You don't seem to consider the option of not using the same actual proc effect for attacks with wildly different speeds or other properties.

By this I mean you have two different "set the target on fire" effects, one with a short duration and one with a long duration; you use the short duration one on weapons with high attack speed and the long duration one on weapons with low attack speed. For example.

This achieves approximately the same goal as trying to normalize the value of the effects by scaling them based on attack speed, but is potentially easier to tweak since you aren't playing with curves applied to many attacks just to try to balance one attack that's a bit out of the power scope. It also lets you make very targeted tweaks to individual skills, and it can presented to the player as the effect ("set on fire"); they don't need to know it's actually two different effects in content.

The downside is that it requires you to build and manage more content, so there's extra complexity there. But personally I think that's an easier problem to tackle than trying to find good scalability curves across so many factors. Is there a reason you don't want to take this approach?

Regardless, I'd suggest you build some kind of spreadsheet or similar mechanism that can give you an overview of the concrete data that goes into balancing this sort of thing. For example, something that applies your calculations on all skills over a certain number of attempts and gives you a concrete "percentage the proc occurred" value and information about damage and damage modification et cetera so you can experiment with various approaches and see how they impact the overall landscape.

I also think it's important to keep in mind that "balancing" something by making all possible choices equivalent in power isn't always the best; it reduces the impact of each choice to essentially a decision about whether or not you like red particles (fire) or blue particles (ice).

Thank you for your answer Josh. I cannot really change the proc type based on the cast-time of the skill for a few pratical reasons. You are proposing something very binary (big proc, small proc) which will not work with my game. First, it promotes some wierd min-maxing weapon choice where you can find find the best value of proc per second and proc strength and choose the weapon accordingly.

Furthermore, as skill's base cast time and various buffs/items/talents can heavily modify a skill's casting time, it is easy in my game to end up with a unit using a slow weapon but using lots of "speed bonus" and fast base cast-time skills in order to end-up with relatively fast skills... possibly even faster than a unit who would use a fast weapon with slow skills and no speed bonus. In this scenario, you could have big/small proc on respectively fast/slow casted skills which is counterintuitive and probably not very good balance-wise.

Ultimately, aside from some clarity on the player's side, I do not see any benefit to the approach you descirbed compared to a normalization one (full or partial). Am I missing something here?

Regardless, I'd suggest you build some kind of spreadsheet or similar mechanism that can give you an overview of the concrete data that goes into balancing this sort of thing. For example, something that applies your calculations on all skills over a certain number of attempts and gives you a concrete "percentage the proc occurred" value and information about damage and damage modification et cetera so you can experiment with various approaches and see how they impact the overall landscape.

I will probably end-up doing a few sreadsheet down the road but i'd like to have a clear design philosophyof the proc system first. For example: Even tought I'd prefer to give procs some usability to all types of units/builds, maybe it is not a good idea and it is better to let fast weapon take adavantage of proc and give slow weapon something else to compensate?

I also think it's important to keep in mind that "balancing" something by making all possible choices equivalent in power isn't always the best; it reduces the impact of each choice to essentially a decision about whether or not you like red particles (fire) or blue particles (ice).

I completely agree. But at the same time having an entire system (that I planned on expanding quite a lot to give lots of customization options) "locked" to fast attackers is also not a perfect design. That is why I like my option 3, where I try to get the best of both world.

My personal instinct is that the player should have to assign (equip?) the proc effect to one ability, because it would be boring and confusing if different attacks could cause the same proc. It would be different if i were something where the player was controlling a whole army of units, but if this is the player controlling 3 or fewer characters then procs should not be character-wide. Then, you can adjust the proc effect when assigning it.

Let's assume I go that way. It would solve the issue of procs scaling differently based on the "base cast-time" of each skill but it does not prevent items/talents/buffs and slow/fast weapon to make the proc scale too nicely with things that improve cast speed.

