I love it when I find a glitch or error, something definitively wrong in some way. I'm not being sarcastic or funny, I actually get a sense of personal satisfaction from solving problems. The bigger the headache you're getting from something the more it excites me to take on the challenge. Which is why I always say that problems are fun.
That in mind what I'm currently dealing with isn't quite a traditional problem, just a lack of forethought. One tester found an almost game ruining strategical oversight. Apparently because my system for releasing bosses is based on the items (souls?) you've collected, and if you chill in one area and get a ton of kills in advance, while you may not get extra keys you do still get these pick ups. You can then proceed into the next level and unlock all the bosses at once, use an overpowered spell to kill them all, and move right along, rinse and repeat.
Right now I'm playing with several solutions, for starters instead of letting you choose when to release a boss it's now immediate with no choice which honestly I'd planned from the start but received some complaints and after talking with them about what specifically they didn't like about it we've now managed a solution I think they'll find quite satisfactory. The second is to have a tracker for each floor of how many you killed on that floor so the doors don't automatically unlock based on that, the number was previously incremental and now it's more of a slightly fixed number altered by the floor number variable (non-multiplicative) and they're happy about the idea of seeing how well they did on each floor though I'll have actual results tomorrow of how much they like it. I've also added a behind the scenes timer tracker that shows after you game over how long you took on each floor. I'm also dumping that info to see how the game progresses from floor to floor to make sure the variables are balanced right, because if everyone consistently dies on the same floor then obviously I've done something wrong.
I also added a little spice by picking bosses at random, there were previously 12 bosses alternating a set to each floor, now it just picks one for each cage with fixed master boss levels every 10 floors in which you encounter 4 of the same boss and a 5th super powered version of that boss, though to my knowledge none of the players have made it to the 120th floor yet without that slight cheat or whatever you'd call it.
I'm doing other minor tweaks to make it easier for them to get that far and thinking about making an actual final level. The final level however will be beyond nintendo hard, I'll make it level 1,000 and have it include every single mega boss, and enemy you've killed that far along with a final gatekeeper with an ungodly health bar, just so you'll have hope that I can crush.
Though not to be in direct defiance of occams razor as it pertains to entertainment (a bit of a stretch I know) I'm also thinking about overstepping the apparent simplicity of my current auto-allocation system available to the player and implementing a more sophisticated system of balancing regen rates, accelerated regen rate, mana tank max storage per rank and overall, along with spell power and related in a more calibrated methodology. Essentially the idea is that you can invest points (again thinking I might just call them souls) into categories and each category will have a graphed curve you can view to see how powerful the spell is, how long it takes to charge / cast (for power spells not streams) and mana cost, and any other variables and based on how you invest determines how the curve changes to reflect these stats. For example you might want a simple power spell to charge really quickly so you'd invest in its early stage mostly, but if you charge it into its second or third stages it means you really want more power so you'd invest in that area for that stage, and perhaps you instead want to cast a standard power spell really cheaply so you'd invest there, and while if you invest enough all the stats can improve you'll likely suffer in other areas for it.
This principal also applies to your actual mana regen, the max mana per tank by rank in which you effectively have 3 mana tanks and each of their regens are based on the investments in them, and you can invest in how potent the magic in them is because they all basically boil down to a value of the first tank after all the math is processed. Right now they're set up so that a single point of mana from the second tank is worth several mana points, which when you use mana from that tank your spells by default charge quicker, are more powerful, etc.
I can tell you right now I will not be letting you cast different spell types together, and even lasting spells like the lava do not interact with other spells and I won't be programming them to either. That is just beyond my current skill and wouldn't really make a huge difference in players. I'm still considering screens for you, I'm just still playing with the HUD, if any of you are familiar with UDK and unrealscript just let me know and we'll collab a bit so you can help me get it displaying right. Right now I'm working on getting the "crosshair" or otherwise known as reticle to be based on a variable which is the current range of the spell whether you're doing an area effect spell, a target cone, or a straight stream/beam. I can manage it slightly by state but nowhere near where I want it at.
I should save some more for another post.... Say something about all this in the comments so I know you exist please.