Goblinworks Executive Founder
|
| 5 people marked this as a favorite. |
Here is the collected notes of The Seventh Veil about the abilities available in PFO.
Not included: Attribute bonuses for purchasing each ability, abilities beyond the first level, or base damage information on any equipment.
Missing: Shortsword, Mace, Shield, Axe abilities.
Those who wish to contribute may request editorial rights.
Woooho very nice!
Complimenti.
|
Goblinworks Executive Founder
|
Wrathful Strike 1 Greatsword 33 2 3.36 3 Improved Critical +20 Open 1r to Attacker Unbalanced 1r if Attacker has Replying
I'm curious as to what the "restriction" Open 1r to Attacker means.
My expectation is that it applies the "Open" debuff to the character who uses it. I anticipate that there are/will be abilities that have "X if Target is Open".
Goblin Squad Member
|
At the moment, no-one can sort it except the editors.
If you make it open (editable by link without login) any changes can be reverted by the owner at any time. You also get a detailed list of any changes that were made, via a pretty much infinite list of versions you can look at and revert to.
You can also lock cells if you want, so that only you can do anything to those.
|
Goblinworks Executive Founder
|
At the moment, no-one can sort it except the editors.
If you make it open (editable by link without login) any changes can be reverted by the owner at any time. You also get a detailed list of any changes that were made, via a pretty much infinite list of versions you can look at and revert to.
You can also lock cells if you want, so that only you can do anything to those.
I'd rather admit new editors by request, since if everyone can sort I anticipate conflicts between e.g. people who want to sort by damage factor and people who want to sort by weapon type.
I encourage everyone to make a copy and play with that one, and also comment (right click-> insert comment) on cells with typos or strange values; sorting would mess up those comments.
|
Goblin Squad Member
|
This is just the beginning of the vast knowledge that The Seventh Veil will collect over time. Now we catch a real glimpse of what a University Settlement can accomplish. Thanks, Decius!
|
Goblin Squad Member
|
"provokes opportunity" appears in the "restrictions" category in the tooltip, so that's the column it appears in. Was that your question?
Some skills grant increases attribute increase.
Consider,
Schools (Abjurer, Conjurer, Enchanter, Evoker, Illusionist, Necromancer, Transmuter)
Rank 1 grants: 0.099 Int.
…
Arcana (all Library?) 1 & 2 grants: 0.083 & 0.093 Int.
…
Power 4, 5, & 6 grants: 0.017, 0.017, & 0.018 Int.
…
Clothing Armor Proficiency 1 grants: nothing.
Personally, I think increasing attributes is important.
|
Goblin Squad Member
|
"Grants" would be in the "Effects" or "Other" column. For example, "Bullseye Shot" applies the "Bleeding" debuff, while "Impaling Shot" applies "Bleeding" situationally.
[Edit] Ah, I see what you mean now. I think that data is incomplete. When I was capturing the effects on Attributes, I was only looking at the effect on Intelligence. We'll try to get more comprehensive data later.
I should add that I won't be LiveStreaming next week when the servers are up Thursday/Friday, as I'll be in Seattle at PaizoCon (and finally meeting the other Stewards face-to-face)
|
Goblinworks Executive Founder
|
Ah: Attribute increases will be reported in a different spreadsheet.
Executive summary: Knowledge skill (lower ranks first) are the most efficient for int in terms of XP/Int, and cleric domains are the best for wis. Information on str, dex, con and per has not been systematically collected.
Other prerequisites include achievements. So far it has not been noted that a specific achievement is required, only a category of achievement. (Adventure, Arcane, Crafting, Divine, Martial). In normal play it appears that in the time it takes to earn the XP required for the second and third levels of feats, most players should be able to get the required achievement(s).
|
Goblin Squad Member
|
And weapon grant strength or dex. I know I am beating on a wall. These will be important for low level character trying to make attribute 11 or 12 to lear another skill (per its prerequisites). Per what Nihimon has pated the information is there, it is just ignored in the spread sheets.
|
Goblinworks Executive Founder
|
And weapon grant strength or dex. I know I am beating on a wall. These will be important for low level character trying to make attribute 11 or 12 to lear another skill (per its prerequisites). Per what Nihimon has pated the information is there, it is just ignored in the spread sheets.
There will be a different spreadsheet indicating attribute increases. It is ignored here because where the information was collected, all of the attack feats for a role provide the same attribute bonus.
