So I’m facing a team that has the Dragon Soul with all traits unlocked and i was struggling to win. Finally get down to that Dragon Soul and beat him. And then he resurrects…FIVE TIMES!! Luckily, my final two troops were krystenax, so i kept summoning silver Drakons. But that’s complete bs. There’s a 25% chance for resurrection and every single time i face a dragon soul with that trait unlocked, it resurrects at least once. Yet i use black manacles (20% chance to devour) and it hits maybe one out of every twenty times it’s cast. Ridiculous.
1 Like
RNG is weird.
I’ve been using Mercy in my “best viable PvP team” for about 2 months now. I play roughly 200 PvP matches/week. Out of all of those, it feels like maybe only once or twice a week do I get in a situation where enough purple is close to yellow to get a free turn.
However, I sure do remember all those matches where the board begins with two yellow/purple clumps, making it impossible for me to prevent the AI’s Mercy from earning a bonus turn!
OP, start tracking whenever you feel like you’re being stiffed by the RNG. Either you’ll find you’re getting fairly treated, or you’ll be able to show exactly how the AI is being favored.
2 Likes
I think we are collectively coming to an understanding that percentage-chance things in this game are not really random.
There are some kind of hidden coefficients and other stuff. This isn’t something to be upset about because on the whole it makes the game better.
@Lyya pointed out yesterday how the devs came out and directly stated that they tinkered with how certain troops can spawn/loop. This was done to improve the game experience on all sides. Think of how many instances of this intereference behind the scenes they are not telling us about.
Take Webspinner and Goblin Shaman out for a ride. Your head will explode trying to make the spawning connect. These hidden corrections to mechanics and Troop pairings, which would otherwise be out-of-control, are pervasive in the game.
I think this will be the topic of my next rant on video 
2 Likes
Conspiracy theories on range are BS. The devs have no incentive to create suspicion and player rage. Plus the sort of coding you are talking about is hard and needs to be deliberate and … no incentive.
In regards to TDS I have had TDS spawn six or seven times in a row before when I am trying to kill it.
I have also had my TDS spawn 5 times in a row and the opps TDS dies way more often than my mine. The spawn rate feels about 25 percent.
TDS respawning 5 times should happen once every 1000 times you kill it. I would kill TDS about 40 times per week (maybe slightly less now it is not meta) which over a year means I should have seen this at least twice. When it was meta I should have been seeing this once every three months.
In regards to mercy I have played it in offense and defense and the bad/good starting boards felt about right. Mercy often misfires for the defense.
5 Likes
A lot of things misfire for the defense because the Ai is as dull as a brick wall. So some troops that are good for invading are not good for defending because they require a brain.
Manipulating the looping mechanism (or the starting board, or the cascades) gives an advantage to defense teams to compensate for AI’s deadbraininess.
Beside, the devs have admitted that they manipulate the looping chances in the game, so we don’t have a conspiracy theory: there’s a conspiracy. How far it goes is open to debate.
I’d just like some kind of game mechanic document/stickied thread that says “Yes, we do manipulate RNG to try and make a better experience” or “No, we leave all RNG decisions up to the RNG implementation.”
If it is left to the RNG itself, then what little I do know about probability is applicable. I can properly tell someone things like, “Look, 25% doesn’t mean you always see 1 in 4. There is a universe where you go 100 trials with no results, deal with it.” I’ve seen some interesting counter-arguments to this, and in a few situations thrown up my hands and said, “You do seem to know more about statistics than I do so I’m going to step back.”
But if the RNG is being manipulated, anything that “feels off” is subject to complaint. In this case it “feels right” that TDS resurrections might cap at 2, or the percentage might drop by 50% each time (25%, 12.5%, 6.25%, etc.). And the game might decide to give me a nice Mercy board if I have, say, 2-3 losses in a row. Etc.
So I’d like to know, because people bitch about the RNG all the time. If it’s “really RNG”, it’s unfounded and can’t really be changed. If the RNG has tweaks, then every complaint is a potential bug report unless the tweak is transparent.
RNG is tweaked for definition, or we will see a much more chaotic disposition on the board, with clumps etc. To what point it is being manipulated is open to discussion.
I remember a time when the new and improved RNG, with much more cascades and bonestorms, was defined as “fair to the AI”. I still cannot manage to wrap my mind around that concept.
…Found it.
I don’t think this can or should happen.
