No xp for killing things - but xp rewards for quests?

[Bruma Hobo]

Fuck GDrama, this thread is why I love the codex

 

[DraQ]

Each skill could be different, in smithy for example you could divide the making process into several parts, making a simple knife could require three steps, one to make the blade, then the handle and then the third step would be to assembly those two.
If you fail at the first you get nothing, if you succeed then you get the blade, failing at the second won't make you lose the blade, you just don't get the handle, you get the handle with a success.
Failing the third part of the process could do nothing or could break one of the two pieces, or break both.
If you know that something is very complex or time consuming you could break it into several other parts, like if you want to make that 2 meters sword, to make the blade would require 3 checks and each one would require you to spend 3 hours trying.

Knife example is good, but what good is a third of the blade in the sword example?

If you want to make the summon remain indefinitely and survive the battle then you could put checks to use your summoning spell to actually regenerate the summon after battle.
Or if your summon gets ready to level up or something you would have to use your summon spell again to make it better. Just throwing some random ideas here.

But you will still use those spells rarely compared to direct damage spells, or, especially, direct damage attacks.

Why make an assumption that all kinds of activities can be optimally broken into chunks of the exact same size?

 

[tiagocc0]

Knife example is good, but what good is a third of the blade in the sword example?

In the case of the 2 meters blade, if you divide it in 3 steps to make 1 blade, failing on any of them would ruin the blade.

But you will still use those spells rarely compared to direct damage spells, or, especially, direct damage attacks.

Well, there's a subject we haven't touched yet and that's training. It's like if the fantasy were real and we were there living it.
How would a person who throws fireballs get better at it and how would a person who summons creatures get better at it?
Would they only use their skills at combat? No, they would probably train at every chance they got and/or when they are alone.
I'm not going to get into detail here because I haven't given much thought about this.

Why make an assumption that all kinds of activities can be optimally broken into chunks of the exact same size?

It's like thinking about perfect games, how would they be?
A perfect RPG could be one you can be anything you want and still be able to enjoy the story. Like you can role play anything and it wouldn't ruin the game.
I know of games like Skyrim that abuse on how much a skill increases because they hadn't enough time to do something better or because they simply didn't want to.
For me then the perfect use base system would have the same rate for all skills.
It's like something that can't be reached, like never, but still it's what I would use as a goal. It's what drives me toward a better system.

 

[DraQ]

In the case of the 2 meters blade, if you divide it in 3 steps to make 1 blade, failing on any of them would ruin the blade.

Then what's the point? Why not make or break it in one go? Hey, you can even randomize the fraction of necessary time spent when failing which will allow you to mess up at any point, rather than just at the end of pre-determined steps.

Well, there's a subject we haven't touched yet and that's training. It's like if the fantasy were real and we were there living it.
How would a person who throws fireballs get better at it and how would a person who summons creatures get better at it?
Would they only use their skills at combat? No, they would probably train at every chance they got and/or when they are alone.

But since summoning is more time/resource intensive a summonner would train less.

If both, maxed out summoning and maxed out destruction are defined as skill levels achieved by dedicating your life time to perfecting the skill, then at the same, maxed out level the summonner will have used his skill much less times than destructive caster.

It's like thinking about perfect games, how would they be?

No, it's like making pointless mechanical assumptions then trying to wrap the in-universe logic around them rather than working out the logic, then building up the mechanics to reflect it. In other words - ass backwards.

For me then the perfect use base system would have the same rate for all skills.

Why would this make it more perfect?

 

[tiagocc0]

Then what's the point? Why not make or break it in one go? Hey, you can even randomize the fraction of necessary time spent when failing which will allow you to mess up at any point, rather than just at the end of pre-determined steps.

That's a great idea, the point is to pass the message that it's a complicated/complex piece to be made, you could randomize the time and quantity of times necessary to make the piece, a master would always make it on one go, a novice could screw up but instead of ruining the piece he could make another try, but each time would decrease the piece quality.

But since summoning is more time/resource intensive a summonner would train less.

If both, maxed out summoning and maxed out destruction are defined as skill levels achieved by dedicating your life time to perfecting the skill, then at the same, maxed out level the summonner will have used his skill much less times than destructive caster.

If it's something complicate it could require more trained, or you could divide in training and study, training where you spend resources and study where you don't spend resources, summoning then would require a great amount of study a maybe a not so great amount of training.

No, it's like making pointless mechanical assumptions then trying to wrap the in-universe logic around them rather than working out the logic, then building up the mechanics to reflect it. In other words - ass backwards.

If it were pointless then the idea would have died out right at the start, instead we had a nice chat about smithy and summoning and how they could be improved. Regardless of the rate being equal or not.

