiola, the trivial is 46. Wiki states the formula for skill with trivials less than 68 as:
MIN(Skill - Trivial + 66, 95)

For Skill = 46, trivial = 46,
SKILL-Trivial + 66 would be 0+66 = 66.
MIN(66,95) = 66.

That all being said, I don't believe the given formulas are correct. I get about a 95% success on crafting when above trivial, although I have never analyzed the results systematically.

Ah thanks! Very interesting. The stuff I have been looking at has triv values greater than 174, which explains why I never noticed this.

I can also state, without a lot of data, that the triv stuff I was crafting at lower triv levels seemed to be pretty reliable, 95% feels right, 65% definitely does not feel right. I don't have actual data to support that though.

Good to know.

[loramin]

Roughly same conclusion I was having larper99. The formula was saying I should have 66% chance but I was having significantly higher with a fairly high sample size that should have ruled out random luck.

As for updating the wiki, that would assume I know the correct calculation. I don't. I just know that what's on there presently looks to be wrong.

People often say this sort of thing about the wiki, but you don't have to know what's right to say that something's wrong. For instance, I added the following to the Tradeskills page based on this thread:

NOTE: This formula is suspect, and may not be correct for Project 1999. See [1].

Now of course, it'd be even better if someone could delete that and the incorrect formula, and put in a correct one, but my point is that even if all you do is add a "this seems wrong" note, you're still moving the wiki in the right direction.

[larper99]

So, finally did a bit more baking this weekend. Took note of my successes.

I was making beer braised wolf (trivial 68) with a tradeskill of 67. My success rate was about 66%.
As soon as I got to level 68, I got 15 successes in a row. So, I jumped from 66% to apparently the 95% level once I reached trivial (although sample size is too small).

Then I did some pickled froglok. Trivial 61. Got 23 successes out of 27 attempts, for an 85% success. Expected (from Wiki) would be 73, so with such a small sample size seems like the formula is correct.

Not enough samples, but research will continue as time and resources permit.

[eqravenprince]

So, finally did a bit more baking this weekend. Took note of my successes.

I was making beer braised wolf (trivial 68) with a tradeskill of 67. My success rate was about 66%.
As soon as I got to level 68, I got 15 successes in a row. So, I jumped from 66% to apparently the 95% level once I reached trivial (although sample size is too small).

Then I did some pickled froglok. Trivial 61. Got 23 successes out of 27 attempts, for an 85% success. Expected (from Wiki) would be 73, so with such a small sample size seems like the formula is correct.

Not enough samples, but research will continue as time and resources permit.

This has also been my experience, once trivial there is a decent jump in success. You'd think it would be more gradual.