About the success/fail formula:
It appears that the success/fail formulae posted by Ishwar (and available in the FAQ's on EQTraders) is indeed correct, including the underlying difficulty. Since it was first was published, a few new twists were introduced, including reduced fail rates on deeply trivial combines and mastery AA that can also reduce fail rates. Here is the detailed success/fail formula. I will be using Ishwar's notation for most of this.
First, the base chance to succeed is determined based on a "difficulty" factor -- the trivial we see in the game is derived from this number. The base chance to succeed is as we have known it:
Next, if you have a mastery AA, your reduction in the chance to fail is calculated as follows:
The minimum chance to succeed is 5%.
Your final chance to succeed is as follows:
A person with zero skill can attempt any tradeskill combine in the game and have at least a 5% chance to succeed.
A person with zero skill and the appropriate crafting mastery 3 will have a 50% chance to succeed.
About the skillup formula:
(This is a direct copy of the handout Tanker passed out at the Fan Faire, with a few corrections of typos. Tanker refused to email me a copy of his document, so there is a risk of transcription errors.)
Background
In order to explain how the formula was changed in as simple a fashion as possible, it's necessary to describe the formula as it sat before the changes made this Spring. The tradeskill formula has 2 parts, or Passes. Every time a tradeskill is attempted that can possibly give you a skillup, i.e., is not trivial, etc.), the formulae are run, and the player must succeed both Passes in order to gain a skillup.
Old Formula, Pass 1
Pass 1 of the skillup formula is as follows. Note that the chance to pass is a value between 0 and 100.
Y = skillup difficulty of the current skill (for tradeskills, this is either 2, 3, or 4 currently). (Kyros note: see below for more on this factor.)
S = Player stat used for this tradeskill. For most skills, it is the higher of your WIS or INT, minus 15. For Smithing, it's the higher of your WIS / INT / STR (no minus 15). For Fletching and Make Poison, it's the higher of your WIS / INT / DEX (no minus 15).
Some items of note about this formula
Pass 2 of the skillup formula was as follows. Note again that the chance to pass is a value between 0 and 100:
%-chance to pass =
Some items of note about this formula:
In order to alleviate some concerns the designers had over perceived sharp decreases in skillup chance at around 160-190 skill, and the lack of change to that chance over 190 skill, the following change was made. Note that there was no indented change to Pass 1, as the change is meant to affect the player at different skill levels and across all tradeskills, not at different stat levels or per tradeskill. (Footnote: As you may know, there was a bug the first patch these changes were included which did change the Pass 1 chance of success. This bug was fixed in a subsequent patch.)
New Formula for Pass 2
The new formula for Pass 2 is as follows:
%-chance to succeed at Pass 2 =
As stated in several posts online during the last few months, there were a few expected consequences of the formula change:
Skillup chance graph
The following graph displays the chances of succeeding at Pass 2 of the skillup formula, based on your current skill.
This chart shows that the chance to succeed at Pass 2 is unchanged at skills under 175.
This chart shows that the new gradual reduction in the chance to succeed at Pass 2 under the new formula at skills over 175. Note the new lower cap to the chance to succeed at 2.5%, and that the chance to succeed doesn't get down to old-formula-post-190 levels until skill hits 269 or so.
Now, one of the main points we got from tradeskiller feedback over the years is that you don't pay particular attention to the "chance to succeed" as a statistic in itself. You all seem much more interested in AC/S (Average Combines per Skillup), and it's on the following two graphs that you can see the dramatic changes to the AC/S progression using the new formula.
This graph again shows the fact that the formula did not change when skill is under 175. Note the assumption that Pass 1 chance is 100%; if your Pass 1 chance is 50%, double these values. This is why we refer to the changes as percentage changes to AC/S, since we can't tell what the actual change will be in AC/S without knowing your Pass 1 chance.
Here we finally get to the most important graph. Note that, under the Old Formula, the AC/S would increase from 5 at skill 160 to 20 at skill 190, then stay at 20 for the next 110 skill points. Removing this dramatic increase and stagnancy in the middle of the skill progression was a goal of the new formula. An interesting note about this graph is that you can take the area under the different lines from the left edge to a vertical line at whatever skill you're interested in and compute the "total number of combines it should take me to get to that skill (on average)". This was how we computed the 15% and 20% numbers stated above. Assuming a 100% chance of success on Pass 1 (which is unrealistic, but can be used to gauge relative change in the formula), the total area under the Old Formula graph line from skill 0-299 (since you can't skill up at 300) is 2789, under the New Formula line it's 2407 (roughly 86% of 2789). The total area under the Old Formula graph line from skill 250-299 is 1000, under the New Formula line it's 1182 (roughly 118% of 1000).