If youre wanting procs such as like a DoT for burning, a fairly decent implementation is either putting a limit on the amount of stacks that the effect can have, meaning if it procs more than that it has no effect which stops things from just getting huge stacks from a fast weapon. (each stack having its own timer) or you can make the effect stack additive on duration.

for effects like a lightning strike or something that proc you can work the same sort of mechanic to have a "only x amount can happen within x time" system

another system is to do something along the lines of Proc % chance * ( time since last proc / scaling factor)

so you could have lets say 5% proc chance per hit and you want it to be back at that after 2 seconds your scaling factor is 2 if you use the timer on a seconds basis .. means faster attacks can proc more, though the % chance is heavily reduced below a certain time factor

Generally speaking, I think a fast attacker should be slightly weaker in aggregate than a slower attacker. If the fast attacker gets better benefit from proc's, this would mean the slow attacker needs substantially higher core damage per turn or health to make up for this additional drawback. I think option 3 would actually start locking characters into roles, where the fast attacker needs to maximize proc output to keep up.

The reason the fast attacker will tend to be weaker is that they have more opportunities to switch strategies. A slow caster has to commit to an action 100 turns from now when the situation may have changed. A fast attacker can swig a potion or start retreating or use a different skill at much shorter notice. So we expect them to get better situational benefits, which needs to be balanced by something else.

Stack limits seem like a natural way to provide variety: The slow proc is stronger, so the slow attacker has an edge one-on-one (offsetting their strategic limitation). However, the fast attacker has more aggregate proc output assuming they can spread it around. So in a crowd situation the fast attacker has the edge in proc power. However, in keeping with his role that requires smart situational awareness: maximizing proc output needs to be balanced against the benefit of defeating opponents one at a time to reduce aggregate damage output. The slow attacker is less flexible but is optimized for the obvious case of "take them down one at a time". The fast attacker can achieve better raw output, but needs to balance that against basic tactics.

If youre wanting procs such as like a DoT for burning, a fairly decent implementation is either putting a limit on the amount of stacks that the effect can have, meaning if it procs more than that it has no effect which stops things from just getting huge stacks from a fast weapon. (each stack having its own timer) or you can make the effect stack additive on duration.

for effects like a lightning strike or something that proc you can work the same sort of mechanic to have a "only x amount can happen within x time" system

another system is to do something along the lines of Proc % chance * ( time since last proc / scaling factor)

so you could have lets say 5% proc chance per hit and you want it to be back at that after 2 seconds your scaling factor is 2 if you use the timer on a seconds basis .. means faster attacks can proc more, though the % chance is heavily reduced below a certain time factor

It seems like the PPM system from World of Warcraft for example who sets a proc rate per minute that takes into account attack speed in order to normalize it so that there is the same number of proc for slow/fast weapon.

Imposing a stack limit on debuff has a similar effect with the downside of not working for all kind of procs. For example, direct damage proc or CC proc would not work under such a stack-limiting system.

Generally speaking, I think a fast attacker should be slightly weaker in aggregate than a slower attacker. If the fast attacker gets better benefit from proc's, this would mean the slow attacker needs substantially higher core damage per turn or health to make up for this additional drawback. I think option 3 would actually start locking characters into roles, where the fast attacker needs to maximize proc output to keep up.

This is interesting :) If I understand correctly the first part of your sentance, you do not refer to proc here but to the relationship between fast/slow weapon. I agree that this is a core question that needs to be answered at the same time. In my current design, fast weapon have more DPS while slow weapon have more DPM (damage per mana). The difference is not huge and not meant to compartimentalize builds but rather to offer some customization options to players.

I'm curious about the part where you think my option 3 (moderate scaling with skill's speed) will "force" players into using faster and faster attack to maximize his procs. Yes, faster attacks will lead to more procs but reacher faster attack speed values will be less effective than for example, gearing to improve the unit's damage or even the proc's output. Since I'm proposing a scaling a procs compared to attack speed that is inferior to a 1:1 ratio, I'm hoping it will prevent that going crazy for attack speed is the only way to use procs efficienctly.

The reason the fast attacker will tend to be weaker is that they have more opportunities to switch strategies. A slow caster has to commit to an action 100 turns from now when the situation may have changed. A fast attacker can swig a potion or start retreating or use a different skill at much shorter notice. So we expect them to get better situational benefits, which needs to be balanced by something else.

