Race Stat Changes
An archived conversation. New replies are closed.
What simulations are you wanting here? Chanter vs dwarf and chanter vs orc? Orc melee vs dwarf melee? I think if we want to see some simulations you need to ask for specifics. 12 races and 12 classes = a shitload of possibilities.
Balance pure melee vs pure melee first of all (races designed for melee without MR i.e. Drake/Sauri/Orc)
Have a typical melee build of those races face each other without classes factoring into it.
This would be done using Nitehawk's attack simulator. If the simulator is too limited to do this then write some sorta scripting to have some characters automatically fight against each other and warp them back.
See if each of these 3 races has a fairly even chance at winning against each other, if they don't then look at what formulas or racial attributes would need to be tweaked to even this up.
If this isn't possible then consider racial stat tweaks as a last resort to bring balance to this.
After this do some smaller tests to make sure these races are still roughly equal when it comes to effects from different classes (i.e. classes with str multipliers, classes with more AC)
After this set of testing look to balance Melee with MR races against pure melee (Goblin, DE, Dwarf, maybe ling if you want this to be useable as melee)
In my opinion Melee with MR races should be able to win 40-45% of the time against the pure melee races.
They can't be completely evenly matched or noone would bother making sauri/orc/drake any longer.
Again, if these tests prove one race isn't competitive as the racial stats stand now then look at what tweaks can be done, but with racial stat tweaks as a last resort.
Maybe also consider some sort of melee "mercy rule", so if people make wacky combinations for fun (gnome slayer, elemental zerker) then these characters still have a chance to win at least 33% of the time
Once melee combat is balanced with each other you'd then have to look at balancing MR using a similar method.
Look to tweak MR formulas first or racial attributes first, with racial stats as a last resort.
Also consider tweaking healing formulas for the "all round races", both for priests and guardians (half elf, human, halfling, dwarf, goblin, and to an extent elf/gnome)
Also consider druid formulas for the "all round races"
Also consider ranger/bard formulas for the "all round races"
I strongly believe racial stat tweaks should be a last resort, and be fully justified, as these could potentially cause a lot of frustration/anguish/players quitting/general bad feeling.
I thought i'd posted all of this 5 days ago (10 pages back), but i think i left it off as my last post was already too long.
Please can people stop putting loads of other wacky ideas in this thread that aren't related directly to racial stats and please put them in a separate suggestion post?
This thread is really hard for most people to keep up with, which means we're only getting very limited feedback on this.
Goblin vs horc in melee simulator? Anything can be melee. So shouldn't we include elemental ninja vs saur ninja as well? *sarcasm
While I am normally a stats guy and agree some stats for fights and comparisons would be nice I am just not sure how realistic they would be.
Even if we just take a look at 1v1 stats which is a very small part of actual game pvp we are looking at tens of thousands of combinations each with different quantitative results.
So first off im guessing we would test 1 class vs 1 class so thats 12 races against 12 races
de vs drake
de vs dwarf
de vs elemental
de vs elf
de vs gnome
de vs goblin
de vs half elf
de vs halfling
de vs half orc
de vs human
de vs saurian
then you got
drake vs dwarf
drake vs elemental
drake vs elf
etc...
I'm guessing 1000 simulated fights for each of the combos would be enough (so you are looking at millions of simulations considering each race and class combo). Then once you get all these results youll need a number system to quantify the win because even once you find which race is the best of the best within the class youll need to do this for each class.
For example if half orc wins 9/12 classes but dwarf wins 7/12 but just barely loses each of its matches by less than the orc loses which is really better? is a win just a win? how do you factor this when you do the inner classes tests and dwarf cav beats 3/12 chanter races and orc beats 1/12 chanter races. i mean the possibilities are endless pretty much.
then youll need to do each class of each race against the others (ie half orc might be best against melee bards but it wont stand up to chanters so you have to take that into consideration as well). thats where the point system will come into place.
tldr:
There are thousands and thousands of possible combinations that would need tested just to get a proper statistical insight into just 1v1 combat (which is not even the most common type of pvp in game).
