These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

Player Features and Ideas Discussion

 
  • Topic is locked indefinitely.
12Next page
 

Do away with skill "levels" and just run benefits for training along a curve.
Author
Previous Topic Next Topic
Corey Fumimasa
CFM Salvage
#1 - 2013-05-15 16:46:31 UTC  |  Edited by: Corey Fumimasa
Skill levels seem so clunky and overbearing. Could the skill progression just be based on a curve without levels?

So if you want to train racial cruisers for a while then you will get bonuses based on time spent training as opposed to having to train for a specific time or level.

To me this just feels more organic and natural.

As for prerequisites they could be based on values that are comprised from the multiplication of various lesser skills. So maybe to get covops you don't need cloaking to be all the way to 5 if you have science, high speed maneuver, and sensor linking at very high levels.

somehow it feels like this would allow for more focused character builds as it would take the average of a players interests and provide a level of ability exactly representative of those interests.
Robert Caldera
Caldera Trading and Investment
#2 - 2013-05-15 16:58:07 UTC
dont want to train interdictors 5 years long just to compete with other interdictor pilots.
Destination SkillQueue
Doomheim
#3 - 2013-05-15 17:12:46 UTC
I'm not seeing the benefits of the system. Capped skills also keep the system somewhat sane and allow new players to catch up and be competative relatively soon. I'm asuming levels just make things easier from CCPs perspective and I don't see any actual problem they cause, so I don't see why they should be removed either.
Tchulen
Trumpets and Bookmarks
#4 - 2013-05-15 17:19:37 UTC
Seems overly complicated from a code perspective and I can't really see a benefit.
monkfish2345
School of Applied Knowledge
Caldari State
#5 - 2013-05-15 17:29:14 UTC
while this seems like a good idea, you have 2 big issues.

1) it's overcomplicated. good ideas tend to be simple ones.

2) regardless of how it looks, it will ultimately be a granular (levelled) system. essentially all your asking for is for more levels. and unless it gave an exponentially diminished return for each step it would lead to massive imbalances.

end of the day there is nothing wrong with the current system. and in it's current form it means in a single ships class it is possible for newer players to compete with people that have been playing far longer. because per hull there is a limit to how skilled you can be.
Kirimeena D'Zbrkesbris
Republic Military Tax Avoiders
#6 - 2013-05-15 17:33:29 UTC  |  Edited by: Kirimeena D'Zbrkesbris
Well, UO had gradual skill system (0.0-100.0%) with titles gained at certain skill %. While there is no need for such system in EvE, it could provide benefits from patrially trained skills. As for prerequisites - an analogy to those titles could be used. A downside is a longer time needed to get same efficiency than in current system (lets say lvl4 in skill vs 80%) unless some modifier (based on current skill%) for training speed is introduced. Similar modifier is already used in NPC corp/Faction standings gain calculation.

Opinions are like assholes. Everybody got one and everyone thinks everyone else's stinks.
Corey Fumimasa
CFM Salvage
#7 - 2013-05-15 23:56:49 UTC  |  Edited by: Corey Fumimasa
I think from a code perspective it would actually be simpler than the tiered system we now use. The values used would just be a percentage of the time trained. That would be one less step in the code as there is no reason to define the separate steps.

They could use the same limits and progression that we have now, so there would be no difference in time to benchmarks or end values.

The thing it clears up is partially trained skills, under this system all time spent training would be of value.

Just a thought I had to post, I think skills are actually pretty good as they currently sit, but this just seemed a bit more elegant and alien feeling at the same time. And I have always had a fondness for analog computing, it brings me back to old firecontrol problems.
Ruze
Next Stage Initiative
#8 - 2013-05-16 00:13:28 UTC
Corey Fumimasa wrote:
I think from a code perspective it would actually be simpler than the tiered system we now use. The values used would just be a percentage of the time trained. That would be one less step in the code as there is no reason to define the separate steps.

They could use the same limits and progression that we have now, so there would be no difference in time to benchmarks or end values.

The thing it clears up is partially trained skills, under this system all time spent training would be of value.

Just a thought I had to post, I think skills are actually pretty good as they currently sit, but this just seemed a bit more elegant and alien feeling at the same time. And I have always had a fondness for analog computing, it brings me back to old firecontrol problems.


Except the current system is coded, established, and working as intended.

Sure, a percentage leveling system is what I personally would go with if I was making my own game. It's not a bad idea at all. But 10 years in, changing systems between a fullly functioning system to another that would, a max training, do the exact same thing ... that's not going to be easier in the slightest.

Your point on partially trained skills is a good one, in my mind. But for many skills the benefit is small, and the math is very clear cut. The machine doesn't have to figure out what percentage of 100 your skill is at every time you fire your turret, it just gives you a 6% increase. Processing wise, it is probably a step or two simpler.