Why would this make it more perfect?

I don't know, I just know that the overuse of the rate system with some skills made too abstract and with increased rate can make the game boring, so my assumption is that going to the other direction could make the game more enjoyable.
Just my 2 cents on how I would like to improve a use based system.

 

[Johannes]

If you put interesting bits all over then the user will want to spend time on it.
Like you would have several checks made until you have a sword forged, each check could fail, have success or have an great success where the player gets a bonus.
Be it a bonus about improving it's skill, a bonus to the item or maybe an idea for a new item.

Having multiple checks is neat... as long as they do something a single check can't. So what would be the point of those extra checks? What would happen if I failed the first, but succeeded at the second and how would it be different than the other way around or failing both?

Have the created items quality depend on how many skillchecks you pass, for example. That way you get more detailed result than just a binary failure-success. But it doesn't really need to be more detailed (creating those hilts, blades, whatever, who cares) than just giving the item a good or bad modifier depending on the skillchecks.

 

[DraQ]

Then what's the point? Why not make or break it in one go? Hey, you can even randomize the fraction of necessary time spent when failing which will allow you to mess up at any point, rather than just at the end of pre-determined steps.

That's a great idea, the point is to pass the message that it's a complicated/complex piece to be made, you could randomize the time and quantity of times necessary to make the piece, a master would always make it on one go, a novice could screw up but instead of ruining the piece he could make another try, but each time would decrease the piece quality.

Thanks, but if the process isn't interactive nor does it allow for different results, then you don't pass any message, whether you have one check or over nine thousand.

Retries are a nice idea, of course that requires a quality system of some sort.

If it's something complicate it could require more trained, or you could divide in training and study, training where you spend resources and study where you don't spend resources, summoning then would require a great amount of study a maybe a not so great amount of training.

But then you need study mechanics (which would be nice), and there is still not guarantee that it will actually help here, because if study doesn't waste resources, then it's naturally subject to grinding.

If anything I'd make study responsible only for learning completely new skills and gaining skill perks, whereas practice would perfect those skills and perks. New perks could also be rarely gained randomly when practicing, based on intelligence. This kind of perk would probably work best if it didn't count towards learning penalty caused by accumulation of perks and skills (to prevent JoAT), but wouldn't confer new skills, only perks for existing ones.

If it were pointless then the idea would have died out right at the start

Many pointless ideas don't alas. Yes, our chat is nice and productive, but at best tangential to the problem at hand - that assumption that each and every skill ever realistically advances at same rate per 'use' is spurious and completely unwarranted.

Why would this make it more perfect?

I don't know, I just know that the overuse of the rate system with some skills made too abstract and with increased rate can make the game boring, so my assumption is that going to the other direction could make the game more enjoyable.

Expound.

 

[tiagocc0]

Thanks, but if the process isn't interactive nor does it allow for different results, then you don't pass any message, whether you have one check or over nine thousand.

It allows for different result if you put failure, exceptional failure, success, exceptional success and the possibility to learn something new at each try.

Retries are a nice idea, of course that requires a quality system of some sort.

Yep.

But then you need study mechanics (which would be nice), and there is still not guarantee that it will actually help here, because if study doesn't waste resources, then it's naturally subject to grinding.

If anything I'd make study responsible only for learning completely new skills and gaining skill perks, whereas practice would perfect those skills and perks. New perks could also be rarely gained randomly when practicing, based on intelligence. This kind of perk would probably work best if it didn't count towards learning penalty caused by accumulation of perks and skills (to prevent JoAT), but wouldn't confer new skills, only perks for existing ones.

You could have to study to be able get somethings you wouldn't only from training.
Set a limit to study that is proportional to your current skill amount. So if you have 10 points you could study for 10 hours, after studying for 10 hours you can no longer study or the study won't affect you anymore, after you get more 5 skill points you get to spend 5 more hours of study if you want.

Many pointless ideas don't alas. Yes, our chat is nice and productive, but at best tangential to the problem at hand - that assumption that each and every skill ever realistically advances at same rate per 'use' is spurious and completely unwarranted.

What I really hate is that complicated skills increase faster, this as unreal as you can get. And this happens because the developer didn't want to spend more time constructing an interesting gameplay mechanics for the skill in question.

Expound.

Skyrim smithy skill.
One button pressed equals one item and a lot of skill increase.

I suck at playing games like Skyrim but I was able to get 100 in blacksmithy early in the game.
And I was always expecting it to be great, but it was exactly the same when I was a novice to when I was a master. It sucked.

 

[DraQ]

Thanks, but if the process isn't interactive nor does it allow for different results, then you don't pass any message, whether you have one check or over nine thousand.