You would have MILLIONS of combinations to get proper statistical insight to actual party based pvp.
I honestly dont think a professional game studio could accomplish this task with months of time and they honestly wouldn't even bother trying. Something they do try is to get analytics from real game matches this however would be somewhat hard unless tracked by koth runners for maybe a year or so.
Even if we just take a look at 1v1 stats which is a very small part of actual game pvp we are looking at tens of thousands of combinations each with different quantitative results.
So first off im guessing we would test 1 class vs 1 class so thats 12 races against 12 races
de vs drake
de vs dwarf
de vs elemental
de vs elf
de vs gnome
de vs goblin
de vs half elf
de vs halfling
de vs half orc
de vs human
de vs saurian
then you got
drake vs dwarf
drake vs elemental
drake vs elf
etc...
I'm guessing 1000 simulated fights for each of the combos would be enough (so you are looking at millions of simulations considering each race and class combo). Then once you get all these results youll need a number system to quantify the win because even once you find which race is the best of the best within the class youll need to do this for each class.
For example if half orc wins 9/12 classes but dwarf wins 7/12 but just barely loses each of its matches by less than the orc loses which is really better? is a win just a win? how do you factor this when you do the inner classes tests and dwarf cav beats 3/12 chanter races and orc beats 1/12 chanter races. i mean the possibilities are endless pretty much.
then youll need to do each class of each race against the others (ie half orc might be best against melee bards but it wont stand up to chanters so you have to take that into consideration as well). thats where the point system will come into place.
tldr:
There are thousands and thousands of possible combinations that would need tested just to get a proper statistical insight into just 1v1 combat (which is not even the most common type of pvp in game).
You would have MILLIONS of combinations to get proper statistical insight to actual party based pvp.
I honestly dont think a professional game studio could accomplish this task with months of time and they honestly wouldn't even bother trying. Something they do try is to get analytics from real game matches this however would be somewhat hard unless tracked by koth runners for maybe a year or so.
Most games throw the changes in and expect feedback and change it based on that, and/or have a test server that has the test build until release.
While I'll test out certain things, there's only so much I can do with a short amount of time, and atm this is eating at players actually leveling or doing stuff so its hurting the community IMO. You can't really take into account potions, player delays, etc effectively, and as it was said, 1vs1 vs 6vs6 with priests etc work differently in this game obviously. So we can get some sample data but it's still paper data and not actual live.
50% fizzle for most spells is 3 to 4% to 1.5% to 2%. It can be a little lower than 50% but 5-10% etc prob makes very little difference.
While I'll test out certain things, there's only so much I can do with a short amount of time, and atm this is eating at players actually leveling or doing stuff so its hurting the community IMO. You can't really take into account potions, player delays, etc effectively, and as it was said, 1vs1 vs 6vs6 with priests etc work differently in this game obviously. So we can get some sample data but it's still paper data and not actual live.
50% fizzle for most spells is 3 to 4% to 1.5% to 2%. It can be a little lower than 50% but 5-10% etc prob makes very little difference.
While I am normally a stats guy and agree some stats for fights and comparisons would be nice I am just not sure how realistic they would be.
Even if we just take a look at 1v1 stats which is a very small part of actual game pvp we are looking at tens of thousands of combinations each with different quantitative results.
So first off im guessing we would test 1 class vs 1 class so thats 12 races against 12 races
de vs drake
de vs dwarf
de vs elemental
de vs elf
de vs gnome
de vs goblin
de vs half elf
de vs halfling
de vs half orc
de vs human
de vs saurian
then you got
drake vs dwarf
drake vs elemental
drake vs elf
etc...
I'm guessing 1000 simulated fights for each of the combos would be enough (so you are looking at millions of simulations considering each race and class combo). Then once you get all these results youll need a number system to quantify the win because even once you find which race is the best of the best within the class youll need to do this for each class. then youll need to do each class of each race against the others (ie half orc might be best against melee bards but it wont stand up to chanters so you have to take that into consideration as well). thats where the point system will come into place.
tldr:
There are thousands and thousands of possible combinations that would need tested just to get a proper statistical insight into just 1v1 combat (which is not even the most common type of pvp in game).
