Getting truly random results is quite different from having heavy streaks that are periodically returned to statistical norm with a temporarily slightly higher than average success rate. Truly random results would have no discernible pattern, or occasionally have what looks like a pattern briefly before returning to randomness. When you have a consistent pattern of unusually long streaks of one result, and then a few of the other occasionally just enough to return it to balance, that’s not random.What it looks like is a bad RNG covered up by a secondary program that tracks the percentage balance and returns it to balance periodically.
This becomes a noticeable problem when something like a timed event is involved. If you only have so much time to finish the last match or two and it happens to align with the time when you will get no successful devours for a while, it wastes a lot of time and can lead to failing to complete the event due to bad RNG design, even if you get returned to “statistical balance” later on in explore or PVP.
I’m lvl 736. I’ve since switched to maw, infernal king, sheggra, mercy for my team, but that doesn’t change the fact that the RNG is wonky, and for lower level players they need to rely on something like devour to be able to deal enough damage to win, and may not have great maw, mercy, or good skull spawners to make great maw work on it’s own. A broken RNG that gives nothing but failure when you need it and works well when you don’t isn’t fair to most players who use devour.
As for the team you are showing, I do not have King Highforge, and can’t trait Glaycion or Bloodhammer yet.
Well you can pop in gorgatha(but as first troop and put the other 3, same order just down 1 place) for highforge and I would try and trait them up. It is a very a solid team that you can use for anything. Of course it may backfire but it has a better success rate then devour does.
I’m not fond of Gorgotha in pet quests. His spell is prone to setting the AI up for major combos as often as it can set me up, and while in a normal PVP his 75% damage reduction can help mitigate that, in pet quests the enemy is attacking with such inflated damage that it’s like taking normal damage for him. I am trying to minimize random gem manipulation as much as possible because a single bad gem spawn easily turns into a single turn TPK.
How can you knock a team you never tried though? Give it a spin once. Hold gorgatha full and loop the bottom 3 troops. Just pay attention to their spells though so you are picking the right colors to loop. 3 converters in that team so it is easy to pick a wrong spell if you are speeding through matches.
Do you really think that they programmed in a bad RNG with all this bolted-on tracking to monitor every little thing and then manually push it back to the correct average? That would be insane. The much more likely explanation is that you are seeing patterns where there are none and the RNG is generating random results with the correct percentages.
It’s not as unlikely as you might think. RNGs are among the more finicky and difficult things to program. It’s well known that computer RNGs have a habit of getting “stuck” on the same result for periods of time and it requires extra work to get them to not do that. It’s actually less effort to build an RNG that doens’t work well and then program in a way to push random results back into statistical balance than it is to build an RNG that behaves completely randomly.
I’ll loop aaaaaaaaall the way back to the beginning of this thread and repeat myself: What exactly is your data set?
If you’re making a claim, have some proof. It’s no one else’s job to prove you wrong until you’ve first proven yourself right.
If you have compelling data that shows that RNG is somehow broken, then I would be thrilled to see it. Every time this discussion comes up - not “most”, not “usually”, but every - the result is either that there is no data, so end of discussion, or that there is data, and it shows the RNG working correctly.
If your position is simply “later pet battles can get very hard for new players”, then…yes. Yes they can. That’s a true statement. It has nothing to do with RNG.
I didn’t type out the exact results because it would stretch the screen but here.
Miss
Hit
Miss
Miss
Miss
Hit
Miss
Hit
Hit
Miss
Miss
Miss
Miss
Miss
Hit
Miss
Hit
Miss
Hit
Hit
Miss
Miss
Miss
Hit
Miss
Hit
Miss
Miss
Miss
Miss
Hit
Hit
Miss
Hit
Miss
Miss
Miss
Miss
Miss
Hit
Hit
Hit
Miss
Miss
Miss
Hit
Miss
Hit
Miss
Miss
Miss
Miss
Miss
Hit
Miss
Hit
Miss
Hit
Miss
Hit
I’d have gone longer, but pet quests take a while between each one and I only get so many devours each time. But it’s showing a pattern of alternating between all miss and 50/50 for most of it with one exception where a single 3 long hit streak occurred.
I just use my dragon team in rescues and only had to retry once. I use gob-chomper x4 to kill the goblins though.
I think the devour is pretty fair because it shouldn’t be something you can rely on.
Your intitial assertion was that Kerberos was devouring at 20-25% for you. You asked if anyone else was seeing something similar and no one has verified your findings, though some, including me, have said the opposite. You have direct dev confirmation that nothing funny is going on in Pet events and your own data says 37% in quite a small sample.
You have a 60/40 chance of not devouring and seem to be surprised that when there are streaks, that they are misses. This is what you should expect. Averaging out to around 40% is also what you should expect. Everything appears to be functioning normally, at least as far as the game is concerned.
We’re talking about a 40% chance here. not 10%. Streaks should happen on both sides still. Not quite at the rate of a coin flip, but reasonably close. The feeling of getting one every four casts is fairly accurate though, given that it seems like you usually get 1-3 successful devours between every streak of 4-5 fails. The streakiness is rather bad.
You had an overall 36% devour rate, which is reasonable for a theoretical 40% chance and a small sample pool. A 40% rate, over 60 trials, has a 74% chance to hit at least 22 times (as you did), so that’s also totally reasonable. Nothing is pointing at anything being wrong.
Sure. And they did:
…you’re just forgetting about them, because a streak of misses sticks out in your memory as “abnormal” (even though it’s totally not at a 60% miss chance), while streaks of hits are forgotten as “normal” because you’re expecting to hit “all the time”.
You’re getting multiple streaks of 2 and 3 hits in a row. At this point, I can’t imagine what distribution of hits and misses could possibly convince you that this is random.
No, it isn’t accurate at all. You had a rate of 37%, which is a lot closer to 40% than 25%. 2 more hits in your sample of 60 would have gotten you to exactly 40%. I mean, if you want to tell me that it “feels” like it is happening one time in 4 when it is actually happening awfully close to 4 in 10, then fine.
Hey, if that’s how you feel, that’s fine. I’m not trying to invalidate your feeling about it. You’re welcome to feel any way you like about it.
However, your feeling was wrong. You did not get one in every four casts; you got almost exactly 40%. I understand and acknowledge that you feel there’s a problem, but you also need to understand that you disproved your feeling yourself with your records.