In all games whether the audience realizes it or not, what they perceive as happening is not what is happening, instead it is a representation or a metaphor of something else. Another way of saying it is that a game on a deeply philosophical level is presenting an illusion to the observer to create an experience while disguising a differing underlying complexity. Magic tricks. Interactive magic tricks is what I’m saying.
This game has grown incredibly complicated and well beyond the scope of what it was ever designed to be. The screws and bolts are starting to rattle loose and the illusion is breaking in many places. In many places it is held together with special care and it shows even to the untrained eye. I don’t know that it will do any of us good to expose the underlying truths on a molecular level. It would kill this game for us.
1 Like
This thread covers this in more detail. Essentially, it was a statement that the game had been cheating too heavily on the player’s behalf, and they reined it in after finding the bug.
I don’t believe luck is being “tweaked” outside of the areas the dev team has stated, but I don’t think there’s any way to resolve this to every player’s satisfaction short of source code introspection or reverse engineering, and as such it will remain a perennial topic.
I’m fairly new to the game and forum and thus haven’t seen any of these topics before. I just think that if it says there’s a 25% of something happening, then it should be that way. For that thing to happen FIVE times in a row is ludicrous. I don’t know what the odds of that happening are, but it has to be pretty damn low. Now if they want to make a stipulation and say it happens a at a different percent for defense teams, that’s fine. Btw, it makes a great case to put all % based troops on your defense team!
I like what you’re saying. What we know is what we are told, and the rest we are left to infer. Drilling to the core of this game to know about every detail does not serve players. Even if we could, I don’t want to and I don’t think, collectively, we should endeavor to.
1 Like
All probability calculations are based on assumptions that the events are random and independent. Specifically, resurrection of TDS is not random and independent. Hence, simple rules of probability calculations cannot be used to estimate the chance in this particular case. See other topics that discuss this issue extensively.
It is quite reasonable. However, there are also a few examples of developers making somewhat inaccurate or incorrect statements which might have some degree of contradiction with the actual state or inability of developers to reproduce occasional bugs or glitches. This happened in the past and might happen in the future. In these cases, IMHO, some deeper delving might be beneficial. Otherwise, I certainly don’t want to run around with cheat engine always on and check every little piece of damage or mana that gets through the game. It can be done but would be rather disappointing as probably 95% of the things are working as intended more or less. 
Fair points!
I hope I’m not morphing the convo too much, but when calls arise for more dev explanations, I temper my desire for that with this: Never once have I seen a player demand an explanation from the devs as to why, for example, a matrix of 3x4 Red gems just fell on their behalf, charged up 2 of their Troops and gave them an ET, when mathematically we all know this to be ‘impossible’ given what we accept to be true about this game.
I think in a sense my underlying feeling is let us not peel back the illusion too far, because we will not like what we see.
2 Likes
That discussion about random and independent is way above my pay grade…lol. So what exactly do you mean when you say it’s not a random and independent event? Does that just mean every time it says “there’s a 25% chance of this happening” that you can’t calculate it based on multiple times? For instance, if you face tds 10 times, he should only resurrect 2 or 3 times on average, yet it absolutely happens more than that.
Been there my friend and can only the AI is in my favour when others battle me and my dragon soul comes back 5 times too 
What he is saying is that, like most other software, this game relies upon the illusion of randomness whenever a random event is called for by employing an algorithm known as pseudorandom number generation, which is “good enough” for most intents and purposes but which does not produce independent results when asked for the next value; rather it provides a human-unpredictable but non-random sequence. Most software pRNG is designed to have even distribution over its sequence, but some generators are better at this than others.
5 Likes
For random and independent events, probability of one resurrection is 0.25 and for two sequential resurrections 0.25*0.25=0.0625 etc. If the events are not random and independent, it might mean that they are random and dependent or non-random and independent or non-random and dependent.
Each of these three cases has to be reviewed separately and results of probability calculations will be different. In either case, if the event is repeated many-many times, it will probably converge to somewhere in the close proximity of 0.25 as the description says. This was not tested by me however, so IDK, it most likely should get there due to Law of large numbers Law of large numbers - Wikipedia This does not mean that event is random and independent. So, one cannot calculate probabilities of two sequential events easily.
1 Like
I love you man. Your posts are teaching me quite a bit, much of which as you know, doesn’t always apply to GoW 
At one time I worked at IC Group doing large-scale giveaways, promotions, sweeps, contests, etc many insured or probability-based. If you ever want to rap let’s do it, I think it would be enjoyable. But lets go off forum.