You would have MILLIONS of combinations to get proper statistical insight to actual party based pvp.
As just discussed in Discord, rather than completely shooting in the dark like we are now then at least attempting to balance 1v1 would give us a good starting point for balance rather than what we're doing at the moment.
I disagree on needing to test every single race against each other.
For example if you balanced Sauris/Drakes/Orcs for melee, and balanced Dark Elves/Goblins/Dwarves for melee, then you could test sauris vs dark elves and drakes vs goblins for example.
I think the main aim here is finding out what races are underpowered or overpowered and then look at ways of rectifying that without going overboard.
c/p from Discord:
Eld - Today at 9:08 PM
i just think that if you want to be able to balance the game properly then you need the tools to help you do it
having an automated fight script/analysis script
would be so invaluable
for everything
and this is the ideal application for it
i just feel it's better to work from a known knowledgebase
we have certain assumptions we've made based on a limited sample size. I know in rok people assumed certain combos were good but were overlooking important things that showed up in testing.
i also remember when i started some very good players kept telling me to make certain combinations because they were great, but the zeitgeist has changed even though no stats or formulas have and they now have different opinions
if we're making big changes that affect everyone then we should be working from fact not hearsay(edited)
i even assume now that 25 str orc slayers are gods, maybe they're not, maybe their hittrate fucking sucks compared to a 22 agi slayer, and maybe a 22 agi/23 str slayer is actually godly.
maybe they're equal
who the fuck knows??
how can we make big changes without knowing???
There's definitely an issue of huge numbers of results from sims. A base of tests is fine, I think, but of course it's going to change depending on race and class. Take orc slayer v drake slayer, for example. Now orc zerker v drake zerker. Those results won't be the same, you know what I mean here? Now which of those do we balance with? We could look at them and try to find middle grounds I suppose, but this is an enormous undertaking.
If we're comparing just orc/drake/lizard and we are comparing let's just say 4 classes (barb, knight, monk, assy). For each class you have 6 permutations to compute (orc/orc, orc/drake, orc/lizard, drake/drake/, drake/lizard, lizard/lizard). x4 classes = 24 simulations to run.
That's only 3 races and 4 classes. It's a lot of work.
I think an attack simulator would be great, but it's still a simulator and NH has a very valid point about adding in all the little racial bonuses and pot use, and lol forget simming what's going on in 6v6 combat with heals. Saying we can balance everything with sims is a bit far fetched, unless you have a godmode sim that can map out 6v6 combat.
My point is I see no way to do this without using input and "feels" from people who are playing every day. I don't need a sim to know an orc slayer can round something and it's far less likely a lizard does that, but if you sim out those 2 fighting each other the lizard may win half the time. Does that make them equal? Not in my eyes - orc is better because of the fact it can round things more often. That doesn't show on a sim.
If we're comparing just orc/drake/lizard and we are comparing let's just say 4 classes (barb, knight, monk, assy). For each class you have 6 permutations to compute (orc/orc, orc/drake, orc/lizard, drake/drake/, drake/lizard, lizard/lizard). x4 classes = 24 simulations to run.
That's only 3 races and 4 classes. It's a lot of work.
I think an attack simulator would be great, but it's still a simulator and NH has a very valid point about adding in all the little racial bonuses and pot use, and lol forget simming what's going on in 6v6 combat with heals. Saying we can balance everything with sims is a bit far fetched, unless you have a godmode sim that can map out 6v6 combat.
My point is I see no way to do this without using input and "feels" from people who are playing every day. I don't need a sim to know an orc slayer can round something and it's far less likely a lizard does that, but if you sim out those 2 fighting each other the lizard may win half the time. Does that make them equal? Not in my eyes - orc is better because of the fact it can round things more often. That doesn't show on a sim.