It allows for different result if you put failure, exceptional failure, success, exceptional success and the possibility to learn something new at each try.

But you don't need multiple checks for degrees of success either and if you just make several checks forging a 2m sword it's a waste of computational effort if the result is the same - either you end up with a blade or a piece of scrap metal.

Multiple checks are only called for if you can fail or succeed differently depending on your input, random factors or environmental influence.

You could have to study to be able get somethings you wouldn't only from training.
Set a limit to study that is proportional to your current skill amount. So if you have 10 points you could study for 10 hours, after studying for 10 hours you can no longer study or the study won't affect you anymore, after you get more 5 skill points you get to spend 5 more hours of study if you want.

If the system was perk based I wouldn't make study give any sort of numerical skill progress, just perks and new abilities. They could influence influence numerical skill progress, but preferably indirectly - for example practising skill with a basic spell or attack technique could make you proficient enough at it so that it always succeeds and is effective on susceptible low level mooks, but completely ineffective against high level foes (capable of deflecting or otherwise neutralizing it if it's a spell, or flawlessly counter melee attack of that type in normal circumstances), in either case you'd gain no skill this way (failure/success probability based skill XP). You'd have to learn and use something new despite there would be no explicit requirement for some perks or abilities to advance the skill further.

Expound.

Skyrim smithy skill.
One button pressed equals one item and a lot of skill increase.

Except that's not the problem with ease of use - unless you propose a crafting system working differently than "select stuff, then commit", any semi-reasonable interface (in terms of general idea) will allow you to produce items in batches.

The problem is that there is no cost associated with use of this skill, no failure preventing natural implementation of difficulty dependant XP scaling, and no cost for this failure.

If you can advance your skill by mass producing shitty daggers, then the one button mechanics is the least of your problems.

If you can only advance the skill by producing shitty daggers in narrow skill range and materials are rare, costly and may be used in potentially more beneficial manner, then you won't grind armorer skill unless it's a big part of your character's build.

Also, for every skill gain, any subsequent skill gain in any skill should be made more difficult to reflect character's limited capacity for learning.

 

[tiagocc0]

But you don't need multiple checks for degrees of success either and if you just make several checks forging a 2m sword it's a waste of computational effort if the result is the same - either you end up with a blade or a piece of scrap metal.

Multiple checks are only called for if you can fail or succeed differently depending on your input, random factors or environmental influence.

I just said that you can have "failure, exceptional failure, success, exceptional success and the possibility to learn something new at each try", which means that exceptional success could mean bonuses for the blade and learn something new like a new recipe for another weapon is always nice.

If the system was perk based I wouldn't make study give any sort of numerical skill progress, just perks and new abilities. They could influence influence numerical skill progress, but preferably indirectly - for example practicing skill with a basic spell or attack technique could make you proficient enough at it so that it always succeeds and is effective on susceptible low level mooks, but completely ineffective against high level foes (capable of deflecting or otherwise neutralizing it if it's a spell, or flawlessly counter melee attack of that type in normal circumstances), in either case you'd gain no skill this way (failure/success probability based skill XP). You'd have to learn and use something new despite there would be no explicit requirement for some perks or abilities to advance the skill further.

I see no problem with this approach. But what I said was that study wouldn't give you skill points, instead skill points that you acquired would give you the ability to study the subject, the studying could give you a perk or new abilities like you said.

Except that's not the problem with ease of use - unless you propose a crafting system working differently than "select stuff, then commit", any semi-reasonable interface (in terms of general idea) will allow you to produce items in batches.

The problem is that there is no cost associated with use of this skill, no failure preventing natural implementation of difficulty dependant XP scaling, and no cost for this failure.

If you can advance your skill by mass producing shitty daggers, then the one button mechanics is the least of your problems.

If you can only advance the skill by producing shitty daggers in narrow skill range and materials are rare, costly and may be used in potentially more beneficial manner, then you won't grind armorer skill unless it's a big part of your character's build.

Also, for every skill gain, any subsequent skill gain in any skill should be made more difficult to reflect character's limited capacity for learning.

I don't mind making all skills harder for every skill gain, or making materials rare.
What I don't like is having a skill increase a lot faster just because the developer was lazy. And the main point is a shitty one click design without failure.

Producing item in batches is horrible, what kind of fantasy world is that where you mass produce items in a blacksmith?
This should only be allowed if you have a skill that is far superior to the item you're mass producing, like a master mass producing swords but that still can't mass produce those 2 meter enchanted blades of doom.

 

[DraQ]

I see no problem with this approach. But what I said was that study wouldn't give you skill points, instead skill points that you acquired would give you the ability to study the subject, the studying could give you a perk or new abilities like you said.

