| Author | problems in damage done by the hero |
|---|
Hi,
Can anyone explain how is it possible for the VetaL's hero to do constantly 21 damage? Don't explain the amount, I know from where it comes. My problem here is the constant number of HP in damage when it should be a random in between a minimum and a maximum.
Combat: https://www.lordswm.com/warlog.php?lt=-1&warid=13101050
Thanks! |
| he does damage other than 21. though about the streaks must just be random. i mean, eventually it's gonna have that particular sequence |
| Look at the end of the combat when he hits my unicorns. Once 20 hp damage and the rest 21. About that I was speaking. I don't know, 7 or 8 hits with 21. |
| it's what you call randomness, not an error:S |
| LOL! I call it error in randomness. :D |
| me tooooooooooooooooooooooooooo:D |
| Why do people always think clusters are not random? The probability for 8 times the same damage from a damage range of 8 is 1 to 2 million, so why shouldn't this happen in 13 million battles? |
| And it's about average of his damage range, so its the most likely damage to repeat ;) |
#8:
Averaging random numbers is waste of CPU ticks. Well, if it is average damage, then they should specify it like that.
#7:
A test of any RNG is the distribution of the points in the range. The more clustered, the worse RNG. |
#5
it doesn't matter what you call it, that doesn't change the fact that it's just random... |
#9
If you make the same experiment sufficiently often you will get every result that has positive probability. The event you described has positive probability so it was bound to happen and an RNG has to reflect this. |
#11:
Unfortunately, a fight is not lasting long enough to allow such experiments, so, I suppose here the RNG should provide less number in the series, but to be better scattered. I think the same problem is with the luck and moral. And that's because you need your probability to finish in a combat, not in 10 (just giving a random number here). It is quite frustrating to fight an important fight and the probability not to get fulfilled, while after that fight to have a stupid hunt and the probability distribution to start paying from behind. |
So your problem is not that you think the values are not random, but you don't like the distribution that is used.
I think the damage is idependent identically distributed between lower and upper bound and what you probably want is to ensure that the average damage you deal is not too far from the expected damage, like it is for luck and morale, but that wouldn't be independent anymore.
This game has random elements and they can certainly deceide a match, if they couldn't we wouldn't have them, if you want the game to be more deterministic write it to I&S, but it's not an error no matter how inconvinient it might be for you. |
#13:
It depends on the point of view. Here, I don't suggest anything, but I am emphasizing a conceptual problem. One may live with that, but still, as concept, it is faulty in my opinion.
I know about RNG's and statistic distributions (I am working in that field) and I know that many distributions show a local cluster even if overall it gives nice distributed points within the function sampled. I don't care about roulette here, but I care about combats. You cannot concept a combat strategy if you cannot rely on having your probability distribution fulfilled in that combat. I am not speaking of a combat of 1-2 turns, but I am speaking about a combat of 20-30 turns. If 10 turns the hero makes the same damage and the damage is high, think how fast a knight can kill your troops using only escort. And if you use luck, but you reach a cluster of low luck probability, then you have no chance. So, conceptually, I consider this a bug and that's why I posted this thread here. |
I don't suggest anything, but I am emphasizing a conceptual problem
So you pretty much want to say that you don't like a feature of this game without giving an alternative.
You cannot concept a combat strategy if you cannot rely on having your probability distribution fulfilled in that combat.
Right you keep seeing this in history all the time:"Let's go home guys, in expectation we win this battle, but probaility might work against us." This goes from Sun Tzu to Patton.
Planning under uncertainty is a major part of strategy. If randomness wouldn't affect the outcome of a battle you could formulate it as an extensive game and in the most interesting case it would be NP-hard to calculate.
Here it's even more deterministic than in the original HoMM V, there morale and luck aren't forced to expection and it doesn't has the 2 miss, 1 hit rule for incorporals. |
So you pretty much want to say that you don't like a feature of this game without giving an alternative.
If I wanted to give an alternative, I would have offered myself to implement it.