U already have know what's underpowered
Sub 21 str
10 wis
Sub 20 end
It's basically a 3 tier system of character types
Bringing them closer in win odds will just make everyone use high Mr melee builds.not to mention closer they are the more Mr formula won't work as intended.
You can make an excel spreadsheet and do all that testing.
Prop 19 fixes a bunch of inconsistencies
Run it
Sub 21 str
10 wis
Sub 20 end
It's basically a 3 tier system of character types
Bringing them closer in win odds will just make everyone use high Mr melee builds.not to mention closer they are the more Mr formula won't work as intended.
You can make an excel spreadsheet and do all that testing.
Prop 19 fixes a bunch of inconsistencies
Run it
Most games throw the changes in and expect feedback and change it based on that, and/or have a test server that has the test build until release.
While I'll test out certain things, there's only so much I can do with a short amount of time, and atm this is eating at players actually leveling or doing stuff so its hurting the community IMO. You can't really take into account potions, player delays, etc effectively, and as it was said, 1vs1 vs 6vs6 with priests etc work differently in this game obviously. So we can get some sample data but it's still paper data and not actual live.
isnt that what we have been doing for the past 9 months? I mean I feel like these suggested changes reflect our playing / testing over these last 9 months. suggestions are input from us players and event runners. I dont think that will stop after these changes either but thats what we have and option 19 pretty much reflects that. its just not going to be possible for everyone to totally agree which is what i think is holding us back atm.
there's too many player options and choices to accurately pinpoint and reflect every problem.
you can balance everything for 1v1 w/o potions thats cool, once you add potions strength is less essential and agility becomes far more important
you could attempt to balance 6v6 by setting characters hps higher than the avg rounds 6 decent characters would do to kill it. however thats a joke and blows 1v1 balance permanantly.
what you have is basic race templates that can be balanced with sims or minimal ingame info. this is more or less completed on prop 19, everyone argued mostly every race's PoV and reflected which stat changes hurt that race's class builds the most. theres no denying proposal 19 looks better than current so just add it.
when i have worked with gaming industries in the past (on a few mmo's, diablo 2, dnd 3.0-4.0) these industries only test averages period.
they dont care if a horc slayer is op until next patch because stats are more or less irrelevant in comparison to classes ability formulas. the reason is, if a horc slayer is OP vs other horcs, its 99.9% likely a saurian slayer is OP vs other saurians, etc so why even compare that races with lesser combative stats. you can crunch that data if you want but the fact is the biggest problem with how the data is coming in is everyone tends to compare all data to saurians, horcs and to an extreme of almost entirely the same class, slayer. these companies patch classes, the race tends to not matter nearly as much.
most games have a 2 tier system for characters. they have melee with varying degrees of defenses vs spellslingers with varying degrees of defenses accomplished by players adding X stats each level. where players have to go defenseless fanatic to 1 shot people, or go heavy into a mixture of defensive stats to be better rounded.
embers trying to accomplish a melee>MR>mage>melee circle with flat stats which already makes it harder. then on top of that you have
some players wanting the game closer to 50/50 win odds no matter what you face or have(lame). if anything at all a human 20 20 20 10 20 14 should be the only choice ingame that 50/50s everything.
look at the game from this perspective heavy melee str/agi build> heavy melee strength build> Decent melee High MR > crappy melee superb MR> Heavy magic > magic with well rounded defense/hps> heavy melee
when you look at it from that viewpoint the game is already a success. PoV from each race has been gathered about each class within that race. anything left off of Prop 19 is either of very little significance or such a massive significance its far above a stat change mattering anyway.