If you'd study the subject, you'd study for specific perk anyway so you generally wouldn't study exact same thing ever again, though individual perks might require certain skill levels.

What I don't like is having a skill increase a lot faster just because the developer was lazy.

Then what about weapons that swing at different rate and make different amounts of damage?
Should halberd be much harder to master than sword, just because halberd swings slower and rarely requires you to swing again?
Should crossbow be harder to master than bow because of lower rate of fire, exactly opposite to how it works IRL?

And the main point is a shitty one click design without failure.

I have stressed it many times that failure mechanics is mandatory for learn by doing system.

Producing item in batches is horrible, what kind of fantasy world is that where you mass produce items in a blacksmith?
This should only be allowed if you have a skill that is far superior to the item you're mass producing, like a master mass producing swords but that still can't mass produce those 2 meter enchanted blades of doom.

Tough luck, because being able to produce items in batches is not a cause but sympthom. There is no reason why good interface shouldn't allow that other than trying to implement difficulty through interface, but then you can still make items in batches, you just have to click a lot more and making something harder through unfriendly interface is shitty idea anyway.
You'd better come up with crafting mechanics that is interesting (not something you simply learn by rote and repeat) and specifically precludes making stuff in batches. Any ideas of a crafting minigame which does not suck?

 

[tiagocc0]

If you'd study the subject, you'd study for specific perk anyway so you generally wouldn't study exact same thing ever again, though individual perks might require certain skill levels.

Agreed.

Then what about weapons that swing at different rate and make different amounts of damage?
Should halberd be much harder to master than sword, just because halberd swings slower and rarely requires you to swing again?
Should crossbow be harder to master than bow because of lower rate of fire, exactly opposite to how it works IRL?

My idea is to have everything evolving at the same pace, not by making something that is less used a lot easier to evolve, actually making it evolving much faster in a unrealistic way. (Games are not realistic, but we still would like some level of realism to make it more enjoyable).
If weapons have different rates of fire, than you probably doesn't want to use each swing you make a way to earn a fixed amount of skill, you could use damage instead.
If a weapons does more damage it will likely be slower. If it's slow and causes low damage you would have to use it more to kill your enemies.
But killing a enemy would give the same amount of skill to any weapon.
See, you are not leveling the rates yourself, you are using another game mechanic to do so.

What I said that started all this conversation was that I do not like something leveling up so fast that it feels unreal.
I don't want to put in a game 'sword increases at 1.0 each each hit, maces increases 1.4, daggers increase 0.7, blacksmith increases 3.5 each item'.
That's what I felt happened at Skyrim. I'm bashing this approach because it's the easiest approach and it ruined Skyrim for me.
Of course a well balanced game could very well be using this system, but I as the player would not know or feel like it was being used like I did in Skyrim.
What I know is that I as a developer would avoid this kind of system, if it can't be avoided completely I would at least make it less noticeable to the player.

I have stressed it many times that failure mechanics is mandatory for learn by doing system.

And I agree with it.

Tough luck, because being able to produce items in batches is not a cause but sympthom. There is no reason why good interface shouldn't allow that other than trying to implement difficulty through interface, but then you can still make items in batches, you just have to click a lot more and making something harder through unfriendly interface is shitty idea anyway.
You'd better come up with crafting mechanics that is interesting (not something you simply learn by rote and repeat) and specifically precludes making stuff in batches. Any ideas of a crafting minigame which does not suck?

Exactly, mass production is needed, but should be exploited.
It doesn't need to be a minigame, but could be a level of skill, or a perk that is required for you to be able to use that interface that mass produces items.

 

[DraQ]

Then what about weapons that swing at different rate and make different amounts of damage?
Should halberd be much harder to master than sword, just because halberd swings slower and rarely requires you to swing again?
Should crossbow be harder to master than bow because of lower rate of fire, exactly opposite to how it works IRL?

My idea is to have everything evolving at the same pace, not by making something that is less used a lot easier to evolve, actually making it evolving much faster in a unrealistic way. (Games are not realistic, but we still would like some level of realism to make it more enjoyable).

But different stuff is *used* at different pace.

If weapons have different rates of fire, than you probably doesn't want to use each swing you make a way to earn a fixed amount of skill, you could use damage instead.

But that's even dumber, because there is no logical relation between damage dealt and skill progress.
Hell, for example during sparring you will deal very little damage, but can learn a lot, certainly more than when skewering stronger enemy with desperate attack (a freak success meaning a lot of skill points and a lot of damage dealt).
Besides, I'd like to abandon notion of damage as single value altogether along with HPs.