And when 3000 players are in a fight, a step or two per player per turret per shot is a LOT of steps.

If you're driven to threaten others with harm or violence because of what they do in game, you can't separate fantasy from reality. That "griefer/thief" is probably more sane than you are. How screwed up is that?
Kirimeena D'Zbrkesbris
Republic Military Tax Avoiders
#9 - 2013-05-16 04:50:19 UTC
Ruze wrote:

Your point on partially trained skills is a good one, in my mind. But for many skills the benefit is small, and the math is very clear cut. The machine doesn't have to figure out what percentage of 100 your skill is at every time you fire your turret, it just gives you a 6% increase. Processing wise, it is probably a step or two simpler.

And when 3000 players are in a fight, a step or two per player per turret per shot is a LOT of steps.

Is there much difference in calculating: damage = damage*(1+0.05*skilllvl) and damage = damage*(1+0.25*skill%)? I'd say there is no difference at all.

Opinions are like assholes. Everybody got one and everyone thinks everyone else's stinks.
Ruze
Next Stage Initiative
#10 - 2013-05-16 05:03:30 UTC
Kirimeena D'Zbrkesbris wrote:
Ruze wrote:

Your point on partially trained skills is a good one, in my mind. But for many skills the benefit is small, and the math is very clear cut. The machine doesn't have to figure out what percentage of 100 your skill is at every time you fire your turret, it just gives you a 6% increase. Processing wise, it is probably a step or two simpler.

And when 3000 players are in a fight, a step or two per player per turret per shot is a LOT of steps.

Is there much difference in calculating: damage = damage*(1+0.05*skilllvl) and damage = damage*(1+0.25*skill%)? I'd say there is no difference at all.


I guess you've got me there. I didn't know that this was the math behind damage calculations, myself. I really haven't dug into the code or programmed anything for the game, so to me, most of this is hypothetical discussion. But I appreciate the point you made.

If you're driven to threaten others with harm or violence because of what they do in game, you can't separate fantasy from reality. That "griefer/thief" is probably more sane than you are. How screwed up is that?
hmskrecik
TransMine Group
Gluten Free Cartel
#11 - 2013-05-16 07:25:41 UTC  |  Edited by: hmskrecik
Kirimeena D'Zbrkesbris wrote:

Is there much difference in calculating: damage = damage*(1+0.05*skilllvl) and damage = damage*(1+0.25*skill%)? I'd say there is no difference at all.

By my counts, the difference is dramatic. It takes exponentially longer to train each skill level while the percentage based calculations assume, you didn't state otherwise, that timed progress is linear.

And if you propose logarithmic curve of percent progression then you propose the same system as existing, only more complicated (100 skill levels instead of 5, more granular SP calculation required).

Edit: also it messes with the reward system. As of now the "Skill training complete" gong means something, esp. if it was month-long bringing some nice or important skill up to V. Even if you scale the system to have the same net effect, you will strip it from this particular form of delayed gratification.
Kirimeena D'Zbrkesbris
Republic Military Tax Avoiders
#12 - 2013-05-16 07:42:17 UTC  |  Edited by: Kirimeena D'Zbrkesbris
hmskrecik wrote:
Kirimeena D'Zbrkesbris wrote:

Is there much difference in calculating: damage = damage*(1+0.05*skilllvl) and damage = damage*(1+0.25*skill%)? I'd say there is no difference at all.

By my counts, the difference is dramatic. It takes exponentially longer to train each skill level while the percentage based calculations assume, you didn't state otherwise, that timed progress is linear.

And if you propose logarithmic curve of percent progression then you propose the same system as existing, only more complicated (100 skill levels instead of 5, more granular SP calculation required).

Edit: also it messes with the reward system. As of now the "Skill training complete" gong means something, esp. if it was month-long bringing some nice or important skill up to V. Even if you scale the system to have the same net effect, you will strip it from this particular form of delayed gratification.

Quote:
A downside is a longer time needed to get same efficiency than in current system (lets say lvl4 in skill vs 80%) unless some modifier (based on current skill%) for training speed is introduced. Similar modifier is already used in NPC corp/Faction standings gain calculation.

Art of quoting self.

Question was not in learning calculations, but skill calculation during combat and that stays the same. Also i agree that "skill training complete" is an achievement, but gradual skill bonus system makes more sense because you cant instantly get better in something - you take your time to learn something and can put new information to use as soon as you get it (but with less efficiency) and not only @ end of course. IMO "skill training complete" will be even more of an achievement if it sounded only @ 100%. Blink

As a side note, change to skill queue will be needed. Maybe limit it by number of skills in queue, or by sum of skill training time modifiers or both. Or maybe give an option(slider) for player to choose time he wants to train each skill in a 24hr timeframe, but no less than 5-10min and an ability to train to certain % (so you could put any skill to train for 23hrs59mins and then add another skill to train afterwards till 100%).

