Title has become very long and I am not sure if this is better place in General Discussion or Feedback. Anyway:
The following wall of texts tries to summarize my general design considerations. The considerations take into account the balance state of Cataclysm Quickplay matches. I will try to refrain from specific examples as the considerations are operatable for weapons, talents and traits in equal manners. Though, do mind that I have specific examples in my head with some of them being painfully obvious. Still, not mentioning them specifically in the OP. Also, as this will be a bit of text (like lots of it), it might be a rather incoherent read as I try to link different concepts to each other. Also, behold my mighty Excel illustrations ^^'.
Power Level & Identity
Not quite happy with the section title …. Anyway, in my opinion there exist weapon, talents, traits in Vermintide which are either (very) overtuned or outright broken. This makes up about 10 - 20 % of all stuff.
Naturally, they are exactly the things a good chunk of the community flocks too because it makes for an easy live. So, if we leave the situation as is, parts of the players tool kit stays underused. That is why there are regular calls for either buffing 50 - 70 % of all things or nerfing the 10 - 20 % overtuned elements. Both are possible though there are limitations and drawbacks to each of them. As mentioned intially, I favour nerfing for various reasons. Let’s look at a very simplified and unspecific example.
Obviously, these two talents are not on the same level. So what are the possible solutions here? Either buff talent 2 or nerf talent 1. How could buffing talent 2 look like?
First option 2a would buff the talent by giving it additional utilities A&C, similar to talent 1. But that is also the issue. While it has been buffed talent 1 and 2 nor we feel very similar. You could say the buff made more talents viable. Though in reality, it decreased variation. Talent 2b is the double down approach which has other issues. For more information on that, please look at the Breakpoints & Stacking section. Talent 2c might seem like the most sensible option. Though what is the point of utility D when A, B & C are killing everything so fast that D is basically useless. It just is variation for the sake of variation though without actually bringing variation. In addition, introducing new utilities always complicates balance even more.
So, up to the other side. How would nerfing talent 1 look like?
Option 1a would be the counterpoint to 2b. Instead of doubling down, you are halfing the first one making it notable weaker. For the comparitive balance of the the two talents it has the same effect however without the issues described in the Breakpoints & Stacking section. Though, if the utilities are decreased to much it might become the same as having no utility at all. Talent 1b is also a straightforward nerf. Though, it has the benefit of focusing on just it one main utility. Now, talent 1b and talent 2 are actually providing clear variations. Nerfing has therefore increased variation. Also, both talents have a distinct identity which is also something I would like to see more in weapons, talents, traits, skills: Clear cut identities. If something is meant to be an utility B thing but can be used for utility A as well despite going against the obvious identity than I think utility A should be decreased, even if unpopular.
Breakpoints & Stacking
I think most are familiar with the general concept of breakpoints. In Vermintide it is mostly used for weapon damage and enemy health. Though, there are also game breakpoints and in my opinion, we are dangerously close to them on Cataclysm based on the - not specified in detail - average Legend&Cataclysm player skill. With certain weapons or with certain set-ups we might be over them already.
What is understood as a game breakpoint here? It means that a certain utility has reached a value where the game would need to throw a considerably larger amount of enemies at you (basically spamming them) to still pose a threat to the player. Extreme examples would be a player moving so fast that no enemy can catch up on him, defense&health being so high that you take close to no damage (percentage-wise) or attack damage so high & fast that enemies drop like flies. Take note that these are the extreme variants and even partial breakpoints can be game breaking.
These game breakpoints are also the reason for the often bemoaned small incremental stat stacks for some talents and traits, etc. Why? Because the easiest way is to stack the same stat over and over again. If you want to break the game, you just to do so on very small front like speed or applied damage. If you want non-Vermintide examples just look at the powergame 1x1 for RPG games like Baldur’s Gate 1. Want to reduce difficult and needed skill, just stack your strength & constitution while decreasing charm to stack even more. If you want to break something, just stack the same stat again and again. Why do I mention this? Because of the discussion in power level & identity section. Sure, you can go with option 2b. But you make it easier to reach game breakpoints, which we are - again - at least dangerously close for Cataclysm. So, in my book it is not an option. We have to keep in mind, the range where a game actual works (as intended) is rather narrow. Because it isn’t that 10 % more power means 10 % more efficiency. If you reach breakpoint tresholds, your efficiency increase not longer linear but exponential and the game has to spam a lot more to come close to providing a challenge on the highest public difficulty.
And even if it does not reach the game breaking points, there is still the issue with redundancy which brings me to the next section.
Capacity & Redundancy
These terms might be a bit more abstract. They are probably used for different subjects. I will use them in the sense as used in engineering a plant. Capacity roughly describes the maximum production rate of a product in a plant. In terms of Vermintide it would come down to player engagement or the simplest term kill power. I will lead the discussion with kill power as it is easier to understand though player engagement is the better choice. Both terms are not exactly describing what I am aiming for. The redundancy on the other hand describes the amount of elements present additional to the purely needed parts to reach capacity.