General Information
Tradeskill difficulties:
The following difficulties were announced by Tanker:
Alchemy: 4
Fletching: 4
Poison Making: 2
Smithing: 2
Tailoring: 2
Tinkering: 2
Note that they mesh well with the values derived by Suani, which leads me to think the rest are probably accurate as well. Begging is now the hardest skill at 28.
Miscellaneous:
The order of success/fail and skillup messages will be reviewed, and if it's feasible, it will be changed to be more consistent. They will also try to incorporate the name of the item you are making into success/fail messages.
Current trophies will not be changed to have a higher skill modifier. The devs consider the stats on trophies to be enough of an upgrade over geerloks.
Loot All buttons on corpses are unlikely; the time it takes to loot a corpse is a game balance consideration.
The devs are not insensitive to the amount of bank space required by DoN cultural armors. They will consider ways to reduce bank space, and they are also considering increasing stack sizes on armor molds and augment patterns to 200.
As previously mentioned, DoN armors will get a review for viability, including upgrades to some slots, if appropriate. In addition, the devs will try to figure out why legs don't get a Type 12 augment.
A tradeskill super-trophy will not be coming soon, but it is still something the developers are considering.
Bag tokens will probably show bag stats with the June 22 patch.
Velium weapon to velium ore conversion are almost done. They will be going to QA next week, and if they pass, they will be included in an upcoming patch.
The Grand Robe / Grand Tunic / Phylactery quest will be reviewed. Making changes to this quest is tricky, though, as all changes have to be approved by Swords and Sorcery, the publishers of the pen-and-paper EQ RPG.
Lubricous Metal Bits are not a real item, and the recipe for them will be removed.
Karana's Heartfruit Salad will be live in the game with the next patch.
Recipes for Taelosian wheat and mountain wheat are also coming pretty soon. My impression was that this would be about two patches out, but I'm not positive. The recipes will be largely based on the player-submitted ideas on the EQLive boards.
The Ice Cream Churn will get a token with the next patch.
Absor said he would send me additional responses to some questions by email, so I'll post those as I get them.
It appears that the success/fail formulae posted by Ishwar (and available in the FAQ's on EQTraders) is indeed correct, including the underlying difficulty. Since it was first was published, a few new twists were introduced, including reduced fail rates on deeply trivial combines and mastery AA that can also reduce fail rates. Here is the detailed success/fail formula. I will be using Ishwar's notation for most of this.
First, the base chance to succeed is determined based on a "difficulty" factor -- the trivial we see in the game is derived from this number. The base chance to succeed is as we have known it:
Success* = Modified Skill + 50 - Difficulty
The modified skill takes into account skill-modifying items like geerloks or trophies, and is always rounded down. This is the number you see in the new UI when you have a geerlok or similar item equipped. Success* is capped at 0 and 100 at the low and high end, respectively.Next, if you have a mastery AA, your reduction in the chance to fail is calculated as follows:
FailReduction = (100 - Success*) * Mastery modifier as a percent
Your Mastery modifier is 0 if you have no AA, 10% if you have one rank of the mastery AA, 25% for the second rank, and 50% for the third rank. This is added to your base success rate: SuccessWithMastey = Success* + FailReduction
Next, your maximum chance to succeed is calculated. This acts as the upper cap for the SuccessWithMastery. The cap is 95% plus 1% for every 40 points that your skill is above the trivial. If you are 39 points over the trivial, you get no benefits. If you are 40 points over, the cap becomes 96%. 80 points is 97%, and so on, up to 200 points or more giving a 100% maximum chance to succeed. This only looks at your raw skill, not your skill as modified by a geerlok or other item. This is your Max Cap. Some recipes can override this cap with a minimum chance to fail. It is not known whether any recipes currently use this.The minimum chance to succeed is 5%.
