Cruel/Merciless should mean more enemy skills and better AI -- not just bigger numbers

I'll just go out and say it: right now Cruel and Merciless are horrible from a game design perspective. It's not that GGG doesn't understand how to implement a difficulty curve; the increasing number of tricks that monsters gain going from Act I to Act II to Act III (all Normal) is somewhere between sufficient and good (could still use improvement, but I'm not really complaining). However, going from Act III Normal to Act I Cruel feels more "Act I" -- going backwards -- than it feels "Cruel" -- going forwards; same thing with Cruel to Merciless.

Yes, the monsters have more life and more damage. Who cares? They're zombies. Braindead, slow, retarded zombies. Cakewalk; boring. Later, you run into monsters that actually have, you know, movespeed, and spike damage becomes a major problem. Until, of course, you gear/level up appropriately, Path of Life Nodes/Affixes style. Then they are also pathetically easy. Gearcheck/treecheck; boring.

They may have hit the gym while you were gone, but they definitely didn't go back to school.

All of this is because of a simplistic "just increase the numbers" pattern for iterative difficulties that may have worked fine in 2001, but is hopelessly behind the times now. Creating new difficulties by simply increasing the numbers on monster stats is no longer acceptable in a premier ARPG.

This has been a problem with the game since its infancy; for example, try reading Dev Diary: Balancing Path of Exile. The entire post is about automating game balance through the use of tools that assign numerical values to everything using a system based around a mathematical variable named effectiveness.

"

Chris wrote:

Monsters are also balanced using the same system. We’ve specified that an average monster currently has 16% of the effectiveness of a player, and they scale at the same rates. Different monsters have different ratios of how much effectiveness is spent on offense or defense. Their strength/dexterity/intelligence alignment also modifies what areas effectiveness is spent on. For example, a pure-strength monster such as a zombie gets entirely life and physical damage reduction. A half-dexterity/half-intelligence monster would get its defensive effectiveness split between life, evasion and energy shield.

Unfortunately, increasing difficulty is not a numbers game; as the Extra Credits video says, it's more of an art than a science, which means it's not something you can write an equation for.

We can, and should, do better. I'm not saying the "effectiveness system" has to go, but it needs to be relegated to its proper role: something that creates a rough draft, with detailing done by hand.

But what should that detailing look like?

Below are some changes I'd make to Cruel, Merciless, and Maps. Keep in mind that difficulty should be a curve, so these concepts should be slowly applied as characters work their way from level 1 to maps.

Decreased Monster Life and Damage (relative to current levels). No, this isn't a QQ thread; I'm not saying that Merciless is too hard, I can't progress, cry cry cry. On the contrary, it's actually pretty too easy to progress in Merciless once you pass the Path of Life Nodes/Affixes gear/level checks. No, the reason for the decreases here is: pretty much every suggestion I'm making below increases the difficulty along another vector, and I'm not trying to turn PoE Merciless into D3 1.00 Inferno here.

Increased Monster Movement Speed/Skills. The higher difficulties should make kiting harder, rewarding players who spec into very high movespeed and/or use crowd control abilities to prevent monsters from reaching them... however, this should not be an uncontested challenge, and right now it's entirely too easy to hold monsters at bay with things like Summon Skeletons Spell Totems. For some monsters this just means more movespeed; for others you need to get a little craftier. For example, fast zombies are just... weird. So instead, increase difficulty with their "burrow" ability; use it only for show in Normal, then have zombies use it a little bit more intelligently in Cruel, then zombies only unborrow when they're in direct melee range of you in Merciless (scary!). Maybe give them a "burrow speed" on the higher difficulties -- invisible underground travel -- so they actually seek you out. Plus other stuff, like increasing attack speed for goatmen leap slam, actually have Flicker Strike monsters follow the cooldown (in Normal) but eventually give the same mobs access to the Frenzy skill (in Merciless)... stuff like that.

Add new monster offensive skills. As you go through Normal, monsters go from being stuck on basic attack to eventually getting more interesting skills. This pattern should progress as they go from Act III Normal to Act I Cruel, and onwards all the way to maps. Basic example: Zombies. When you're first learning the game, these monsters should be easy and stupid. However, in Cruel maybe some "Drowned" zombies gain Glacial Hammer. There's an attack that will make you think twice.

Give monsters "support gems." Players are able to go around firing off multiprojectile chaining Lightning Arrows; there's no reason monsters shouldn't be able to do this too. Actually, those Lightning Arrow skeleton archers should probably be doing exactly that by the time you get to Merciless (take that, summoners!). In addition to adding AoE and other unique effects to attacks, support gems should also provide a better model for how to increase monster damage; it should not be a pure linear increase. With some monsters, elemental damage (as if through WED) should be the primary method; with others, increased critical chance (as if through Increased Critical Strikes); with others, through resistance penetration (as if through Fire Penetration). By increasing monster damage in a wider variety of ways, monsters would find holes in player defenses, and players would find themselves occasionally in difficult spike-damage situations, which actually makes for entertaining gameplay; what's not good gameplay is when the system relies on that tactic alone. By the time you get to maps, every monster should have both a proper active skill with a few proper supports (even if the supports are pure damage-based).

Give monsters more curses. If you ask me, the map suffixes that curse players with Enfeeble, Temporal Chains, et cetera, should not exist. None of them. Why? Because by the time you get to maps, so many enemies should have curses of their own that you can expect to be cursed pretty much all the time anyway. Right now a few monsters use curses, which in the current implementation I feel is a bad thing; curses, from a monster perspective, are probably advanced enough that they shouldn't be applied until around the end of Act II Normal. Increasing the frequency of monster curses at the highest difficulties encourages more dynamic gameplay where getting one curse overwritten by another changes the player's entire decisionmaking process; the point isn't players being continually cursed, the point is players having the curses on them continuously changing.