How does this play into Vermintide? If you want to kill all enemies in the time the AI director gives you, you need a capacity of at least 100 %. This means with four players, each player needs a capacity of 25 % (assuming average skill level of the player). If you want to go the extreme route, you could make it so that each player can only kill certain enemies which guarantees that each of them is always engaged to some regard and it would demand a lot of co-op play to pull this off. But this would be a bad idea as one can see quite quickly. Because as soon as one player dies, the team effectively can no longer react to all threats/enemies. Even with the limitation of one player, one enemy-type it would be an issue for the same reason. In addition, the 25 % are assumed as average skill. If you get a team slightly below average skill, they are also doomed to die. So what would be a good player capacity? Obviously subjective but in my opinion roughly 50 % player capacity are a good middle ground. It means even when two players die, it is still possible to keep the game running. When alone you are doomed (as it should be). But it also means that you have already 100 % redundancy within the game. Why is to much redundancy a bad thing? Because it means that some players will have lowered player engagement and the game is simply not fun any longer for them. Which is where most discussion about “fun” split in the forum. The game currently has set-ups where you can reach even with low player skill a redundancy level of 70 % or higher on Cataclysm. This means they get a high engagement and have a lot of fun. But the other three players suffer from this. As such redundancy has to be kept at a reasonable level or the game will become unfun. Even under the assumption that all players suffer the same amount of reduced engagement because of redundancy, it would still be bad as the game then no longer is a fight for survival but a damage race. Instead of being cooperative it is competitive. Sure, people will become aggressive and arrogant and start dying. So some might say, look it is still difficult. But no, people die for the wrong reasons.
And this is already the case if we look at overall capacity and redundancy. Things become even more extreme if we look at niche capacity. If the niche is small enough, people will not care as much like for example a special sniper. Though if the niche makes up a lot of the player engagement, his fun while playing the game will suffer significantly. Imagine for example there would be a ranged weapon which has no mobility penalty, high damage, rather quick reload, no damage drop-off and literally unlimited ammo. With such a weapon a player could easily dispatch 70+ % of all ambient enemies with no one being able to keep up. The area they go through would be more empty and boring. Even if the other player dies after a third of the map because he is to far away because of his power-rush. It still means that for a third of the map, the rest of team is less engaged. The other player dying also doesnt make the weapon somehow balance, only the player in question stupid. And now imagine the same weapon is also very useful against hordes, close up armored enemies, specials and monsters. It would be to much. Another example would be a weapon which might have issues long range but can dispatch 50+ % of the incoming horde with a few clicks before they come into melee range. One-shotting large amounts of the horde again lead to lowered player engagement for the rest. It leads to frustration and unfun situations. This is why redundancy has to be kept in check. The only sensible option here would be nerfing.
For completion’s sake: Obviously, you can counter redundancy by making enemies stronger. But balance in most games shows that enemies are usually a rather fixed point and normally the player tools are getting balanced. So for this discussion, I will assume that enemy strength is a fixed value and no variable. Obviously, you can also buff stuff and simply let the game spam more enemies overall. You have 200 % redundancy? Okay, here are three times the enemy numbers. This is a solution but has also some issues. The first is rather mundane. Hardware limitation. Not everyone can compensate higher enemy counts because their desktop will break down under the strain. So, it could result in less players. Also, it is not said that the AI director can manage this high amount of enemies effectively to pose a threat. Also, the nature of the game would shift. Instead of a medium-paced game where you develop skill and spatial awareness to overcome your limitations and potential clunkiness (the slightly more tactical variation with space for defense and support gameplay), it will become a game based around smoothness, mobility and fluidity where you react to the incoming flood of enemies (the more reaction based variation with focus on aggressive gameplay). Both are possible, and it is Fatshark decision which direction they prefer.
Player Skill & Efficiency
And with this, I come to the last section (honour to you dear reader who managed to get through all of this). The discussion based so far around the average player skill which for obvious reason can not be specified. So everything in this thread can be ignored as me just making up numbers (though it is more about the concepts than the specific numbers). Most players would assume that average skill results in average efficiency.
Though in reality, this is only the case if the weapon/talent/trait has a linear progression. But this is not always given for several reasons. So let’s take a simplified look at which progression curves a weapon, etc. can have:
First option is linear, while the second option is exponential (though with a low start) and the last one is square root function (more or less, the curves are not exactly pretty). So from a player perspective, option 3 would be the way to go. Even with low skill, they can reach fantastic results. On the other hand, most players would simply ignore option 2. Even those, going through the pain will feel cheated. That is why there is a simple rule of thumb to these progression curves. If it is more difficult to learn it should end up more powerful than the linear option. The more powerful the longer the initial slow slope is taking. On the other hand, the easier to learn option should end up less powerful than the linear option. It serves as a good entry weapon for players but is not the best option if skilled. And the higher the initial rise is, the lower it should be in overall power. Because imagine you had this ranged weapon killing all ambients and it would also be super-easy to learn and be similar on maximum power than the most powerful option. Obviously, it would be a horrible state of balance. Or the weapon would be the best weapon available and easy to learn.
Naturally, as developer you would like to have these different progression curves (as long as the rule of thumb is kept) simply to reward those taking the time to develop the skill and to make weapons feel more diverse. Obviously, this also means that not all weapons will be equally viable at the highest skill set (and subsequently difficulty). This drawback has to be accepted. Otherwise, you will create even more redundancy with below average skill, leading to even less fun for a significant part of the playership.
And that’s it. Again, if you mentioned to read through all of this, congratulations I appreciate it. We can try to discuss about certain things if you have questions or see things differently (not sure how much time I will have in the next few days). We can also discuss about more specific examples in the comments because I wanted to avoid them in the OP.
I hope I did copy everything correctly and did not forget parts. Due to the text walls there will also be several typos. I will try to ignore them unless critical or to annoying. Rambling ending.