itch.io

This is cool for a prototype but needs a lot more development for a sustainable game loop.

• Numbers need to have an impact beyond "bigger == better".
  • Some numbers could be linked in pairs.
  • Extending that, while breaking with the visually cool concept of card snippets, I think having something that increases (and/or lowers) multiple stats could make building a card much more interesting.
  • Battles (or a future relic mechanic) could come with modifiers that prevent you from playing certain numbers based on your or the enemy's last turn(s).
  • Powers could interact more with numbers
• HP need to grow in order to foster more interaction.
  • Having a character die after 1 turn should be a deliberate (or desperate) choice, not the default for anyone without damage resistance.
  • Likewise, the only alternative to damage resistance is being able to oneshot multiple enemies, who lack HP as well.
  • Against bosses/elites, both HP and turns are basically irrelevant and only damage matters.

If you think that is fun, read this thread we had about the Steam demo.

True.

I have been trying to grind that combo a bit after you told me, but so far always fell short, not getting to hunt-proof it in time. That saidI found you can use the Dragoon and infuse it with Duelist + Growing Aspect + Adaptive Aspect, and it'll actually scale to the moon. I believe the upper cap you can hit is 25 Hounds, each with 4 attacks, so 101 attacks total.

So every battle you get 2 HP, which then get swapped (and doubled) 101 times. So after just the first attack cycle you should have 5e30 damage gained per battle. On top, you can infuse Crow onto your Dragoon to deal random damage equal to his attack, plus once per Hound using the Moth of Materializing and you can hit pretty much the same amount.

I am somewhat disappointed by the fact we'd only overflow a float after about 67 million floors. Unfortunately also while I don't quite know the growth formula, at a 1% growth per floor we'd still only last about 16'000 floors. Exponential growth man. Works for the enemies, too.

Anyway, I think I've also found the closest thing to an actual infinite. Again use the Dragoon. You'll have to use an Artificer so it can have 5 aspects. Then infuse it with Duelist and Adaptive Aspect again. This time also infuse it with a Paladin, so it only takes 1 damage, and then have it be Fate's Beloved itself. For the final 2 aspects, give it 2 more attacks. The key to finishing the puzzle is infusing a super Moth of Weaving. It does grant the buff to the Dragoon itself. Let's assume our weavers buff for 10, then with 3 attacks and the amount doubled that's +60/+60. Even with the Black Knight ?, the HP gained will be halved to 30, and since enemies won't hit for that much, you'll just regain it.

Unfortunately even then, you'll eventually be unable to beat Black Knight or False Sun. With 2 enemy choices out of 10, you'll automatically lose one in 45 hunts, so even going infinite won't even stall for another 2000 floors.

Edit: I've finally nailed the comp, and unfortunately some things changed/weren't as aspected, so the maximum is unfortunately only 18 Hounds with 4 attacks each for 73 attacks total. Still bonkers though. That said, there might also be a way to get 24 Hounds with 4 attacks each, but that requires even more bullshit to pull off.

You'll be stopped by game performance long before that XD. I've reached the 600s and ultimately gave up due to lagg.

I am curious about 5k tho. I think enemy growth is exponential, so unless you found a way to scale exponentially, they'll catch up by then. My linear 8 million damage and 500k HP per floor growth should've started being outmatched by floor 1000 or so. And sure, those two values are getting multiplied again for total damage. But still not exponential.

Steam Demo is fun as well, I already see a couple more stupid combos with the new starter.

Time to update your E notation.

Jokes aside, this took me waaaaay too long to achieve. Easily dozens of tries before I finally got every piece assembled. Fun fact: the guy at the end gains 8 million damage and 500k HP per combat.

IT IS POSSIBLE TO HAVE 9 FOLLOWERS!!!

Skip a hunt, lock the +1 follower slot item. Eventually another shows up. Buy both. Profit.

When you use the infinite reroll + set 14 faith followers, getting infinite rerolls causes an issue with memory. Game went horribly slow.

Edit: found even more bugs.

• With high enough stats (5k in my case) distributing your stats among the team will crash because the game runs out of memory.
• 2x/3x speed does not affect floor transition, which does feel quite long eventually.

That's of course also a pretty cool way to change it.

However just in case you missed it: there's currently also a bug/feature where the witch's effect persists while it's dead, which would render the change relatively pointless if not fixed as well.

Heyo, I have on more proposition (if I didn't make it already).