Finally, if I'm shooting an assault rifle in single shot mode I'll probably learn more about firing accurately than when spraying in full auto, despite dealing less damage.
Same with firing a sniper rifle (single target, high damage, hard to use effectively) VS grenade launcher (AoE attack, very high damage to multiple targets, relatively easy to use), or sniper rifle VS a machinegun using the same type of ammo.

What I know is that I as a developer would avoid this kind of system, if it can't be avoided completely I would at least make it less noticeable to the player.

Well, of course such system shouldn't be used as cop-out and noticeable stuff is usually the stuff that isn't working properly, but this is more about not using mechanics unless it's necessary rather than declaring a priori that it shouldn't be used.

It doesn't need to be a minigame, but could be a level of skill, or a perk that is required for you to be able to use that interface that mass produces items.

So the players will still mass produce items without perk, but will have to click more and generally be annoyed by clunky interface? An excellent idea.

You see the problem here is that with crafting mechanics consisting of choosing components and item to craft, then clicking 'commit' is that being able to do the same again is not even a logical extension of the interface, but logical core functionality.

It can be mitigating from resource management level, by making resources and time valuable enough for pointless activities to be not worth it, but to fight it from level of crafting mechanics itself is more difficult.

First, you need to create crafting mechanics that will be something more than what I described above.
Second, this mechanics will have to be interactive. Since unlike stuff like stealth and combat that is intertwined with variety of other mechanics (for example you need to pick a lock, but there are guards patrolling around, so you have to do this quickly and silently - good lockpicking or hacking mechanics should NEVER pause the game proper) crafting is kind of isolated activity due to being performed in controlled environment it will be a minigame completely disjoint from the game proper and will not be able to rely on outside disturbances to make it interesting, it will have to be interesting on its own. Now, the problem with any mechanics, is twofold:
-it needs to bear some relation to the activity it handles
-it needs to be interesting enough to warrant interactivity. To be interesting a mechanics must introduce some degree of unpredictability requiring making decisions on the fly rather than in advance. This works with combat (if it's any good), and this may work with even very simple mechanics if they are perturbed by outside factors. I honestly can't see it happening with smithing, so the question is: Can you make a smithing minigame that would remain nontrivial and will prove impossible to learning by rote? If so, how?

 

[tiagocc0]

But different stuff is *used* at different pace.

Just because you use something less than something else doesn't mean it has to increase faster.
So just because the player is more likely to use swords than crossbows you 'have to' increase the crossbows rate?
So if a player just uses crossbows and another just uses swords, the player with the crossbow will be benefited because the system makes it easier for him to develop the skill while the player who uses only swords will have to spend more time to achieve the same skill level killing the same enemies.

But that's even dumber, because there is no logical relation between damage dealt and skill progress.
Hell, for example during sparring you will deal very little damage, but can learn a lot, certainly more than when skewering stronger enemy with desperate attack (a freak success meaning a lot of skill points and a lot of damage dealt).
Besides, I'd like to abandon notion of damage as single value altogether along with HPs.

Finally, if I'm shooting an assault rifle in single shot mode I'll probably learn more about firing accurately than when spraying in full auto, despite dealing less damage.
Same with firing a sniper rifle (single target, high damage, hard to use effectively) VS grenade launcher (AoE attack, very high damage to multiple targets, relatively easy to use), or sniper rifle VS a machinegun using the same type of ammo.

It's not dumber, you're going to use the damage that the enemy receives not the damage the weapon makes.
So if you get 10 skill points by killing an enemy with a sword, why would you give more experience or less experience to someone using a crossbow killing the same enemy?
Sparring causes damage as well, not real damage as to hurt the enemy, but you have to have a way to measure your performance, thus a damage that does not subtract from the enemy health, bingo, sparring equals the same as a real battle.

The concept of a grenade or a fireball depends, on games that give exp for killing an enemy this acceptable to give skill based on damage done to enemies, if this is not acceptable to you then the damage system does not applies, that simply, it just works like most games.
Using grenade launchers is not a easy feat, if you choose to do so you're at the penalty of killing yourself, you won't be able to fire at close ranges, your rate of fire is probably slow as hell so you have to hit your enemy or you will probably die right after and to be able to hit lots of enemies at the same time is something that requires great timing and accuracy to spot the place where you would cause the most damage. Generally grenade launchers are not used to kill enemies, but to cripple them.
So thinking like this I think it's acceptable to give exp based on the damage done. Because even if you shot it in middle of five enemies where 1 would die, the second would get 70% damage, the third and fourth would get 40% and the last 25% damage, you would get lots of exp at the expense of using a weapon that could get you killed easily. If grenade launchers were really that useful everybody would only use them instead of other weapons.

