Can someone explain how...

[Tim28213]

I am beginning to believe that the devs recycle old code. The reason I think this is that something will get fixed then a few releases later that same bug shows back up again but they are able to correct it quickly the second time, and the third time, etc... until they finally get that item out of all the code pieces that are in development. Insert many recurring bugs that come and go with new releases and it makes you wonder.

So you team that worked well with the current AI might find that against the updated, or possibly outdated, AI it plays slightly differently. Then the next update comes and your team works well again because it is back to AI v2379.8 instead of AI v2378.6.

I'm speculating of course, but I can't imagine how some of the bugs keep coming and going if they aren't reusing at least some pieces of old code in the development. And as complex as this game is I can't imagine not reusing code to keep the development pace moving.

[DeathDealer-5301]

It's possible that by leveling up talismans and whatever else you did, the autoplay AI is now using toon abilities in a different order.
So it used to be toon A would always go first and use his power move, but now toon B does, and toon A never gets to.
Could try to put your toons in a different order? not sure how it selects that and if order will help, but worth a try.

[Yup]

Anything you do to any of your team members can change things.... power up a talisman, add a skill shard, swap a talisman, buy a new relic... I believe all of these factor in to opposition AI strength. And there are times where something like a relic purchase pushes you from difficulty B to difficulty A making things MUCH different until you power up more to sit in that difficulty level.

Like running BNW III compared to BNW X... when you get to X initially it can be very hard, but as you power up characters, talismans, and relics, it gets mind numbingly easy.

I think most dungeons follow this overall methodology to a degree. This seems true, at least to me, because having all characters, all skill sharded, all relics, and all full power talismans SHOULD mean there's nothing in the app that is ever really "difficult" - however that's not the case. Insanity dungeons can be challenging regardless of any player's specific resources. So there's some weighted AI factor in there. Granted I think this is done to keep the app more interesting overall and it's not done to "screw" or "mess" with players maliciously. But... when you jump a difficulty level it can be VERY noticeable at fist.

I also think new dungeons have some other stuff going on... perhaps the "old code" as Tim speculates. I can see that taking place as well. Since new dungeons typically crash like crazy for several days, then eventually those crashes start to subside. I can only assume it's because the code-base has been "massaged" in some way. I couldn't even run Tiny Tim's Madness or Insanity levels on day 1 or 2 because the app crashed before completing even a single run.. now I can at least run it 5-10 times before a crash.

[Bakura]

Well, I agree with Tim about recycling code.
They obviously do it as far as dungeon graphics.
Last years "Phantom of the Opera" dungeon had snow added to it and became last years "Christmas dungeon," then w/ out snow became the "Daily & Weekly dungeons," and currently the "Labyrinth dungeon" is the same but w/ an added background of an inlet or lake. Oh, and also this year's Cyborg dungeon has the same dungeon as those just mentioned.... I think it was basically the same as last year's cyborg dungeon too (regarding the scenery.)

So, basically it's rehashing the same stuff but with a few tweaks. I'm sure the AI coding gets the same treatment.