Witch is pretty overpowered against any random damage team, which is cool in general but incredibly unfun when you don't get any attack buffs. I'm taking 10  minute battles. Maybe make it randomly reduce damage by 0-10, that way random damage is still reduced by 90% but can still end the fight if strong enough. And it's actually stronger against other comps rather than just a thing to be ignored.

Adding to that, an accidental discovery I made is that slow-mo on hitting a ramp shot would be a really cool glyph, since it lets you easily switch the lights for the mission.

Happy to be of help.

And yes, it really makes the game feel better paced and makes buying additional balls truly worth it. With these changes I managed to beat 1st difficulty as well as 2nd difficulty with Inferno ball. And it was really fun!

In those, I also had a couple of other, though much less significant thoughts:

• The tutorial could mention that you can switch around the active lights of each mission. That one took me some time to figure out.
• The Multiball mission is probably still a net negative. You get 3 balls, so unless the centre bumper has less than 3 HP, you get something that helps no one except speedrunners. My personal idea would be to alternate it with a mission that restores your protective bumper to half of the HP it had the last time you triggered Multiball + 10% of its base HP.
• Saving a ball that goes out at the side funnels feels really cool, so it's sort of disappointing to me that the side protective bumpers (usually, at my skill level) last longer than the centre one.
• The slow motion effect is really cool and fun, so I see a lot of potential for fun Glyphs and Blessings unlocking additional situations when it triggers. E.g. once a minute when a ball is about to be lost, whenever you cradle a ball, or when a ball moves downwards at a very high velocity.
• There's some potential for a couple cool temporary damage boosts from Glyphs in return for doing risky things. E.g. remove side blockers and gain temporary bonus damage after a ball enters a side funnel, while holding one or both flippers open, or while the centre protective blocker is gone.

This game is great and tickles a bunch of childhood itches, all while making it your own and coming down with smooth graphics, gameplay and turning it into a fun roguelike.

That said, I'm terrible at the game, and while it's fun to play regardless, my skill issues have unearthed a rather frustrating part of the game loop: the game goes swimmingly while the centre protective bumper is active, usually lasting till act 4, though I've made act 6 with good resets. But as soon as they run out, I lose one ball after the next, throwing 10 balls away within a single act and losing.

This feels incredibly frustrating and lopsided. Extra balls don't even matter at my skill level, and getting multiball is a debuff since its chaotic nature kills the bumper faster relative to the progress you're making. I really only have 1 life and it's the protective bumper.

So for the sake of us worse players, I'd advocate for making the protective bumper have significantly fewer HP but resetting every time you get a new ball.

This more gradual loss would really help give (at least me) a better measure of where we are, make the game less dependant on a bumper reset, and also give a more rewarding feedback for keeping a ball going for longer. Plus of course no more losing 5 balls within 20 seconds.

Thank you for hearing me out, and of course for letting us play such an awesome demo. Good luck with your Steam release!

Ran some super barrier stacking to test Maze. Didn't pose a significant problem with a sufficiently redundant build, thank you very much for the change.

Although it made me realise 'Lonely Witch?' should work slightly different. Not so much for balance, as more the fact that with my random damage based team combats against them took 10 minutes every time. My suggestion would be to reduce damage by a random amount between 0 and 10.

Oh, this is a bunch of nice quality of life changes.

Although I would add that the problem with Pallisade was mostly the crazy scaling. Make the damage cap be based on its initial HP and it's fine.

What I have noticed is problematic though (I might've mentioned it before, not sure) is the Maze. Originally I thought it should be good, but I've noticed it screwed up my team order way more often than I expected. I even did a team that went xABxABx (where x are any combination of units, but a B unit must always be behind any of the A units to function) and I still lost back to back maze encounters where both got swapped badly.

So I did the math, and it turns out that if you swap any 1 unit with any 1 other unit, the chances of that happening are 5%. But as I understand it, the card swaps 2 units with another 2, so it's already a 25% chance to swap into a bad permutation.

So yeah, even with redundancy, the Maze hands you a free loss every 4th encounter. Making it to floor 100 with e.g. a Barrier based strategy is unlikely. Making it to floor 200 is basically a statistical impossibility.

My proposed solution would be to make the maze shuffle (one of) the unit(s) in the slot opposite itself to a random different slot. So a redundant strategy can still win while still having a hard time.

Lag issue with Reliquary mode seems to be fixed, unless it only affects specific relics. Thank you very much.