Basically, the goal of the research project was to learn about the attack abilities, and the ability score information was collected because it was easy to do so. A future project will be to collect as much information about ability score increases as possible. When gathering information about ability scores, the assumption will be made that attack abilities of a given tier provide the same ability score increase, and the hypotheses that gathering, refining, and crafting skills of the same level provide the same increase will be tested.
The current goals are to finish the attack ability data (mace, shortsword, axe[?]); gather preliminary qualitative results on PvP, crafting, and synergy between abilities; and defeat an escalation.
|
Goblin Squad Member
|
I tried playing with some math to see if I could figure out the system for how much weight each aspect of an attack cost in terms of the total budget when designing the attack.
The simplest case is to assume that everything is linear, ie. damage factor of 3 uses twice as much of the budget as 1.5, that distressed 1r 50% costs twice as much as 25%, etc.
However, that assumption is wrong as it leads to inconsistent results. So the weighting system must include some form of non-linearity...
|
Goblin Squad Member
|
Ah: Attribute increases will be reported in a different spreadsheet.
Executive summary: Knowledge skill (lower ranks first) are the most efficient for int in terms of XP/Int, and cleric domains are the best for wis. Information on str, dex, con and per has not been systematically collected.
What is the scientific method currently used for determining attribute increases? Please list all steps / math. I would like to be consistent with others who are posting.
Unlike Nihimon and DeciusBrutus, I will be playing this week (checking out gathering / crafting skills) and would like to be able to determine if / how much those skills increase which attribute.
|
Goblinworks Executive Founder
|
Get to the point where a higher attribute is required to acquire a new feat. Mouseover text on the 'unavailable feat' list indicates current ability score as well as required ability score. Base attack 2 is a good choice, since it will show all ability scores if they are all <11, but other feats might be more convenient in specific circumstances.
Each time you buy a feat, check your ability scores. Record the initial and final, and subtract initial from final to find the delta. Yes, this sometimes means running back and forth a lot. To improve breadth, if several skills in a category (e.g. shortsword attacks) have identical ability score increases, it is permissible to buy a known number of similar feats and simply confirm that the total ability score increase is consistent.
|
Goblin Squad Member
|
Does "Fighter 2" convey any mechanical benefit, or is it purely a checklist completion item?
Is "Fighter 2" a prerequisite to anything other than "Fighter 3"? i.e. are there some skills or feats which cannot be trained until you advance your role level? Or could you in theory train a selected set of abilities to 20 without ever raising a named role?
|
Goblin Squad Member
|
I know in clerical feats you must purchase more advanced cleric skills from the cleric trainer (at the cleric building: temple?) in order to become eligible for more advanced orisons from the seminary orison trainer. I'd infer that fighter 2 is required for more advanced fighter feats than are available at fighter1.
|
Goblin Squad Member
|
Does "Fighter 2" convey any mechanical benefit, or is it purely a checklist completion item?
Is "Fighter 2" a prerequisite to anything other than "Fighter 3"? i.e. are there some skills or feats which cannot be trained until you advance your role level? Or could you in theory train a selected set of abilities to 20 without ever raising a named role?
If Goblinworks doesn't publish something before I get to it, I'll be trying to capture as much information about the prerequisites as I can the next time I'm in Alpha (which won't be this weekend).
|
Goblin Squad Member
|
What is the scientific method currently used for determining attribute increases? Please list all steps / math. I would like to be consistent with others who are posting.
It varies greatly based on what Feat you're buying.
Knowledge Feats (Arcana, Dungeoneering, etc.) increased Intelligence by 0.083 for Rank 1, 0.093 for Rank 2, and 0.099 for Rank 3.
Wizard Attack Feats (Boiling Blood, Deafening Roar, etc.) increased Intelligence by 0.018 for Rank 1. I did not get data for Rank 2.
Decius and I both did fairly significant raw data dumps in First Impressions.
If you're playing a Wizard, I think you'll find my data dump in particular very useful for planning your progression to Wizard 2.
|
Goblinworks Executive Founder
|
If Goblinworks doesn't publish something before I get to it, I'll be trying to capture as much information about the prerequisites as I can the next time I'm in Alpha (which won't be this weekend).
Going to try and get some of this done this weekend, specifically for cleric and maybe some other classes.