Your final chance to succeed is as follows:
5% <= SuccessWithMastey <= Max Cap
Interesting notes:A person with zero skill can attempt any tradeskill combine in the game and have at least a 5% chance to succeed.
A person with zero skill and the appropriate crafting mastery 3 will have a 50% chance to succeed.
About the skillup formula:
(This is a direct copy of the handout Tanker passed out at the Fan Faire, with a few corrections of typos. Tanker refused to email me a copy of his document, so there is a risk of transcription errors.)
Background
In order to explain how the formula was changed in as simple a fashion as possible, it's necessary to describe the formula as it sat before the changes made this Spring. The tradeskill formula has 2 parts, or Passes. Every time a tradeskill is attempted that can possibly give you a skillup, i.e., is not trivial, etc.), the formulae are run, and the player must succeed both Passes in order to gain a skillup.
Old Formula, Pass 1
Pass 1 of the skillup formula is as follows. Note that the chance to pass is a value between 0 and 100.
%-chance to pass = ( (S * 10) / (Y * D) ) / 10.0
D = 1 if the tradeskill attempt was successful at making the primary result, 2 otherwise.Y = skillup difficulty of the current skill (for tradeskills, this is either 2, 3, or 4 currently). (Kyros note: see below for more on this factor.)
S = Player stat used for this tradeskill. For most skills, it is the higher of your WIS or INT, minus 15. For Smithing, it's the higher of your WIS / INT / STR (no minus 15). For Fletching and Make Poison, it's the higher of your WIS / INT / DEX (no minus 15).
Some items of note about this formula
- Higher WIS / INT / Possible tertiary = easier Pass 1 success, up to a point.
- Your chance to succeed at Pass 1 is up to twice as good on a successful recipe combine than on a failure.
- Certain skills are "harder to skill up in" than others, in that they have a higher chance to succeed at Pass 1 (via a lower Y value), given the same S and D values. (Kyros note: I think Tanker meant "higher" Y values, since the higher Y gets, the harder it is to skill up.)
Pass 2 of the skillup formula was as follows. Note again that the chance to pass is a value between 0 and 100:
%-chance to pass =
For current raw skill <= 15, chance = 100
Otherwise, chance = (200 - K) / 2.0 (note the floating point divide, see glossary)
K = current raw skill, capped at 5 and 190. 5 <= K <= 190 is always true. (Kyros note: Tanker did not include a glossary, but a floating point divide means you keep the decimal part of the result. An integer divide (where you divide by a whole number) discards the decimal part. This also applies to the division in the Pass 1 formula.)Otherwise, chance = (200 - K) / 2.0 (note the floating point divide, see glossary)
Some items of note about this formula:
- Once a player's skill reached 190, the chance to succeed at Pass 2 sat at 5% forever.
- From skill = 16 to skill = 190, the chance to succeed at Pass 2 moves linearly from 92% to 5%.
In order to alleviate some concerns the designers had over perceived sharp decreases in skillup chance at around 160-190 skill, and the lack of change to that chance over 190 skill, the following change was made. Note that there was no indented change to Pass 1, as the change is meant to affect the player at different skill levels and across all tradeskills, not at different stat levels or per tradeskill. (Footnote: As you may know, there was a bug the first patch these changes were included which did change the Pass 1 chance of success. This bug was fixed in a subsequent patch.)
New Formula for Pass 2
The new formula for Pass 2 is as follows:
%-chance to succeed at Pass 2 =
For raw skill <= 15, chance = 100
For 15 < raw skill <= 175, chance = (200 - K) / 2.0
Otherwise (raw skill > 175), chance = 12.5 - ( (10 / 125.0) * (raw skill - 175) )
Some things to note about this new formula:For 15 < raw skill <= 175, chance = (200 - K) / 2.0
Otherwise (raw skill > 175), chance = 12.5 - ( (10 / 125.0) * (raw skill - 175) )
- There is no change to the formula unless your raw skill is > 175
- At 175, your chance to succeed at Pass 2 falls linearly from 12.5% to 2.5% at 300.
As stated in several posts online during the last few months, there were a few expected consequences of the formula change:
- Skillup chance would no longer flatline at 190 skill, providing a more gradual decrease in skillup chance all the way through 300 skill.