When it comes to Reliquary mode, I've also noticed how I'd definitely enjoy some more drive to play it and all its variations. Even just a simple counter of total corruption earned when beating dragon or corruption * floors climbed would go a long way. But some achievements/challenges would be even cooler.

Normal mode only starts lagging at around stage 90 in Endless mode. Next time I play I'll try see if it's only certain reliquaries or the mode in general.

Also you might've considered already but the endless amount of aspects for that one guy is broken. I've almost (barring some misplays) gotten him to oneshot 100+ endless waves. It's surprisingly good for entertainment though since almost anything can pivot. Maybe make that unit cost 20 for symbolism.

Another wonderful update. Stuff from the first run of the patch:

• Undying Aspect puts the unit to the back of your team, which gets rather annoying after some time. Ultimately I love this rework though, makes it really cool and unique.
• Barrier is amazing as well. Makes for some fun HP strategies and I can't wait to try it out some with some 16x HP builds.
• WebGL (Firefox) seems to have some bug that significantly slows down the game. GPU is barely stressed, and 2x speed is twice as fast as normal so it doesn't seem CPU capped either (FPS counter might help for analysis). But for some reason the game plays out in a very slow speed. Feels very much like some gamespeed variable has been multiplied by 0.25x or so.

I am really enjoying this game, but you should absolutely reduce the default resolution for in-browser plays.

The game seems to start at 1080p, which is still larger than many modern (laptop) monitors. A lot of people might just see the start screen, be unable to interact with it, and never touch the game again.

Found a bug:

Sculptur didn't summon any scuptures for some reason. I don't quite remember the circumstances, but I had no other supporters when I bought it, and I had an Artificer rerolled to trigger adjacent follower start-of-combat effeccts twice.

Edit: I've disabled it again. There was no obvious answer, but I've rerolled a combat trait again, so that might be it no matter the outcome.

I really love this update. You're doing a really amazing job here.

• I love the "choose to give this follower +1/+2" event since it's so much more reliable (unless you have Cleric on board which, in a perfect example of Murphy's law, becomes the default target).
• The Mirror event is really, REALLY strong. Definitely worth more than the alternative, most of the time it gives you two of the best units you could have. I'd honestly rate it at the power level of an elite reward.
• Ritualist is really cool and fun to play.
• Unfortunately the "dead unit lagg" still exists (confirmed for Firefox). It also affects Supporters and e.g. having a follower, then 3 supporters, then a followers will create a pause between attacks, meaning it's quite likely the delay is caused by the units taking attacking turns even when they can't attack.
• The mini-elite versions in endless feel a lot better, though still a tad too oppressive. Maybe halve Barricade and Black Knight again and have them scale a bit more in return. As it stands, I feel they both require me to slot in a bunch of tokens into every endless run.

I've also, through experience, noticed some more units that seem to consistently over/underperform:

• Hound is the best stat stacking unit in the game. It doubles (or triples, with Obelisk) any stats it gains multiplicatively with other multipliers such as double attack or Growing aspect. Compare it to higher tier units such as Ranger. Both units double up on the attack, but Hound gets an extra HP pool, is cheaper, and is better early due to taking overkill damage.
• Sentinel is worth 1 Faith at best. It's useful for exactly one thing, and that's tanking 3 Hobgoblin attacks. The "can't attack" downside has no payoff and the 0/6 statline itself isn't worth more than 3 Faith even without the downside.

Especially with Emissary, that interaction is actually not literal to the wording. As Emissary says "whenever faith is gained by another follower, gain 1 faith", which suggests that the Emissary is the one gaining faith (and thus should trigger Raven again). So as you've intended, the Emissary's wording should change to "when another follower gains faith, that amount is increased by 1". Although personally I'd recommend Emissary to be its own instance and trigger Raven an extra time since it's sort of a really subpar unit as long as Supplicant isn't nerfed and also doesn't scale well in general.

Overall though, I'd probably carefully recommend to make separate instances their own triggers. As it stands, triggers are unintuitive and also don't prevent abuse, as one effect still works while just one or two others don't. It's just a knowledge check, plus a bit more RNG involved.

On top of that the units in question don't really scale as well. They are harder to scale multiplicatively than just a stat-based strategy and they don't scale at all with stats gained inbetween rounds. Compare that to an Adaptive Familiar next to another Adaptive unit together with some permabuffs and you can easily gain +8+8 per round or more.

Ultimately, as far as balancing concerns go, I think you might actually be able to achieve the opposite with some clever wording. In permitting multiple triggers by default, you can then easily nerf overtuned units by making those triggers happen only "once per attack", or instead of being two separate effects, you can make the second effect increase the value of the first one, etc.