My planned approach, and I am open to suggestions, is to work based off the list that Stephen provided for advancement. Take note of what feats increased what stats, which feats were a prereq for others, and which trainer I had to obtain it from. Probably will do most of this the second or third day of alpha this weekend so I have time for my XP to build up.
|
Goblin Squad Member
|
My planned approach, and I am open to suggestions, is to work based off the list that Stephen provided for advancement. Take note of what feats increased what stats, which feats were a prereq for others, and which trainer I had to obtain it from. Probably will do most of this the second or third day of alpha this weekend so I have time for my XP to build up.
I highly recommend creating a stable of characters to start earning XP as early as possible.
I was able to gather a lot of data because logging out reset my XP so I could spend it all again. I think that's going to be fixed this time around.
|
Goblinworks Executive Founder
|
There are fears that have role levels as prerequisites. I don't recall at the moment which ones, only that they made sense at the time. Regular players will probably not be impacted by the restrictions I noticed, or they will be better off for being told what they should get first.
|
Goblinworks Executive Founder
|
I think the hard part is going to be figuring out what achievments are needed to advance a role, unless these are also mentioned in the reqs for various feats on the tooltips. I will do some checking.
Hopefully I will be able to make 10 or more characters to sit and gain experience so I have a wide test pool.
One question I have, specific to the Alpha since we gain xp over time, assuming there isn't another character wipe for at least one or two server opening, if the server is off do we still gain experience? Currently it has only been noted you gain experience over time whether you are logged in or not, it didn't note what the case was for when the server was off. Obviously this isn't a real issue during the live game (except for maybe during maintenance periods). In any case, just curious how the xp over time calculations are handled.
|
Goblin Squad Member
|
| 1 person marked this as a favorite. |
Try assuming an exponential scaling, based either on damage factor or +1 DF.
The second part seems to work better. I had to bust out fitting algorithms to avoid needless repetition in solving equations, but I may have a solution now.
Basically I assigned a weight (ie. a cost) to each attribute of an attack and found a collection of 7 attacks which could all be described by 6 attributes. I then picked 6 of the 7 attacks and solved a set of equations for what the weights are. Then I swapped one of the attacks out and solved again. The solution for the weights should be the same, so I made a metric which measured the difference of the weights. Ideally the difference should be 0, but perhaps there is some rounding going on somewhere so maybe not exactly 0 (eg. maybe the budget says a damage factor of 0.76776 but it gets rounded to 0.77 for convenience).
But how to scale weights with increasing effects? Ie. how does the cost of distressed 25% compare to 50%. Is it double the cost? sqrt(2) the cost? 2^2 the cost? So assign an unknown parameter to be the relevant power of the scaling. Assume that damage factors have a different power than all other attributes, so that's a second parameter. Now assume that damage factors don't necessarily start at 0, but rather at some unknown value.
That gives 3 parameters to play with. Throw that into a fitting algorithm (minuit) and ask it to find the set of parameters which minimizes the differences in the weights.
Solution: all 3 parameters are 1.
The 7 attacks were: bright rune, eel's touch, flare, grasping shadows, icicle, shard spray, and false edge. Note that they all have the same cooldown and stamina cost.
It appears that you get a damage factor of 1.0 for free when budgeting an attack. Then for any attribute you apply if you double the effect of the attribute it costs twice as much (50% costs double 25% etc.)
It appears that the following are all roughly equal in terms of how much of the attack's budget they use (something like 32.5% of the budget):
*each 25% chance of distressed
*being at ranged 20 and provoking opportunity (as opposed to being melee and not provoking)
*each 5 points of oblivious
My results hint that you get back about the same 32.5% of the budget when doing a common energy damage type (fire, cold, electric) as opposed to physical. Or equivalently the budget for fire/cold/electric attacks is 32.5% bigger than physical damage attacks. Edit: actually, a better way of putting it is that the 3 above abilities cost ~25% of the budget and being a physical damage attack also uses up 25% of the budget.
5 points of slow and increases in damage factor appear relatively cheap by comparison. Damage factor is budgeted based on how much different than 1 it is.
Further tests are needed to confirm (or contradict) these results and to determine other costs.
|
Goblinworks Executive Founder
|
Suspicion: "Physical Damage" doesn't cost extra, but "Damage of a type other than normal for the weapon used" does.
Additionally, consider that different effects might have different costs with different weapons.
