Book of Heroes

Devs, please explain your RNG algorithm method

I've been hearing a bunch about this, and this is not the only game that's been brought to my attention. I know very little about Random Number Generation, except it's something used in loot drops for games, casino machines, and in this game, battles. A short google search, going to wiki, has this short paragraph:

"Several computational methods for random-number generation exist. Many fall short of the goal of true randomness, although they may meet, with varying success, some of the statistical tests for randomness intended to measure how unpredictable their results are (that is, to what degree their patterns are discernible). However, carefully designed cryptographically secure computationally based methods of generating random numbers also exist, such as those based on the Yarrow algorithm, the Fortuna (PRNG), and others."

Now I'm sure places like casinos, online and walk-in, use very strict and highly accurate randomness. Except of course the ones that are being paid off, or not being watched, to where owners can put in certain chips to change the percentages illegally. i.e. corner store machines.

When it comes to RNG in BoH, PvP seems to be on the front lines. Many have mentioned how several ability percentage modifiers ultimately don't mean much if the algorithm is poorly randomized. And that many, many players, me included, come out with about 50% win-loss ratio, give or take a few. What exactly does your system do with all the percentage modifiers and how does it determine the outcome? I was told, in simple terms, that the pool used is supposed to reset after each generation, and after all modifiers are factored in. Example, after factoring in player A's dodge gems, any passive dodge abilities, and player B's accuracy gems, each player should have a percentage chance of hit/miss and get hit/block/dodge. These end percentages shouldn't change throughout the battle unless another modifier kicks in. After each player turn, does the pool reset? To me that seems to be fairly accurate. Or do you keep the same percentages pulling out of the same pool, which it seems would actually self-change it's percentage the longer you stay in that pool. All of what I said sounds fairly legit, let me know if I am wrong about something, I'm trying to stay general in explaining, and BoH players, especially ones that have played since the first season, should have a right to know if their gameplay/style actually does what they think it might do? Or does your software have hidden tricks that makes a lot of our decisions null? This is not an attack on you guys or the game, I love the game, always have. It's a good thing if we can continue to enjoy playing it, knowing that the creators listen and respond to it's players' concerns.

Illy_the_Red

ModeratorSat May 14, 2016 3:47 am
One highly visible issue with BoH's RNG is the strings of unusual results you can get, such as 9/10 status effects being resisted by an opponent that you find actually has lower SR than your toon. This would be fine if it weren't observed by too many people in the game, and too often for it to be simple percentages going against you. There's also the long strings of ability activations (eg ice shield) that also seem to defy their stated proc % in some matches and then never appear in others.
Most players are observant enough to see when something is unusual.

Derpy_as_I_wanna_be

Sat May 14, 2016 6:49 am
It would be interesting to have an optional little window that shows the "dice rolls" that caused the outcomes.
As someone who does this for a living, 99.99% chance that they are using the built in RNG in Android and iOS and maybe on their server. Almost nobody writes their own any more (nor should they). Probably the developers know nothing more than they call this function and it gives a random number back. The built in ones are very high quality.

As far as what is going on, you are way overthinking it.

As far as strings of bad luck, when billions of RNG rolls are done (and this game has done that over its lifetime), it's going to happen. For example, let's say the opponent has 30% dodge. 6 dodges in a row is going to happen .0725% of the time or 1 in 1327 matches.

The problem with BoH PvP isn't the RNG itself. It is that the PvP design is such that RNG plays too big a factor in the result. It's more like Backgammon and less like Chess.

And by the way, I do work with casinos from time to time. Slot machines and the like are set up for the operators to change the odds - no "illegal" chips required. In fact, some casinos change the odds at different times of day and even advertise that they payout more during a happy hour (a slow time of day they are trying to get more business). When you hear that one casino has looser slots than another, it's probably true and an intentional configuration choice by both casinos.

Illy_the_Red

ModeratorSat May 14, 2016 8:01 pm
Bbbb wrote:As someone who does this for a living, 99.99% chance that they are using the built in RNG in Android and iOS and maybe on their server. Almost nobody writes their own any more (nor should they). Probably the developers know nothing more than they call this function and it gives a random number back. The built in ones are very high quality.

As far as what is going on, you are way overthinking it.

As far as strings of bad luck, when billions of RNG rolls are done (and this game has done that over its lifetime), it's going to happen. For example, let's say the opponent has 30% dodge. 6 dodges in a row is going to happen .0725% of the time or 1 in 1327 matches.

The problem with BoH PvP isn't the RNG itself. It is that the PvP design is such that RNG plays too big a factor in the result. It's more like Backgammon and less like Chess.

And by the way, I do work with casinos from time to time. Slot machines and the like are set up for the operators to change the odds - no "illegal" chips required. In fact, some casinos change the odds at different times of day and even advertise that they payout more during a happy hour (a slow time of day they are trying to get more business). When you hear that one casino has looser slots than another, it's probably true and an intentional configuration choice by both casinos.

I don't think the observations match the 'it's just how luck turns out' explanation. I've had runs of dodges in excess of the 6 in the example many more times than the odds would suggest, and I've fought far fewer matches than would be required for that kind of luck. Without making huge amounts of notes on observations in PvP it can't be dug into properly, but I'm not convinced all the people making observations on this mechanic are mistaken.
Illy_the_Red wrote:
I've had runs of dodges in excess of the 6 in the example many more times than the odds would suggest, and I've fought far fewer matches than would be required for that kind of luck.

If posts are going to make an argument that ignore the basic tenets of the field (probability theory), there is no point discussing this any further.

Illy_the_Red

ModeratorSun May 15, 2016 4:59 pm
Bbbb wrote:
Illy_the_Red wrote:
I've had runs of dodges in excess of the 6 in the example many more times than the odds would suggest, and I've fought far fewer matches than would be required for that kind of luck.

If posts are going to make an argument that ignore the basic tenets of the field (probability theory), there is no point discussing this any further.

Whatever. The % chance stated doesn't reflect the observations. Without recording all instances since pvp began its hard to be absolutely sure, but it's very easy to remember that I've had a lot of runs where I've dodged 8-10 attacks in a row. But then post-beta it's been par for the course that what people are observing doesn't match what they're being told.