On endless, I also just noticed another soft-nerf option, which would be to put most of the special units in one of the first two spots so countered builds have the chance to kill them quickly. Except the Barricade, which should obviously be further towards the end.

There's a couple of bugs with the Mirroring Raven (deal 4 random damage when a follower gains faith):

• It doesn't work with the unit that gains an extra faith whenever a follower gains faith.
• Copying a "gain faith at the start of combat" unit will still only proc this effect once.
• Adding a "gain 1 faith at the start of combat" buff onto the same unit will also not trigger it a second time.
• Nor will the "gain 2 faith on death" unit trigger Mirroring Raven twice if it's enchanted with "gain 1 faith when damaged" (unless of course it survives the hit)
• What will work however is the unit that gains faith on each attack. Make it attack 4 times and Mirroring Raven is gonna blast 4x as well.

I don't know if these interactions may be intended, but they should probably all work given they're nothing special. Except the "gain 1 more faith" unit (after all one might argue it just increases the faith gain) but I think that one especially should work since it's pretty bad.

Another notion on endless mode: I'm not sure how good it is to have random Doppelgangers, Barricades, Firework Masters, and Black Knights in there. They're hard counters to certain strategies (e.g. hyper-buffing a single unit) which leads to fewer strategies being viable there, exactly the reason why you hesitated implementing endless in the first place. I'd rather see enemy units with a lot of stats in the first place, and from there focus on softer effects like multiple attacks, revives, or not taking more than 25% of their HP in one hit.

I definitely agree that reroll is a hard decision. The game's limits, at least for the earlier parts of the game, really don't work well with a reroll mechanic. The best approach might just be abilities or shop items granting rerolls. For a guaranteed reroll, maybe if you skip a shop with 20 faith? That's a high cost, not easily achievable, but also exactly when you'd want  a reroll. There's also always the option of banishing units to make the game a tad more reliable in the lategame.

As for supplicant, my personal opinion is that it'd work perfectly fine at just +1 faith.  Whether you gain faith at round start or a unit's death makes little difference. The former is a bit more reliable and can be doubled up on but the latter actually combos with more options. That said, it might also be more unique to keep Supplicant on 2, and on death combos feel like they need that bit extra eco anyway. Alternative nerfs might be dropping Supplicant to 1 Faith but it gains a second Faith if another unit dies before or after it. Increasing the cost to 4 would go against the cost design but break the aspect of immediately going to 6 Faith. One really unintuitive nerf would be "buffing" it to 4hp since it'd be less likely to die, although that might also cause frustrations with players since it cant die round 1.

I really love the new update. There's some cool new combos, a bunch of great bug fixes, and overall the game feels a bit more challenging as well.

A few things that I've noticed still need some polish are:

• The speed up button should remember its value between runs and also affect the time spent between floors.
• If you have 20+ followers (even if they're dead), the game significantly slows down and with enough of them eventually takes 10+ seconds to calculate your turns.
• Towards the later game and especially in endless mode, a reroll option is starting to feel much more necessary. At the very least I've now gone two runs in a row without seeing even a single copy of a major core unit, constantly having 20 Faith and only incompatible options available.

In terms of balance I'd also like to add that the "on death gain +2 faith" unit feels too strong. It brings you to 6 faith immediately the round after, something only the 1-cost vanilla unit does, while also giving more faith per round thereafter than even most 6-cost faith generation options.

We were given a link to an unstable test version with the bug fixed, I answered in regards to that, not the current live version.

Can confirm the new build works.

Followers no longer attack simultaneously and on-death effects work now.

Comma checks out, German is default language and keyboard on the machine used for testing.

OMG that is exactly what's happening. This build wonderfully fixes pretty much every issue I've come across playing the game.  Funnily you caught me in the middle of me writing a list of them which is now redundant.

Only pointers that actually stay relevant with that build:

• I noticed summoning a lot of followers through stacking start of turn effects slows down the game a lot before attacking or after killing all targets. Not a big issue but it might run into scaling issues with your plans for endless mode or just a longer game in general. (Browser, Firefox)
• Hunts would do better if you could give chase. As it stands they're not really a dilemma and more of a stat check. Either you can beat them and you get free stuff or you can't.  There'd be a lot more decisionmaking if it was more about how much damage one can afford to take chasing after them. What might also help is more "power-now" followers with drawbacks for the lategame. That niche could definitely broaden the possible strategies a lot.