Give monsters more keystones. I'm actually pretty strongly against rhoa charges being unevadeable -- well, with the ability to score critical strikes at the same time, at least. Give rhoas Resolute Technique, and include it in their text box, then I wouldn't really have a problem. In general, however, monsters should gain keystones as difficulty level increases. Maybe there's a caster-type or two that gains Pain Attunement, some monsters with Arrow Dodging, another with Unwavering Stance. I think the Blackguards in Act III should gain a small amount of cold damage in Cruel, then in Merciless get Elemental Equilibrium; it would definitely make for some more interesting teamwork with their Arc-casting brethren, as well as with Gravicius.

Let unique monsters play with remote skills a little more. I understand that totems often exist independently of actually summoning in the monster world, but just as monsters can spawn other monsters, monsters should be capable of using a variety of remote skills. Remember how Sawbones in the Upper Prison casts Summon Skeleton? By the time you hit Merciless, he should do what players do and support that with Spell Totem. Imagine the challenge of killing a unique that lays multiple Ice Spear mines, then detonates them if he's on low life for more than one second. Although it's probably not a good idea for normal monsters to be using remote skills, Merciless and Maps (non-quest) unique monsters should be a showcase of some of the tricks remote skills are capable of.

Give monsters better AI. A lot of people suggest this on a global scale, but I actually believe the AI is just fine for Normal; while players are getting used to the game, it's okay if they have some chances to outsmart the monsters. Instead, it would be better if monsters gradually use better strategies as difficulty increases. In Normal, monsters might bottleneck at a door, leaving them vulnerable to AoE destruction; in Merciless, those same monsters would analyze the available pathing and split into three groups, get into position behind all three doors leading into the player's room, then launch their assault simultaneously. This would lead to lots of cool "oh, they do that now" moments, as well as cool moments of superiority in Normal. (For reference, currently monsters use an inefficient dispersal algorithm when dealing with doorways, which does thwarting mass death by AoE but also leads to inexplicable idle time that makes the monsters still much less of a challenge.)

Make more monster affixes challenges to be overcome, not simple gearchecks; include visual queues so player is aware without reading a wall of text. A "challenge to overcome" means that the affix is very strong if you don't take specific action against it, but has a bypass that makes it much easier to deal with. I think this point is best explained by example, so let's look at some current monster affixes:

• Hexproof. This affix is too restrictive; it gives an unbreakable immunity to curses -- that isn't a challenge to be overcome, it's something you just have to accept and live with. Instead, it should be something like "Recovers from curses 500% faster." Even a level 20 curse, with a normal duration of over 11 seconds, would last less than 2 seconds against this, meaning it's still a very efficient protection. However, you'd have some interesting ways to counter it, if you so chose: you could use Increased Duration and Faster Casting to assist in reapplying the curse; you could support the curse with Spell Totem so the totem is continually reapplying it instead of you; or you could take a particular keystone to be able to ignore the affix completely. For a visual queue, curse symbols over these monsters should appear different than usual; this would clue the player in on the fact that the curse will be gone in a second or two.
• Elemental resistance aura. This affix is not restrictive enough; the game offers numerous means (support gems and curses) to penetrate elemental resistance, which are already in popular use. It also already has a visual queue: the elemental resistance aura itself. This monster aura should be capable of giving monsters a 100% resistance -- in other words, an immunity. With curse immunity removed (but curse resistance still in the game and strong), this could make for some interesting encounters, and even pure-elemental builds wouldn't really have trouble finding ways around it. Care should be taken, however, to ensure that no monster can have more than one elemental immunity at once, even if it can be penetrated.
• Increased life regenration. This affix is too restrictive; there is no bypass, you just try to out-DPS the monster. A much better replacement affix would be "+500% more energy shield cooldown recovery" -- that's one second between damage and recharge. Granted, it would be ES-specific, but that would give the players a clue as to when it might pop up; giving monsters (and players) a blue shimmering animation when their energy shield is recharging would also serve as a visual cue. This recharge would be very frustrating for kiters, but could be countered by facetanking the monster while attacking it, or using a damage-over-time ability to prevent the recharge. Clever players would even notice the "more" wording, and determine that Vulnerability would still be a rather effective; at level 16, Vulnerabilty would increase recharge time to 2 seconds.
• Reflects physical/elemental damage. There's a lot of QQ on this one, but it's actually kind of close to what we are looking for; although it's a very punishing affix, it can be countered with remote skills like totems/traps/minions, heavy damage-over-time emphasis, the Aegis Aurora unique shield, the Arctic Armour skill, or even a combination of an emphasis on physical attack damage, high life leech, and the Vaal Pact skill. I think the only fix that would be needed is to limit its application rate to 20% of your maximum life per second, similar to how life leech normally works (removing the need for Vaal Pact in the physical damage example). Basing it off of life would also make it much less deadly for energy shield builds (CI would probably base the reflect rate off of the pre-CI life, similar to stun), although it would still prevent ES cooldown recovery from triggering -- for a very long time, thanks to the low rate of application.

Speaking of Reflects Damage, implementing changes like these wouldn't be a "nerf," it wouldn't be "making the game easier" -- because we'd essentially be adding more Reflects Damage type mosnter abilities. Implementing such changes properly would still mean plenty of QQ from players about how the game is too hard and they can't progress. To some extent, we want that -- GGG's intent (as stated in the Dev Diaries and elsewhere) is for this to be a hardcore ARPG with truly difficult content. What I'm saying is: let's try to have difficulty in more than one dimension. Trying to achieve the desired difficulty solely through the means of spike damage doesn't really make the game harder; it just makes it gearcheck masochism.