Well, of course such system shouldn't be used as cop-out and noticeable stuff is usually the stuff that isn't working properly, but this is more about not using mechanics unless it's necessary rather than declaring a priori that it shouldn't be used.

Sorry, I didn't get this one.
My point was that I don't like it because of the games where I played and this system was used it ruined the game for me. So I don't like it, I wouldn't use unless necessary and I certainly wouldn't abuse it.
Am I being that wrong? What do want to convince me to? That there's nothing wrong with it or that even though there's some flaws it is better than most systems?

So the players will still mass produce items without perk, but will have to click more and generally be annoyed by clunky interface? An excellent idea.

Oh sure, at the beginning of game you have a weak character, but at the end you are strong and can kill the last boss.
So just because latter you're able to kill the boss doesn't mean I should let the player at the beginning of the game kill it.
Just because you can produce lots of items doesn't mean I should facilitate it to you right at the beginning, you have to fight your way trough the game to be able to have it easier.
You could say that the player is producing items at the forest so he can't mass produce, but if he have access to a forge later in the game he could mass produce.
You could say that producing an item is very tiring so the player can only produce 3 items a day, but later you get a skill that let's you produce 30 so that mass produce button appears in the interface.
There are lots and lots of ways to make it more interesting than always giving the player the easiest way to do everything in the game.

You see the problem here is that with crafting mechanics consisting of choosing components and item to craft, then clicking 'commit' is that being able to do the same again is not even a logical extension of the interface, but logical core functionality.

It can be mitigating from resource management level, by making resources and time valuable enough for pointless activities to be not worth it, but to fight it from level of crafting mechanics itself is more difficult.

First, you need to create crafting mechanics that will be something more than what I described above.
Second, this mechanics will have to be interactive. Since unlike stuff like stealth and combat that is intertwined with variety of other mechanics (for example you need to pick a lock, but there are guards patrolling around, so you have to do this quickly and silently - good lockpicking or hacking mechanics should NEVER pause the game proper) crafting is kind of isolated activity due to being performed in controlled environment it will be a minigame completely disjoint from the game proper and will not be able to rely on outside disturbances to make it interesting, it will have to be interesting on its own. Now, the problem with any mechanics, is twofold:
-it needs to bear some relation to the activity it handles
-it needs to be interesting enough to warrant interactivity. To be interesting a mechanics must introduce some degree of unpredictability requiring making decisions on the fly rather than in advance. This works with combat (if it's any good), and this may work with even very simple mechanics if they are perturbed by outside factors. I honestly can't see it happening with smithing, so the question is: Can you make a smithing minigame that would remain nontrivial and will prove impossible to learning by rote? If so, how?

I honestly can't see it happening with a smithing minigame.
When I played Diablo 1, I played it most for two reasons, being able to find magical items and fighting those random bosses.
When I said that you can divide the smithing process, you could do so that each part of the process could create bad, normal or great results, you could learn something during the process, you could maybe ending up with a magical item.
You have to pay attention to every part because every part could have a result that the player wishes, thus the player would not automatically want to mass produce items.
You could even say that you have two methods, if you mass produce you won't get the chance to learn something new or at least it's probability is a lot lower and you won't get great, magical items from it.
If you choose to make the item individually then you get all those bonuses we talked about.

 

[DraQ]

But different stuff is *used* at different pace.

Just because you use something less than something else doesn't mean it has to increase faster.
So just because the player is more likely to use swords than crossbows you 'have to' increase the crossbows rate?
So if a player just uses crossbows and another just uses swords, the player with the crossbow will be benefited because the system makes it easier for him to develop the skill while the player who uses only swords will have to spend more time to achieve the same skill level killing the same enemies.

Well, crosbows do fire slowly (unless they are those Chinese repeater crossbow or Greek chaingun ballistae) and are easier to fire than bows.

Another solution would be to split skills vertically, so that crossbow and bow would use the same base skill for aiming, but would have different skills for actual operation, with bow's additional skill also influencing accuracy, while crossbow's only affecting reload time and stamina expenditure, but in such case crossbow should increase the shared skill at even slower rate than it would based on ROF alone.

It's not dumber, you're going to use the damage that the enemy receives not the damage the weapon makes.

It is dumber, because damage dealt in no way reflects difficulty of a task.

If you attack a skilled swordsman in frontal combat, you're unlikely to deal much damage with any given attack, but if you survive you will probably learn a lot. If you stab him in the back or while he's down with the exact same weapon, you will deal a lot of damage, but won't learn much.

Additionally basing learning on damage dealt removes ability to learn from your mistakes.