In online play, it's actually every single upon death effect that doesn't trigger.

A little bit of feedback, hope it helps:

• If there's a way to shorten the animation between floors, I missed it.
• Multiattacks provide too little feedback, it was really hard to see if they stack or not.
• Also found a bug when summoning too many units (2x start of turn effect + the dog thingy + 3 units with tentacles) your units just vanish and attack from off--screen.

I've very much enjoyed playing this and hopefully I'll come back ta bunch o enjoy it plenty more.

Sweet game, just a really chill implementation of the genre.

As for feedback, I can't really make good use of the laser. The upgrades only buffing damage is impractical and doesn't help. Lower cooldown or multiple beams would definitely feel better. The stunlock needs to be much shorter or have some other offset so you don't get surrounded or worse, killed when using it.

Good job, looking forwards to see where this goes.

Good work! Here's some quick feedback from my first impression, but do note I didn't get to see nearly all there is:

• A button to reset the save would be appreciated, especially for experimenting.
• My ADHD-wrecked brain demands a speedup slider.
• Consider increasing base health. You can absorb, on average, 10 attacks. So if 4 units arrive, your base is toast even if you kill 1 enemy per attack cycle.
• By making units retarget when new enemies arrive, you can also protect the base a bit and make its HP a strategic resource you can reliably use rather than merely a buffer in case 1 or 2 enemies slip through.
• Consider having fewer levels with more radical changes and instead different difficulties for each level to have the same amount of effective levels. That way progression feels more noticeable.
• High tier units need some buffs. Having 2 units is effectively quadruple the strength to 2. HP get multiplied by more than 2 (since enemies might waste damage) and attack by a bit less than 2 (since attack drops when the first of the two is killed). Meaning 2 boars are just better than a bear, and a rat is just a cheaper boar.
• Meteor Storm is significantly superior to the other two spells, killing 2, 3, or even 4 waves of enemies. And those who don't die are so low, they stand no chance against your units. The other 2 spells beat 1 or 2 waves with heavy support from your own units.
• Being able to cast your spell only once per level, on top of only a limited number of enemies spawns coming at a very steady, unvaried pace gives the game less of a strategy and more of a puzzle feel.

Row 1: shielded & buffed Succubus

Row 2: Lv3 Bard + Lv3 Bone Book

Row 3: Nurse + shielded Big Boned

Sweet little game.

I like this, enough to give it a numbers crunch:

• Most unique augments are unfortunately too weak. The starting values are all just so low that you'd get more from a random silver upgrade. I'd argue some actually become downgrades as they dillute the pool with bad upgrades that only even draw even on their investment by the time the game is over.
• Accuracy has a very extreme curve of usefulness. The initial 30% greatly increase DPS, the last 30% barely do anything.
• The legendary accuracy upgrade is gimmicky. The maximum you can get is 45%, and even that is just barely 1.5x a gold upgrade. And even a flat 60% wouldn't be the world due to the diminishing returns on accuracy.
• The legendary attack speed upgrade is gimmicky in the other way. You deal x1.5 damage. Initially the upgrade is worth less than a gold upgrade, but with enough attack speed it's the single best upgrade by far.
• Armor shred - outclassed by Assassin in 3 separate ways: doesn't increase damage for the 1st hit, requires more upgrades, and has no secondary benefit.
• Blue Fire - The scaling is way off. The first few upgrades do barely anything. Going from every other to every flame doubles its damage. Before that upgrade it's literally worse than just picking damage upgrades. Should max out at every 4th or 5th flame but the damage scaling should easily be 10x.
• Multicast - bad feedback, are the flames glitched into each other?
• Magma puddles - initially weak, but cubic scaling. Ignores armor. Scales with size, duration, and damage.
• Freeze - too low chance to be good initially but quickly becomes a monster. The slow becomes 99% with just 8 gold upgrades that also give HP. Absolutely turns you immortal in the lategame. I recommend the slow affect speed as speed / (1 + slow) and the initial chance go up a little bit. Also glitched, as some enemies in the lategame randomly do not seem to get slowed.
• Vampire - Healing for free is already really good. The initial value is actually good, just simply because there is no other healing. Scales absolutely bonkers to turn you immortal. I'd absolutely tune down the scaling, and also offer more alternative healing, e.g. a slow regeneration.
• Lightning, Barrage, Assasin, and Orbits look pretty good, although all of them need to have their base damage improved.