Prop 19 is ready.
you can balance everything for 1v1 w/o potions thats cool, once you add potions strength is less essential and agility becomes far more important
you could attempt to balance 6v6 by setting characters hps higher than the avg rounds 6 decent characters would do to kill it. however thats a joke and blows 1v1 balance permanantly.
what you have is basic race templates that can be balanced with sims or minimal ingame info. this is more or less completed on prop 19, everyone argued mostly every race's PoV and reflected which stat changes hurt that race's class builds the most. theres no denying proposal 19 looks better than current so just add it.
when i have worked with gaming industries in the past (on a few mmo's, diablo 2, dnd 3.0-4.0) these industries only test averages period.
they dont care if a horc slayer is op until next patch because stats are more or less irrelevant in comparison to classes ability formulas. the reason is, if a horc slayer is OP vs other horcs, its 99.9% likely a saurian slayer is OP vs other saurians, etc so why even compare that races with lesser combative stats. you can crunch that data if you want but the fact is the biggest problem with how the data is coming in is everyone tends to compare all data to saurians, horcs and to an extreme of almost entirely the same class, slayer. these companies patch classes, the race tends to not matter nearly as much.
most games have a 2 tier system for characters. they have melee with varying degrees of defenses vs spellslingers with varying degrees of defenses accomplished by players adding X stats each level. where players have to go defenseless fanatic to 1 shot people, or go heavy into a mixture of defensive stats to be better rounded.
embers trying to accomplish a melee>MR>mage>melee circle with flat stats which already makes it harder. then on top of that you have
some players wanting the game closer to 50/50 win odds no matter what you face or have(lame). if anything at all a human 20 20 20 10 20 14 should be the only choice ingame that 50/50s everything.
look at the game from this perspective heavy melee str/agi build> heavy melee strength build> Decent melee High MR > crappy melee superb MR> Heavy magic > magic with well rounded defense/hps> heavy melee
when you look at it from that viewpoint the game is already a success. PoV from each race has been gathered about each class within that race. anything left off of Prop 19 is either of very little significance or such a massive significance its far above a stat change mattering anyway.
Prop 19 is ready.
there's too many player options and choices to accurately pinpoint and reflect every problem.
you can balance everything for 1v1 w/o potions thats cool, once you add potions strength is less essential and agility becomes far more important
you could attempt to balance 6v6 by setting characters hps higher than the avg rounds 6 decent characters would do to kill it. however thats a joke and blows 1v1 balance permanantly.
what you have is basic race templates that can be balanced with sims or minimal ingame info. this is more or less completed on prop 19, everyone argued mostly every race's PoV and reflected which stat changes hurt that race's class builds the most. theres no denying proposal 19 looks better than current so just add it.
when i have worked with gaming industries in the past (on a few mmo's, diablo 2, dnd 3.0-4.0) these industries only test averages period.
they dont care if a horc slayer is op until next patch because stats are more or less irrelevant in comparison to classes ability formulas. the reason is, if a horc slayer is OP vs other horcs, its 99.9% likely a saurian slayer is OP vs other saurians, etc so why even compare that races with lesser combative stats. you can crunch that data if you want but the fact is the biggest problem with how the data is coming in is everyone tends to compare all data to saurians, horcs and to an extreme of almost entirely the same class, slayer. these companies patch classes, the race tends to not matter nearly as much.
most games have a 2 tier system for characters. they have melee with varying degrees of defenses vs spellslingers with varying degrees of defenses accomplished by players adding X stats each level. where players have to go defenseless fanatic to 1 shot people, or go heavy into a mixture of defensive stats to be better rounded.
embers trying to accomplish a melee>MR>mage>melee circle with flat stats which already makes it harder. then on top of that you have
some players wanting the game closer to 50/50 win odds no matter what you face or have(lame). if anything at all a human 20 20 20 10 20 14 should be the only choice ingame that 50/50s everything.
look at the game from this perspective heavy melee str/agi build> heavy melee strength build> Decent melee High MR > crappy melee superb MR> Heavy magic > magic with well rounded defense/hps> heavy melee
when you look at it from that viewpoint the game is already a success. PoV from each race has been gathered about each class within that race. anything left off of Prop 19 is either of very little significance or such a massive significance its far above a stat change mattering anyway.
Prop 19 is ready.
Word.