Opinions are like assholes. Everybody got one and everyone thinks everyone else's stinks.
hmskrecik
TransMine Group
Gluten Free Cartel
#13 - 2013-05-16 09:26:22 UTC
Kirimeena D'Zbrkesbris wrote:

Question was not in learning calculations, but skill calculation during combat and that stays the same. Also i agree that "skill training complete" is an achievement, but gradual skill bonus system makes more sense because you cant instantly get better in something - you take your time to learn something and can put new information to use as soon as you get it (but with less efficiency) and not only @ end of course. IMO "skill training complete" will be even more of an achievement if it sounded only @ 100%. Blink


Ok, so you assume logarithmic progression. But even then if we scale the system to have exactly the same effect so instead of training to IV in one day you get to 80% in one day, and to 100% in next four days, what is actual gain? Please note the bonuses will have to be scaled accordingly, thus instead of 5% per skill level you'll have 0.25% per skill percent - I have my doubts whether it qualifies as "instantly getting better".

The only effect obvious to me is already mentioned delayed vs. instant gratitude.

I mean, I'm not totally opposed to this idea, I just don't see clear advantages which could justify CCP's effort to implement it. This part of EVE is not broken and you better make REALLY good case before insisting it should be fixed.
Kirimeena D'Zbrkesbris
Republic Military Tax Avoiders
#14 - 2013-05-16 10:03:36 UTC  |  Edited by: Kirimeena D'Zbrkesbris
hmskrecik wrote:
Ok, so you assume logarithmic progression. But even then if we scale the system to have exactly the same effect so instead of training to IV in one day you get to 80% in one day, and to 100% in next four days, what is actual gain? Please note the bonuses will have to be scaled accordingly, thus instead of 5% per skill level you'll have 0.25% per skill percent - I have my doubts whether it qualifies as "instantly getting better".

The only effect obvious to me is already mentioned delayed vs. instant gratitude.

I mean, I'm not totally opposed to this idea, I just don't see clear advantages which could justify CCP's effort to implement it. This part of EVE is not broken and you better make REALLY good case before insisting it should be fixed.

Yes, it will make current 5lvl system into 100(1000 or more if they implement fractions of %)lvl system which is much more flexible. Bonuses may stay the same, but descriptions will have to be changed to show full bonus at 100% instead of perlevel bonus.

As an additional advantage of gradual system i can think of an easier/better way to balance skills and ships. Right now if you have skill that gives you +1 of something per level (ex. controlled drones per lvl), you cannot tweak it without doubling the bonus (which will be OP), but with % based system you can set that skill to give +6 or +7 @100% meaning you'll get +1 of something every ~17(16.67) or 15(14.29)%.
As for players' advantage: lets say you lack 1-2% pg to fit desired ship (even with implants), but dont want to lose a month and train AWU5 for that. With new system you can partially train it and get needed effect while saving on time.

This is not subforum to report bugs/broken things to be fixed, but for features and ideas.

Opinions are like assholes. Everybody got one and everyone thinks everyone else's stinks.
seth Hendar
I love you miners
#15 - 2013-05-16 10:17:55 UTC  |  Edited by: seth Hendar
Kirimeena D'Zbrkesbris wrote:
Ruze wrote:

Your point on partially trained skills is a good one, in my mind. But for many skills the benefit is small, and the math is very clear cut. The machine doesn't have to figure out what percentage of 100 your skill is at every time you fire your turret, it just gives you a 6% increase. Processing wise, it is probably a step or two simpler.

And when 3000 players are in a fight, a step or two per player per turret per shot is a LOT of steps.

Is there much difference in calculating: damage = damage*(1+0.05*skilllvl) and damage = damage*(1+0.25*skill%)? I'd say there is no difference at all.

there is, because skilllvl would be an integer value (so just a DMA operation), while skill% would not, and would need to be calculated on-the-fly thus requiring another set of operation (DMA + FPU operations) just to acquire skill% (because the toon could very well be skilling said skill during the fight, meaning between his first shot, and the 2nd, skill% would have changed

so processing wise, using skill% is, in fact, more ressource consuming than processing with levels
Velicitia
XS Tech
#16 - 2013-05-16 10:41:37 UTC
honestly, something like this *could* be a useful idea ... with a few fixes (OK, so we're scrapping the original idea).