- Average Combines per Skillup (AC/S) would no longer dramatically increase in the middle of tradeskill progression (around 160-190). Instead, it would increase more gently all the way up to 300.
- The Average number of combines necessary to skill up from 1 to 00 would decrease by about 15%.
- The Average number of combines necessary to skill up from 250 to 300 would increase by about 20%.
Skillup chance graph
The following graph displays the chances of succeeding at Pass 2 of the skillup formula, based on your current skill.
This chart shows that the chance to succeed at Pass 2 is unchanged at skills under 175.
This chart shows that the new gradual reduction in the chance to succeed at Pass 2 under the new formula at skills over 175. Note the new lower cap to the chance to succeed at 2.5%, and that the chance to succeed doesn't get down to old-formula-post-190 levels until skill hits 269 or so.
Now, one of the main points we got from tradeskiller feedback over the years is that you don't pay particular attention to the "chance to succeed" as a statistic in itself. You all seem much more interested in AC/S (Average Combines per Skillup), and it's on the following two graphs that you can see the dramatic changes to the AC/S progression using the new formula.
This graph again shows the fact that the formula did not change when skill is under 175. Note the assumption that Pass 1 chance is 100%; if your Pass 1 chance is 50%, double these values. This is why we refer to the changes as percentage changes to AC/S, since we can't tell what the actual change will be in AC/S without knowing your Pass 1 chance.
Here we finally get to the most important graph. Note that, under the Old Formula, the AC/S would increase from 5 at skill 160 to 20 at skill 190, then stay at 20 for the next 110 skill points. Removing this dramatic increase and stagnancy in the middle of the skill progression was a goal of the new formula. An interesting note about this graph is that you can take the area under the different lines from the left edge to a vertical line at whatever skill you're interested in and compute the "total number of combines it should take me to get to that skill (on average)". This was how we computed the 15% and 20% numbers stated above. Assuming a 100% chance of success on Pass 1 (which is unrealistic, but can be used to gauge relative change in the formula), the total area under the Old Formula graph line from skill 0-299 (since you can't skill up at 300) is 2789, under the New Formula line it's 2407 (roughly 86% of 2789). The total area under the Old Formula graph line from skill 250-299 is 1000, under the New Formula line it's 1182 (roughly 118% of 1000).
General Information
Tradeskill difficulties:
The following difficulties were announced by Tanker:
Alchemy: 4
Fletching: 4
Poison Making: 2
Smithing: 2
Tailoring: 2
Tinkering: 2
Note that they mesh well with the values derived by Suani, which leads me to think the rest are probably accurate as well. Begging is now the hardest skill at 28.
Miscellaneous:
The order of success/fail and skillup messages will be reviewed, and if it's feasible, it will be changed to be more consistent. They will also try to incorporate the name of the item you are making into success/fail messages.
Current trophies will not be changed to have a higher skill modifier. The devs consider the stats on trophies to be enough of an upgrade over geerloks.
Loot All buttons on corpses are unlikely; the time it takes to loot a corpse is a game balance consideration.
The devs are not insensitive to the amount of bank space required by DoN cultural armors. They will consider ways to reduce bank space, and they are also considering increasing stack sizes on armor molds and augment patterns to 200.
As previously mentioned, DoN armors will get a review for viability, including upgrades to some slots, if appropriate. In addition, the devs will try to figure out why legs don't get a Type 12 augment.
A tradeskill super-trophy will not be coming soon, but it is still something the developers are considering.
Bag tokens will probably show bag stats with the June 22 patch.
Velium weapon to velium ore conversion are almost done. They will be going to QA next week, and if they pass, they will be included in an upcoming patch.
The Grand Robe / Grand Tunic / Phylactery quest will be reviewed. Making changes to this quest is tricky, though, as all changes have to be approved by Swords and Sorcery, the publishers of the pen-and-paper EQ RPG.
Lubricous Metal Bits are not a real item, and the recipe for them will be removed.
Karana's Heartfruit Salad will be live in the game with the next patch.
Recipes for Taelosian wheat and mountain wheat are also coming pretty soon. My impression was that this would be about two patches out, but I'm not positive. The recipes will be largely based on the player-submitted ideas on the EQLive boards.
The Ice Cream Churn will get a token with the next patch.
Absor said he would send me additional responses to some questions by email, so I'll post those as I get them.
Comment