Make change, if it works awesome, if it dont, we can redo, people make this more complicated then it should be
Yes prop 19
Yes prop 19
Just a reminder you're getting the views of 6? people who are fatigued from arguing over this for a month/28 pages.
I'm not sure how the wider playerbase feels, but I'd be very frustrated if i was an average player who stopped levelling after 2 characters.
I'd personally be very keen to change the races of several of my 25ers, it'd be nice if this was an option with say a 25% exp penalty.
I'm not sure how the wider playerbase feels, but I'd be very frustrated if i was an average player who stopped levelling after 2 characters.
I'd personally be very keen to change the races of several of my 25ers, it'd be nice if this was an option with say a 25% exp penalty.
The good thing for me is that I have a wide variety of races. My characters race/class combo were made based on a theme/names rather than best combinations of race/class <img src="{SMILIES_PATH}/icon_e_smile.gif" alt=":)" title="Smile" />
I think terron said it best. ...
I think I disagree with Eld/JF on this one... PROVIDED that there is appropriate player compensation being offered from NH (eg some racial respecs given - of a generous enough nature that most ppl generally won't feel burned by it)
Simulation testing would take a lot of time and probably not reach any consensus either (due to ppls vested interests etc). It's especially ineffective basing simulations off 1v1 which is very rare outside of formally structured events. Group pvp is where most of the gameplay happens and it would be pretty tough on resources/time to simulate the number of possible permutations of that. Perhaps if everyone collected real group pvp logs for awhile as samples for NH it could help, but then even analysing that data would be a mess as it would be a skewed sample.
Therefore if NH is willing to compensate players with enough racial respecs (or at least something allowing ppl's exp investment not to be wasted entirely) so that players generally end up leaving this change happy rather than burned, then I say go ahead and implement the best guess fix (seems to be proposal 19 for now) and then fix again later if needed. However if inadequate compensation is likely to be offered to players, then it's dann risky changing anything and id probably say that NH mayking any changes would be "at his own peril".
Simulation testing would take a lot of time and probably not reach any consensus either (due to ppls vested interests etc). It's especially ineffective basing simulations off 1v1 which is very rare outside of formally structured events. Group pvp is where most of the gameplay happens and it would be pretty tough on resources/time to simulate the number of possible permutations of that. Perhaps if everyone collected real group pvp logs for awhile as samples for NH it could help, but then even analysing that data would be a mess as it would be a skewed sample.
Therefore if NH is willing to compensate players with enough racial respecs (or at least something allowing ppl's exp investment not to be wasted entirely) so that players generally end up leaving this change happy rather than burned, then I say go ahead and implement the best guess fix (seems to be proposal 19 for now) and then fix again later if needed. However if inadequate compensation is likely to be offered to players, then it's dann risky changing anything and id probably say that NH mayking any changes would be "at his own peril".
Reiterating that I think at the least everyone should get free respecs. I also think offering 1-2 free racial changes would go a long ways towards keeping people happy.
I've been following whole development and agree we can't sim every possible combination for balance. There will always be races that are better situated for 1v1 and others 6v6. Prop 19 looks good to me as well and giving each account 1-2 free racial changes should keep everyone happy.
Dunno about the "price" of changing the race of a char, but a free "stats-only" reroll must be given for each character if their race is modified in the update.
Dunno about the "price" of changing the race of a char, but a free "stats-only" reroll must be given for each character if their race is modified in the update.
I am not totally sure on this but i believe the plan is for every character to get 1 free stat allocate. On top of this there is a chance for 1-2 free race changes. Anything beyond that would be at a cost or not even an option.
I think prop 19 is good, except that orcs should probably get -1 end here? They don't really need a buff in any form and not nerfing their HP along with the heavy melee races is an oversight imo. Dwarf should have the HP advantage as they are the tanky race in theory.
Basically orcs are gaining 20hp which seems unnecessary.
Everything else seems in line to me.
Basically orcs are gaining 20hp which seems unnecessary.
Everything else seems in line to me.