I keep hearing people say the RNG is broken. Of course I highly suspect this is an incorrect statement. However, can you please publish the average for the past 10,000 game rolls or so? I'm assuming that's stored in the data, but maybe not. Or maybe you can put a tracker on it for a day. I'd like to be able to say to these people, "Actually, the devs monitored 10,000 rolls and the average was 3.500035 or similar. And I suppose if that isn't the average then maybe the RNG algorithm needs to be tweaked. Pretty sure that's the case though. EDIT: Obviously this needs to be the unadjusted rolls (like before "Hard to Block"), which might be harder to obtain. Not sure if/how data is stored.
I'm sure if they've tested it and they say it works, it works. Why would they lie? And if they did lie, they could just as easily falsify the data. I think the problem is actually that even when working as intended, the RNG design for Cardhunter works against people's enjoyment of the game. Rolling for a block is fun, hit it, miss it, it's all part of the game. But missing 3 blocks in a row, or your opponent making several unlikely rolls in a row is not enjoyable. It subverts any kind of tactical decision-making by the players, and leaves one or both players feeling a bit off (the loser because he may have won, and the winner because his win was tainted). My suggestion is to give every player a joker card, playable once per match. The card can be used to adjust a dice roll by 2 in either direction, and the opposing player may not play their joker on the same roll. Here's my rationale: 1) Players would need to decide when to use their joker, making it a tactical decision and therefore in keeping with the rest of the game. 2) Allowing players to adjust only one die roll means that they can interrupt an unlucky streak, but can't create a lucky streak. 3) Adjusting the die roll by 2 means you can undo an unlucky/lucky roll (e.g. turn a missed block on a parry or defending into a block) but can't create an extreme roll from nothing (e.g. make an unreliable block unless you were pretty close anyway). 4) It changes almost nothing about the game - in all other respects the game will play as it currently does, including block/armour rolls.
I've not read anything saying it works, but I'm pretty new here. Obviously it's my assumption though. That's kind of the point of the thread, for someone to say it works.
It shouldn't be too hard for them to add the raw rolls (and name the modifiers) to the full logs. Then whenever you think the RNG is cheating, you can check the logs yourself for that session and see what all the rolls actually were.
People might also want to try rolling lots of real dice or playing other war games on their own. I've seen a ton of really weird rolls and outlandish dice series happen. But that is kind of the point, you remember the time Parry failed 3 times in a row, because it is unusual, but all the times it works it's not remarkable, and you don't tend to remember it. But even failing 3 Parry's in a row is still only 1 out of 216 odds (1/6 x 1/6 x 1/6). It's not great, but it is likely to happen if you play 50 games and average 4 or 5 rolls a game (this actually seems low for some builds). And honestly, I still kind of remember the times I saw it happen once or twice in the last year. But that's not a faulty RNG, that's human software remembering the memorable, and the nature of statistics. I like dice rolls in the game because it adds that element of uncertainty and chance. Managing that randomness is part of the fun of the game and strategy. If that's not someone's idea of a good time, there are a lot of other games out there, like chess, that eliminate that element. Good Luck
http://forums.cardhunter.com/threads/random-generator-my.7926/ http://forums.cardhunter.com/threads/beta-feedback-on-rng.2386/ https://forums.cardhunter.com/threads/enemy-rolls-seem-anything-but-random.727/
I might quibble that the Java random number generator isn't quite up to perfectly shuffling a 36-card deck (48-bit seed => 2^48 initial states ~ 2.81*10^14, which is less than 36! or ~3.72*10^41) so most permutations must be unreachable, but I would estimate approximately zero chance of anybody actually exploiting that.
You're right about that. Although, maybe the random number generator isn't reseeded at the start of each match. Is the internal state bigger than the seed? If so, then maybe those other permutations are reachable after all. One time I was doing some work on physics simulations which relied heavily on random numbers. The short version of the story is that the results I was getting were a bit weird and it turned out to be due to the small seed being used in the random number generator. That was the first time I properly appreciated the limitations of random number generators. If you're using a lot of random numbers, and you're looking for the probability of highly unlikely events - then you really need a good random number generator. I ended up using a seeds of around 400 bits, and the problem was solved. That said, most of the time, any old random number generator will be good enough. And the kinds of complaints that people typically make about RNGs in games are nowhere near the real limitations. If you're playing a game and paying attention to things like block rolls and armour rolls, then you're basically watching dozens of random numbers. But if you really want to see patterns emerge in even a low-quality random number generator, you're going to need to see thousands or millions or more random numbers.
The theoretically chemist at our university use hardware randomn number generators for their experiments.
The more I talk with other game developers about this, the more I hear that they cheat their random number generators to make them less streaky. Digital strategy games are training us to expect random systems to produce unrealistically uniform results!
Yeah, despite all the horrors i've seen it do, i actually like the "untempered randomness" guarantee.
I'm less concerned about this. It's the non fixed deck leagues that are more of an issue in that regard.
Dice are random... 3.5 is a good average. I have never had "real" issues with the rolls. Frankly, I try to remove dice rolls from my deck if possible. Use reliable over crafted mail and what not. Tend to use blocks that are 3+ over 4+ and so on. I don't like losing to dice rolls. Oddly, dodge being a 5+ seems to work every single time for me.