I don't have the answers, and I barely have the questions. I also think that including all of those variables results in the dataset we have being insufficient to analyze.
|
Goblin Squad Member
|
Bug fixed. (Words that have never come back to haunt anyone ever ;) )
If I assume that everything is linear and that different weapons have different budgets I can get arbitrarily good fits to the data, but (now that I've also added boiling blood into the mix) I get a negative budget for 'burning' which is obviously wrong. This is regardless of whether the extra budget means a different total budget for different weapons or everything costs proportionally less/more for different weapons.
Linear scaling seems to work the best of all scalings I've tried (raised to some power, exponential, logarithmic), but I'm not convinced it's necessarily the actual scaling used.
If the costs vary in different ways between different weapons then my method won't find it Eg. slowed costs x and burning costs y for a wand, but for a staff they cost 2x and 0.5y respectively. If they varied together then this method would find it, but varying in different ways is beyond the scope of my fit and the results returned would be nonsense (which is what I seem to be finding).
Or maybe there isn't a strict set of rules for the budget at all and the budget is instead based on feel. That would be easier to balance long term as the devs wouldn't be constrained to a given equation.
|
Goblin Squad Member
|
Hmmm...
So all the above is based on a 'budget' method, ie. spend X% on one attribute of an attack, Y% on another attribute, etc. Didn't get that to agree with data.
So next I tried a 'product' approach where you multiply things together. Every attribute then scales you up or down by X amount. Every attack should then work out to the same number in this system. Haven't got that agreeing with the data either.
|
Goblinworks Executive Founder
|
My guess right now is that delay determines max stamina used, and damage factor is a function of stamina used, unconditional effects, range! restrictions, and conditional effects adjusted for the condition.
I'd suggest a power relationship with regard to % chance effects.
Anticipate that there might be some weird multiplication: two 2x multipliers become a 3x multiplier in PF, (when they're allowed to stack at all). Two things that reduce DF by 50% each might be 25% together! or they might be 0. Increases and decreases don't have to follow the same rules.
|
Goblin Squad Member
|
I've tried power relationship and let the fitting program tell me which power works best in the above schemes. None work well enough to describe the data exactly.
The attacks I'm looking at all have the same stamina and cooldown, so neither of those variables should affect the fitting. If we can understand how the scaling works for attributes then the next step is definitely understanding stamina and cooldowns. I agree that stamina, cooldown, and damage factor all seem to be closely related.
I'll try the weird multiplication rules later today. A few things to take care of first...
|
Goblin Squad Member
|
I have some data collected on Cleric.
Starting Stats: Strength 10.108, Dexterity 10.000, Constitution 10.000, Wisdom 10.298, Intelligence 10.000, Personality 10.000
Hit Point skill raises Hit Points by 40 each rank with cost of: 1, 5, 25, 78, 189, 393, 728.
Religion did not appear to increase any stat.
Skill ----> Rank 1 Cost / + Wis ----> Rank 2 Cost / + Wis ----> Rank 3 Cost / + Wis
- Holy Implement ----> 1 / 0.095
- Focus Weapon -----> 49 / 0.083
- Divine Attack -------> 36 / 0.082 ----> 290 / 0.090
- Willpower Bonus --> 54 / 0.033 ----> 434 / 0.037
- Minor Cure ---------> 40 / 0.036 ----> 320 / 0.039
- Glory domain ------> 40 / 0.099 ----> 346 / 0.109 ----> 1217 / 0.115
- Power --------------> 2 / 0.014 --------> 8 / 0.015 ----> 35 / 0.016
- (ranks 4 -6) ------> 63 / 0.017 ----> 147 / 0.017 ----> 211 / 0.018
- Base Attack ------> 145 / 0.029 ----> 772 / 0.031
- Crusader -----------> 8 / 0.091 -----> 53 / 0.100 ----> 189 / 0.106
I am not listing all the domains or spells that could be purchased, but all the domains I bought provided the same amounts of + Wis as Glory and all spells provided the same as Minor Cure. Base Attack provided its + to all stats not just the Wis I have listed. I wasn't looking at Power, but it likely raised all Stats like Base Attack does.
As has been noted elsewhere, some skills or ranks of skills require additional requirements beyond prior rank and sufficient XP. Quick notes reveal Crusader requires Heavy Armor proficiency, and higher ranks of skills required Domain points and a Stat = Stat 9 + Rank of Skill (Glory domain rank 3 needs 12 Wis). Power did not follow this progression (looked to be 10 + 1/4 Rank of Skill)
Easiest way I found of getting Divine points was to do the Focus Achievement (killing with a spell).