Introduction
The information on this page is based on testing, hints provided by wizards, analysis of a large database of level gains sent to me by many players and lot's of guess work. I'm not entirely certain about many things, also I can't guarantee I haven't made mistakes. Some of the logs are old, and there might have been changes in guilds meanwhile that rendered some thing not up to date.
I'd like to thank to Avala, Cliff, Dinadan, Hallvardr, Keylos, Loktai, Somnusmors, Syndic, Taos, Tashi and possibly many others I forgot about for sending me their advancement logs. I gathered over 2000 guild levels information. This was a great effort! Still, this isn't that much. Things vary from one guild to another, and for many guilds I have logs from just several levels. This is often not enough to learn much about a guild. Many guilds are not included at all.
Primary, secondary, tertiary and dprime guilds
Primary guilds get the the largest number of stats. As a rule of thumb, secondaries/dprimes get 50% of that, and tertiaries 75%. E.g. a tert with "small" stat potential will give you in average 75% of points that a primary guild with "small" stat would give.
Twenty levels of the guild in which you start count as primary guild. The order doesn't matter. E.g. if you start as a fallen, and after level 15 join necromancer, and later take the last 5 levels of fallen, all 20 fallen levels will still count as primary. Tertiary guilds are those designed with 9 levels. Anything else (e.g. necromancer levels in the example above) counts as secondary/dprime, regardless of how many levels you take or in what order.
This means that in terms of total stat gain 2 levels of a tert are equivalent to 3 levels of secondary/dprime.
TL;DR
The tables below show how many points you may expect to get from a guild level. Mean column represents typical value for guilds for which I have available logs. Width represents typical level by level variation.
Stat
| potential | primary | secondary/dprime | tertiary | |||
|---|---|---|---|---|---|---|
| mean | width | mean | width | mean | width | |
| tiny | 1.9 | 1.4 | 0.61 | 0.89 | 1.3 | 1.2 |
| little | 4.0 | 2.1 | 1.9 | 1.4 | 2.9 | 1.9 |
| small | 6.0 | 2.8 | 3.2 | 1.7 | 4.8 | 2.4 |
| average | 5.1 | 2.1 | 7.2 | 3.0 | ||
| more | 6.8 | 2.4 | 10 | 3 | ||
| many | 8.2 | 2.4 | 12 | 4 | ||
| lots of | 14 | 4 | ||||
Hp or Sp
| potential | primary | secondary/dprime | tertiary | |||
|---|---|---|---|---|---|---|
| mean | width | mean | width | mean | width | |
| tiny | 1.8 | 1.6 | 2.8 | 1.8 | ||
| little | 8 | 3 | 3.8 | 2.0 | 6 | 3 |
| small | 13 | 5 | 7 | 3 | 9 | 4 |
| average | 21 | 7 | 11 | 4 | 16 | 6 |
| more | 13 | 5 | 21 | 7 | ||
| many | 30 | 10 | 24 | 7 | ||
| lots of | 17 | 5 | 30 | 9 | ||
Ep
| potential | primary | secondary/dprime | tertiary | |||
|---|---|---|---|---|---|---|
| mean | width | mean | width | mean | width | |
| tiny | 1.1 | 1.0 | 1.3 | 1.3 | ||
| little | 3.9 | 2.0 | 2.1 | 1.4 | 3.0 | 1.8 |
| small | 7 | 3 | 3.3 | 2.0 | 4.8 | 2.5 |
| average | 10 | 4 | 5 | 2 | 8 | 4 |
| more | 10 | 4 | ||||
| many | 13 | 5 | ||||
| lots of | ||||||
Hpr or Spr
| potential | primary | secondary/dprime | tertiary | |||
|---|---|---|---|---|---|---|
| mean | width | mean | width | mean | width | |
| tiny | 0.34 | 0.63 | 0.15 | 0.38 | 0.23 | 0.48 |
| little | 0.7 | 0.8 | 0.31 | 0.50 | 0.6 | 0.6 |
| small | 1.2 | 1.0 | 0.50 | 0.62 | 0.6 | 0.7 |
| average | 1.9 | 1.0 | 0.8 | 0.7 | 1.1 | 0.9 |
| more | 2.4 | 1.2 | 1.1 | 0.7 | 1.4 | 0.8 |
| many | 2.8 | 1.2 | 1.5 | 0.6 | 1.7 | 0.9 |
| lots of | 1.7 | 0.5 | ||||
Epr
| potential | primary | secondary/dprime | tertiary | |||
|---|---|---|---|---|---|---|
| mean | width | mean | width | mean | width | |
| tiny | 0.13 | 0.33 | ||||
| little | 0.35 | 0.56 | 0.11 | 0.33 | 0.16 | 0.40 |
| small | 0.6 | 0.7 | 0.36 | 0.57 | 0.29 | 0.53 |
| average | 0.7 | 0.8 | 0.36 | 0.55 | 0.45 | 0.56 |
| more | 0.6 | 0.6 | ||||
| many | 1.5 | 1.0 | ||||
| lots of | ||||||
Skill/spell categories
| potential | primary | secondary/dprime | tertiary | |||
|---|---|---|---|---|---|---|
| mean | width | mean | width | mean | width | |
| tiny | 0.08 | 0.28 | 0.02 | 0.15 | ||
| little | 0.19 | 0.44 | 0.07 | 0.26 | 0.11 | 0.32 |
| small | 0.29 | 0.51 | 0.13 | 0.34 | 0.15 | 0.36 |
| average | 0.44 | 0.59 | 0.13 | 0.33 | 0.26 | 0.44 |
| big | 0.46 | 0.50 | ||||
| large | 0.33 | 0.47 | ||||
| huge | 0.23 | 0.42 | 0.44 | 0.50 | ||
"Large" category bonus seems to be smaller than "big" – I guess that's some unfortunate fluctuation in the data I have.