So if you get 10 skill points by killing an enemy with a sword, why would you give more experience or less experience to someone using a crossbow killing the same enemy?

Because crossbow is a point & shoot weapon that is relatively simple to use while going against the guy with a sword and surviving is likely to require much skill.

The concept of a grenade or a fireball depends, on games that give exp for killing an enemy this acceptable to give skill based on damage done to enemies, if this is not acceptable to you then the damage system does not applies, that simply, it just works like most games.
Using grenade launchers is not a easy feat, if you choose to do so you're at the penalty of killing yourself, you won't be able to fire at close ranges, your rate of fire is probably slow as hell so you have to hit your enemy or you will probably die right after and to be able to hit lots of enemies at the same time is something that requires great timing and accuracy to spot the place where you would cause the most damage. Generally grenade launchers are not used to kill enemies, but to cripple them.
So thinking like this I think it's acceptable to give exp based on the damage done. Because even if you shot it in middle of five enemies where 1 would die, the second would get 70% damage, the third and fourth would get 40% and the last 25% damage, you would get lots of exp at the expense of using a weapon that could get you killed easily. If grenade launchers were really that useful everybody would only use them instead of other weapons.

Grenade launchers are unselective, imprecise, useless in close quarters (unless packing some sort of buckshot loads and working as huge ass shotguns) and need bulky, expensive ammo.

Still, they don't require nearly as much skill to hit the target, only basics of operation and knowing how to use the ladder/quadrant sight, since you have about 2-5m margin of error, while with sniper rifle you may not have even 5cm margin of error. Yes, all weapons have effective min and max ranges, but duh. Also, grenades launched from GL generally need to travel certain minimum distance (usually measured by number of rotations) in order to arm themselves, so you are unlikely to kill yourself by by banging grenade against nearby obstacle rather than distant target.

Anyway, the main problem is that experience should be based on difficulty of the task you've set to achieve (you learn more by failing tasks you're good at and succeeding at tasks you're bad at, than vice versa), not on damage dealt. With ranged weapon the task is either hitting the target or critting the target and given that game has limited amount of means to tell what you were actually shooting at, it should just take any valid target if hit, and do nothing or try to 'guess' based on where you aimed the weapon if you missed. Difficulty can be roughly calculated from precalculated (because it's not much data but needs much calculations to obtain, so it's better to just cache it in advance after doing hitboxes) average size of target's hitbox and distance, modified by weapon's own stats.

You don't learn anything directly from dealing damage, you learn from hitting the target or subtargets (like vital organs) inside it.

Sorry, I didn't get this one.

If you get mugged by a black dude it's ok to refer to him as filthy nigger but not okay to refer to all black dudes in this way based on this single event.

Replace this particular black dude with 'Skyrim', general population of black dudes with 'all systems using different skill levelling rates' and mugging with cop-out/sloppy implementation.

Oh sure, at the beginning of game you have a weak character, but at the end you are strong and can kill the last boss.
So just because latter you're able to kill the boss doesn't mean I should let the player at the beginning of the game kill it.

Except you can still kill the boss in this case, except you just have to click more.

If I have basic skill to make an item and appropriate tools, then if I can make one item, I can just as well make thirty if resources allow. Making me click more to make those thirty items than I absolutely have to is shitty interface.

You could say that producing an item is very tiring so the player can only produce 3 items a day, but later you get a skill that let's you produce 30 so that mass produce button appears in the interface.

But why hide the button if you already limit it with a resource in the form of PC's stamina?

Besides, it's not 'mass produce' button, it's normal 'ok' button, except you click it repeatedly, because good interface design dictates that components/resources stay in their respective slots so that player doesn't need to select them all over again. It has nothing to do with character or gameworld.

The thing is that you can't reasonably curb this possibility with interface as it isn't the problem with interface and one button item making.

The problem is that such activity is desirable.
If you curb this desirability, you will curb the activity even if the interface makes it just as easy to make over 9000 shitty daggers.

For example you could make crafting skills consist of large number of knowledge based perks with few levels each making repetition pointless. You could make resources expensive and bulk quantities of items difficult to sell to make it unafordable. And you could make improvement in any skill cost you potential of improvement in all the others making player not engage in activity not fitting their build and tactics.

I honestly can't see it happening with a smithing minigame.
When I played Diablo 1, I played it most for two reasons, being able to find magical items and fighting those random bosses.
When I said that you can divide the smithing process, you could do so that each part of the process could create bad, normal or great results, you could learn something during the process, you could maybe ending up with a magical item.
You have to pay attention to every part because every part could have a result that the player wishes, thus the player would not automatically want to mass produce items.
You could even say that you have two methods, if you mass produce you won't get the chance to learn something new or at least it's probability is a lot lower and you won't get great, magical items from it.
If you choose to make the item individually then you get all those bonuses we talked about.