If randomness wouldn't affect the outcome of a battle you could formulate it as an extensive game and in the most interesting case it would be NP-hard to calculate.
And you suggest that planning to get a cluster of low damage is a good way of thinking? Then why are the luck and the moral in the talent wheel? Just for lucky ones who will receive it when they need? Hmm... then better to play roulette or to go killing myself in a battle because I know that I will have a lot of luck in the next battle (just being a little bit ironic, no offence, please).
I am a little bit surprised by your affirmation: it doesn't has the 2 miss, 1 hit rule for incorporals. In this game, I got even 4 misses in the row. For ghosts, usually, you are right, but for apparitions it doesn't work. And how do you know about here the luck and the moral are forced to expectations? Did you check? Did you have anyhow to check that? Then how do you explain a necro with luck parameter 1 to have overall more luck befalling on the troops than an elf with 3 points at the luck parameter? I am expecting in a combat if I have luck +x and my opponent luck +y, me to be able to achieve (x/y +/- error) times more luck in that combat. But I don't expect the error to compensate a difference in +1 or +2 points in the luck parameter. |
If I wanted to give an alternative, I would have offered myself to implement it.
And there's my problem, it's pretty easy to say "change it", but quite hard to tell how.
And you suggest that planning to get a cluster of low damage is a good way of thinking? Then why are the luck and the moral in the talent wheel? Just for lucky ones who will receive it when they need? Hmm... then better to play roulette or to go killing myself in a battle because I know that I will have a lot of luck in the next battle (just being a little bit ironic, no offence, please).
It's pretty good summed up by "Expect the worst, hope for the best, capitalize of what comes."
I got even 4 misses in the row
Well I never experienced it and it's in the game description, only thing I can think about it is that magic resets the counter.
And how do you know about here the luck and the moral are forced to expectations? Did you check? Did you have anyhow to check that? Then how do you explain a necro with luck parameter 1 to have overall more luck befalling on the troops than an elf with 3 points at the luck parameter?
Again it's in the game description and it makes the number of luck and morale converge faster to expection than the usual iid. If this were not the case than there would certainly be a thread like yours for someone's mainstack having 8 morale/luck in a row. |
And there's my problem, it's pretty easy to say "change it", but quite hard to tell how. - there is a way to estimate the number of turns one combat can have. one can implement this feature and have the probabilities forced to be fulfilled. :P
It's pretty good summed up by "Expect the worst, hope for the best, capitalize of what comes." - i pretty do that. only that i had enough of it to happen (that's the reason i started this thread). :D
it's in the game description - also the 40% hp recovery was in the game description and just after telling Shebali that it doesn't work, it started to work. :)
it makes the number of luck and morale converge faster to expection than the usual iid - any support for that? any confirmation?
8 morale/luck in a row - i wouldn't be surprised to hear that with this random engine. :) |
there is a way to estimate the number of turns one combat can have
You can't estimate the number of turns, an agressive strategy concludes the battle much faster than a conservative and it also depends when high damage/luck/morale occurs.
i pretty do that. only that i had enough of it to happen (that's the reason i started this thread). :D
Well if you want to win certainly you can try to catch thieves with a partner or only join group battles with players two levels lower.
also the 40% hp recovery was in the game description and just after telling Shebali that it doesn't work, it started to work. :)
Then perhaps you have to tell her how you want the game to be. ^-^
any support for that? any confirmation?
It's not unusual for people to have 3 or 4 luck so chance for 8 luck in a row with iid is 1/6561 and I'm certain people would notice and report such an anomaly. |
You've started to become sarcastic when I thought I have someone who can give me arguments. It's a pity because I could go on with contra-arguments, but I can see that I have nobody to discuss with here. You can post as much as you want, I keep this thread alive for the admins to get a point (I hope they don't think like you).
Only one more thing I will answer to you, fusei. There is a way to estimate the number of turns of a combat and that is using hyperbolic functions. The method is old enough for not being known by the gamers (especially the developers of this game). |