Details
Stats, Hp/Sp/Ep maxes, Hp/Sp/Ep regens and skill/spell bonuses supposedly are rolled independently.
Stats
By stats I mean: Agi, Dex, Per, Con, Str, Cha, Int, Wis. The number of stat points you gain in a level results from two random processes and a value that is not fully known to a player.
- First the total number of stat points is rolled. Possible numbers of points are:
| guild type | total number of stat points | average number of stat points |
|---|---|---|
| primary | 10–29 | 19.5 |
| secondary/dprime | 5–14 | 9.5 |
| tertiary | 7–21 | 14 |
Every value in these ranges is equally probable (single dice roll).
- Second, each of those stat points is assigned to a specific stat with a probability defined by stat potential. Stat potentials given in guild descriptions do not represent the precise value of the probability, but vary from one guild to another. E.g. tiny Wis from Discordian is much tinier than tiny Wis from Toximancer. The text descriptions represent the following ranges:
| potential | probability range |
|---|---|
| tiny | 0.0–0.1 |
| little | 0.1–0.2 |
| small | 0.2–0.4 |
| average | 0.4–0.6 |
| more | 0.6–0.8 |
| many | 0.8–0.9 |
| lots of | 0.9–1.0 |
In order to determine precise value for each guild one needs to analyze multiple guild levels from every guild. The data I gathered is barely sufficient to give rough estimate for many of them.
In the attachment you will find a file stats.zip with best estimates of potentials for individual guilds. Number in brackets is the number of levels in my db. Error bars represent estimated uncertainty of each number: the less data I have, the larger uncertainty.
As you can see, the some guilds look mostly consistent e.g.: primary tiny is 10% chance (though e.g. Fallen Per seems lower), little is 20%, small is 25–30% while secondary and tertiary guilds vary much more one from another. This makes me guess that almost all potentials in primary guilds were assigned the same value, e.g. fighter little Dex (19±1%) and mage little Agi (22±2%) are both in fact 20%.
In order to calculate expected average level gain multiply the stat ratio by average number of total stats. E.g. necromancer assigns about 0.1 of all points to Wis, thus primary necromancer will in average grant 0.1×19.5 ≈ 2 Wis points per level, secondary necromancer will grant 0.1×9.5 ≈ 1 Wis point per level. I attach also files that show actual stat values, but due to additional randomness they will provide less accurate values.
The consequences of this algorithm are the following:
- The stats are somewhat anti-correlated. This means that if you get a good amount of one stat you are less likely to get good amount of another.
- There is quite some randomness. It averages out at higher levels… but not entirely. E.g. single primary guild level yields 19.5±5.5 stat points total (5.5 is standard deviation of uniform distribution from 10 to 29). 20 primary guild levels yield 390±25. individual stats are subject to additional randomness, the smaller stat potential, the larger relative randomness.
- Occasionally you may see values much different from the average.
- There are better and worse levels. This means that one could consider reincing several times and advancing up to level 20 or 40 (or whatever they consider cheap) until they get a favourable roll.
Points
The points, or maxes, that is Hp, Sp and Ep gains are more difficult to interpret. As you can see distribution of sum of maxes is somewhat flatish and squarish, but there are no sharp edges, as for stats.
Possible explanations
- The potentials don't sum up to 100% and some points are either lost or assigned to something else?
- There is something related to the fact that Ep gains are twice lower than Hp and Sp. Yes I tried multiplying Ep by 2, no it didn't help. Yes, I tried to look at guilds that don't give Ep, but I don't have enough data to draw conclusions.
The ratios of single max divided by sum of all of them (Hp+Sp+Ep) seem to fall into similar ranges as those for stats (Ep are twice smaller), but since I don't understand how the sum is rolled, I show distributions of the actual stats.
Regens
The story is similar as with maxes. It's not common, but some levels give no regen points.
Skill/spell category bonuses
This is the most tricky part.
- Primary and secondary guild levels can grant bonuses to any skill they provide
- Tertiary guilds, as they provide bonuses to a predefined list of skills, which you may have or not
- You only see the bonus upon advancement if you have the skill already trained. For this reason you very often don't see the individual skill bonuses you get
- Until mid of last year there was a bug that allowed you to get bonuses to skills from other guilds. Majority of the logs I have are affected by this.
For these reason I gave up completely on individual skill bonuses, and I focused only on skill category bonuses. The results are not great, as since these bonuses are small, the fluctuations are large. An interesting thing is that only primary guilds have chance to grant up to two category bonus points per level; secondaries are limited to 0 or 1.
Not studied
The following things may need dedicated testing:
- What are stats/maxes/regens at birth and how they depend on guild and race
- How individual skill/spell bonuses work. I'm not sure if it's feasible at all to study based on level advancement logs. Perhaps a better strategy would be to take all 20 levels of a primary guild, train everything a bit and display the summed skill bonuses.