But then there is no point in making items individually more involved affair than a button press after selecting components and no point in artificial, interface level separation between mass and non-mass produced item.

Because nothing unexpected can happen that would require constant monitoring during crafting, so you can just frontload all the information and click the single button to commit.

 

[tiagocc0]

Another solution would be to split skills vertically, so that crossbow and bow would use the same base skill for aiming, but would have different skills for actual operation, with bow's additional skill also influencing accuracy, while crossbow's only affecting reload time and stamina expenditure, but in such case crossbow should increase the shared skill at even slower rate than it would based on ROF alone.

Very interesting, I haven't thought about that.

Also, grenades launched from GL generally need to travel certain minimum distance (usually measured by number of rotations) in order to arm themselves, so you are unlikely to kill yourself by by banging grenade against nearby obstacle rather than distant target.

If you want to get that technical you must know that there are several types of grenades.

Replace this particular black dude with 'Skyrim', general population of black dudes with 'all systems using different skill levelling rates' and mugging with cop-out/sloppy implementation.

Then please name a few games that are better using this kind of system so I can play them.


From the rest of the text we are just repeating ourselves so I will leave at that.

 

[DraQ]

Another solution would be to split skills vertically, so that crossbow and bow would use the same base skill for aiming, but would have different skills for actual operation, with bow's additional skill also influencing accuracy, while crossbow's only affecting reload time and stamina expenditure, but in such case crossbow should increase the shared skill at even slower rate than it would based on ROF alone.

Very interesting, I haven't thought about that.

Well, it's an extension of attribute->skill->perks logical structure into attribute->skill->skill->skill.... with perks branching off. You can even blur the distinction between perks and skills.

Also, grenades launched from GL generally need to travel certain minimum distance (usually measured by number of rotations) in order to arm themselves, so you are unlikely to kill yourself by by banging grenade against nearby obstacle rather than distant target.

If you want to get that technical you must know that there are several types of grenades.

Yeah, but delayed arming mechanism is typical safety measure against blowing yourself and your squad with a GL. Now, an RPG is actually dangerous to use but it depends on situational context (squadmates behind, obstacles behind) rather than on skill. A retard with an RPG is unlikely to blow himself up in the open, a skilled RPG user won't be safer than retard IF he tries to fire it from a cramped closed space like a bunker or car with other windows up.

Then please name a few games that are better using this kind of system so I can play them.

Name a few games with well implemented use based, name a few RPGs with non-shitty level mechanics, etc.

Even if all the attempts to date failed, if I can make something workable on paper it will work unless you can show me where does it fail in practice.

If some gameplay mechanics, be it use based or skill dependent levelling rate sucks, it's suckage is due to particular reasons, not because all previous attempts also sucked.

Shit, most RPGs to this date, including dumbed down modern cinematic shooters/slashers with awesome buttons AND all monocled codexian RPGs past mid nineties are burdened with shitty mechanics completely unsuited for the PC platform that traces back to PnP RPGs. Does that mean that all cRPGs must suck and developing a good cRPG is an unrealistic goal?

From the rest of the text we are just repeating ourselves so I will leave at that.

 

[tiagocc0]

Well, it's an extension of attribute->skill->perks logical structure into attribute->skill->skill->skill.... with perks branching off. You can even blur the distinction between perks and skills.

I will definitely try to use this on my game. Even though it's a turn based 4x space game, I'm planning on having leaders that you're able to 'raise' that will have skills/perks.

Name a few games with well implemented use based, name a few RPGs with non-shitty level mechanics, etc.

Even if all the attempts to date failed, if I can make something workable on paper it will work unless you can show me where does it fail in practice.

If some gameplay mechanics, be it use based or skill dependent levelling rate sucks, it's suckage is due to particular reasons, not because all previous attempts also sucked.

Shit, most RPGs to this date, including dumbed down modern cinematic shooters/slashers with awesome buttons AND all monocled codexian RPGs past mid nineties are burdened with shitty mechanics completely unsuited for the PC platform that traces back to PnP RPGs. Does that mean that all cRPGs must suck and developing a good cRPG is an unrealistic goal?

I just thought you had at least a few games that even tough are not perfect they shine on one area or another.


I feel the same about turn based 4x space games.

 

[Major_Blackhart]

Haven't read the thread lately, but here's a thought that just popped into my mind, not that I don't mind the idea of getting rid of XP for combat altogether.
XP for combat is useable only for combat skills (i.e. you can only use experience gained from killing someone to improve your guns or something).

Either way though, it's not too big of a deal.