1. still cap the skills at whatever SP they still have (e.g. 1.5 million for Battlecruisers).
2. Divy it up into 5 "levels" still, which just sit at a certain point on the SP bar. e.g. 500 SP for L1.
3. tweak the queue a bit, such that if it's the only skill in the queue (or last skill), you will train straight through from 0 to 1.5 million SP (i.e. L0-5). If you throw in a bunch of skills, then they step one level, and move to the next one in queue, though you can also choose to make it train L1-3 of a skill, then L4 of a different one.

Skill benefits are still given at the integer levels, and switches still leave the partial SP (e.g. you're halfway between L4/5 on Battlecruisers, so you'll sit at 1.25 million SP)

One of the bitter points of a good bittervet is the realisation that all those SP don't really do much, and that the newbie is having much more fun with what little he has. - Tippia
hmskrecik
TransMine Group
Gluten Free Cartel
#17 - 2013-05-16 10:41:58 UTC
Changed quote order to make point.
Quote:

This is not subforum to report bugs/broken things, but for features and ideas.


True but let me repeat: skill system of EVE is in my opinion (and I have reason to believe that for many too) one of those things in the game which work as intended, work well and are quite balanced. To change something like that it's not enough to propose just incremental change, you'd have to present something which blows the old one off the water.

Kirimeena D'Zbrkesbris wrote:

As for players' advantage: lets say you lack 1-2% pg to fit desired ship (even with implants), but dont want to lose a month and train AWU5 for that. With new system you can partially train it and get needed effect while saving on time.

...and this one, sorry, does not. It's quite good argument and I see it's advantage. But it's not enough to demand change in the core gameplay.
Kirimeena D'Zbrkesbris
Republic Military Tax Avoiders
#18 - 2013-05-16 10:54:57 UTC  |  Edited by: Kirimeena D'Zbrkesbris
hmskrecik wrote:
True but let me repeat: skill system of EVE is in my opinion (and I have reason to believe that for many too) one of those things in the game which work as intended, work well and are quite balanced.

Kirimeena D'Zbrkesbris wrote:
While there is no need for such system in EvE, it could provide benefits from partially trained skills.
Sry for quoting myself again, but i do agree that if something works - better to leave it alone. But that doesnt mean that it is a bad idea overall and it provides much more smooth character progression.

hmskrecik wrote:
...and this one, sorry, does not. It's quite good argument and I see it's advantage. But it's not enough to demand change in the core gameplay.

You omitted skill/ship balancing part of argument =\

Velicitia wrote:
honestly, something like this *could* be a useful idea ... with a few fixes (OK, so we're scrapping the original idea).
1. still cap the skills at whatever SP they still have (e.g. 1.5 million for Battlecruisers).
2. Divy it up into 5 "levels" still, which just sit at a certain point on the SP bar. e.g. 500 SP for L1.
3. tweak the queue a bit, such that if it's the only skill in the queue (or last skill), you will train straight through from 0 to 1.5 million SP (i.e. L0-5). If you throw in a bunch of skills, then they step one level, and move to the next one in queue, though you can also choose to make it train L1-3 of a skill, then L4 of a different one.

Skill benefits are still given at the integer levels, and switches still leave the partial SP (e.g. you're halfway between L4/5 on Battlecruisers, so you'll sit at 1.25 million SP)

How is that different from what we currently have?

Opinions are like assholes. Everybody got one and everyone thinks everyone else's stinks.
Velicitia
XS Tech
#19 - 2013-05-16 11:09:05 UTC
Kirimeena D'Zbrkesbris wrote:

Velicitia wrote:
honestly, something like this *could* be a useful idea ... with a few fixes (OK, so we're scrapping the original idea).
1. still cap the skills at whatever SP they still have (e.g. 1.5 million for Battlecruisers).
2. Divy it up into 5 "levels" still, which just sit at a certain point on the SP bar. e.g. 500 SP for L1.
3. tweak the queue a bit, such that if it's the only skill in the queue (or last skill), you will train straight through from 0 to 1.5 million SP (i.e. L0-5). If you throw in a bunch of skills, then they step one level, and move to the next one in queue, though you can also choose to make it train L1-3 of a skill, then L4 of a different one.

Skill benefits are still given at the integer levels, and switches still leave the partial SP (e.g. you're halfway between L4/5 on Battlecruisers, so you'll sit at 1.25 million SP)

How is that different from what we currently have?


Effectively, it's just "if no new skill in the queue, then auto-train the next level of the last skill you were training, provided you weren't training L5 already" ... just pre-caffeine Cool

One of the bitter points of a good bittervet is the realisation that all those SP don't really do much, and that the newbie is having much more fun with what little he has. - Tippia
Inna Cristiana
GucciGang
#20 - 2013-05-16 11:10:23 UTC
This sounds silly. Sorry.
12Next page