I know doing the numbers balance can be a chore so I hope this helps.

On second glance, the memory leak just seems to be very situational garbage collection. The RAM usage goes up to about 1.5 GB over the course of a dungeon, but apparently gets cleared if RAM runs out. Plus it's definitely cleared at the end of the dungeon. So all good there.

For everything else, I'll try my best to give some productive playtest feedback:

Witch Doctor's mind control ability should definitely be limited in the amount of uses. Everything else feels fine about it. I've noticed that about 3 uses is the break point where you actually run out sometimes, but not all the time. However 4 uses makes it hard to run out and 1 use would make the ability worthless to upgrade. So my recommendation would be to keep 2 uses, and instead of increasing the amount of uses, that upgrade could increase the range (since range is not that strong if you only have 2 uses anyway).

The Priest's extra turn eliminates Recovery, that alone makes it an incredible support ability. You can even chain it on consecutive turns (if you skip your Priest's extra turn), basically eliminating its downsides. The rest of the Priest kit is very meh though. The lightning attack is by far the coolest attack if it hits, but sucks in practice since it's just too easy for the enemy to make it hit yourself.

Mercenary's secondary abilities are meh. The bomb is just a basic attack with limited uses and the net is an average CC option. But its main attack is the gold standard if you ask me. Would be cool to see other characters with a reliable primary.

Barbarian is probably the strongest character in the game. A limited use instant 2-damage attack is better than most primaries. It should probably be more limited in range & uses. The grapple is great, since you can move enemies 2 tiles in an unorthodox way, but not broken. The part that makes it really strong is eliminating recovery. Here's a suggestion for an alternative: whenever the Barbarian kills a unit, it heals 1hp.

As for the base attacks: maybe I'm simply too obsessed with them because I try to take as little damage as possible and they become a lot better if you accept that you'll take damage here and there. I definitely like the underlying tactical idea though and once there's only 2-3 enemies left, they're pretty enjoyable. More enemies means that there's always someone in range to thwart you, with multiple abilities that can mess you up in different ways. Especially the AOE attacks suffer since it's actually really easy for the enemy to make them hit yourself. On most characters it also generally feels like I'm  just 1 upgrade down to make good use of them. Either the damage, or the movespeed, or the utility of the attack is lacking, and that's also while not having upgraded any of the secondary abilities.

I'll hopefully get to play the game a bit more some of these days, and maybe get to pay some attention to how strong each character's primary really is when making best use of it.

That's indeed a nifty place for the ultimate description, and I absolutely love how it updates damage. Some more things I've noticed while reading through there though:

• The movement reduction from Cane of the Ancient only shows up in the detailed view, but in the quick view it does not show the speed being halved.
• The Burn/Freeze/Shock passives from wands should probably also show up in the passives tab.
• Some artefacts (e.g. Poison Vial) don't have their passives shown either.

And just as an inspiration, you could also track lifetime damage, healing, and ultimates fired, as well as lifetime CC inflicted. Don't overwork yourself though with all those awesome updates you keep coming.

The Guide section doesn't include the mana reduction for Shocked, and the ultimates section is also missing.

This is a lovely game that managed to enthrall me for multiple hours. It's very fun to experiment with different character combos and upgrades and see what works best for each dungeon.

I've also apparently found some memory leak-ish behaviour, over the time of a dungeon the RAM usage drastically increases (Firefox on Linux). Only at the end of a dungeon it's cleared.

As far as gameplay is concerned, there's really only 2 issues:

1) Mind control is too strong. 3 Witch Doctors with +1 speed and mind control upgrades are basically invulnerable. Meanwhile the most damage I ever took in a turn was when a Mind Flayer mind controlled my Mercenary from a mile away with no counterplay.

2) Most main attacks are too weak. The delay hurts too bad until only 1 or 2 enemies are left. Even the "strongest" ones like an upgraded Wizard's or Priest's are actually just as likely to hurt your own units as the enemy due to all the control available. It's much more efficient to use other moves to make enemies hit one another or drop into traps. So most of them need a little more oomph or utility at a baseline or on an upgrade to grant more incentives to use them.

I like the difficulty scaling. It's fun to get more upgrades in a harder game (and honestly, I'd love to see even longer dungeons).

Some of my favourite characters were the Barbarian with its great crowd control, an attack without delay and the ability to skip Recovery; the Priest with the potentially massive lightning AOE and an amazing support ability; and the Mercenary with the only main attack that doesn't have a delay.