I've taken care of the advantage of using a fast weapon (easier to react to rapid changes) in the general damage/speed ratio of weapons. Having procs be flat out better with fast weapon would have to be factored in my initial weapon balancing. This is completely feasable. It may however make slow weapon only good for damage and fast weapon bad of dealing damage since procs can be very powerful on very fast builds and thus the damage output of fast weapon would need to be really low. :/

Stack limits seem like a natural way to provide variety: The slow proc is stronger, so the slow attacker has an edge one-on-one (offsetting their strategic limitation). However, the fast attacker has more aggregate proc output assuming they can spread it around. So in a crowd situation the fast attacker has the edge in proc power. However, in keeping with his role that requires smart situational awareness: maximizing proc output needs to be balanced against the benefit of defeating opponents one at a time to reduce aggregate damage output. The slow attacker is less flexible but is optimized for the obvious case of "take them down one at a time". The fast attacker can achieve better raw output, but needs to balance that against basic tactics.

This seems to refer to a full normalization of proc power (option 2). If I understand your reasoning, you are saying that this is a good option because it is balanced against the different battle set-up when slow attackers will be better than fast one depending on what kind of enemies / formation of enemies they encounter? This is a good call in theory and something that would likely help achieve a smoother balance of the proc system :) Sadly in my game, there is no such thing as fights are more like a Final Fantasy 7 battle where situational awareness, positioning & reflexes are not factored in battle.

A friend of mine suggested to make procs based on the skill's final damage. This is how most procs works in the game Path of Exile for example. In this game, a "Stun proc" does stun the unit according to how much damage you did to her. s damaging proc is also there a function of the base skill's damage. As skill's damage and its relation to cast-speed and weapon attack speed is already balanced, it means procs are also balanced as well. This however forces units that would like to enjoy proc to focus on damage heavily. I'm not a fan of their proc design to be honest. It kills most non-damage builds that would like to proc things. I want a design where a player can build a purely supportive character that procs things even if he deals no damage at all initially. I also want proc to work on defensive skills, for example a proc when casting a healing spell on a target. Or a proc when casting a simple buff on a friendly unit, ect...

Sometimes you will not know a "proc" is broken until players get to interact with it and combo it will a few things you had never thought of. Think of WoW and "Proc" from "Procs" :)

Oh yeah, I expect plenty of these :) I'm designing my character customization process to be very open and let player do crazy stats / items / skills associtation. I like a game where players can create uninteded strategies by pushing the game to its limits. The paradox is that to properly support all that crazyness, I need an initial design of every stats & mechanisms in my game that is well balanced & polished. Otherwise, everything will hit the fan way too quickly and will just require me to re-balance things left and right.

UPDATE:

On the three parameters that can influence the cast of a skill, I've been able to rule one out. There is no need for me to take into account every stat/buff/item/talent bonus that would increase cast speed when balacing procs. The reason is simple but somehow eluded me before. Every bonus in that category come with an associated cost already. If you use an item that grants +100% cast speed for example. It comes at the expensce of not using an item with +100% damage or + 100% to proc X strength (procs are scalable roughly by the same stats as damage in my game). So if I compare a unit with +100% cast speed or one with +100% damage (which apply to his proc), they both end-up with the game ratio of proc's power per unit of time in the end. So it's ok and balanced if unit are using these stats to push there attack speed to insane values.

This does not solve the general proc issue at all but it still reduce the scope of the problem. There is only the base cast time of a skill and the weapon speed that I need to worry about. This also reduce the difference between the mininum and the maximum efficiency of a proc based on the skill's final cast speed by an order of magnitude, which can only help me achieve a better proc balance (even with one of the options I already described).

Edit: My weapons will have a maximum speed difference factor of 2 between the slowest and the fastest. The factor is about 10 for skill's base cast time in most cases. So w/o any scaling, I'm looking at a *20 difference between fastest / slowest cast time for a skill in the end. As a reminder:

• Option 1 keeps the factor 20 of difference between fast/slow unit and makes procs really bad for losw units
• Option 2 make that factor 0 since procs's efficiency no longer depend on attack speed at all
• Option 3 aims at reducing that factor 20 to something like 5 or 10.

Thank you for your input everyone ! I'm very gratefull of every post so far, it is really giving me new ideas and making me rethink the problem on new angles :) Keep it up !