Can the RNG be improved?

Infrared
Infrared Posts: 240 Tile Toppler
Every so often, people will post about how the AI cheats, or how a ridiculous series of unlucky turns made them want to smash their phone. And then others will rightfully point out that it's all actually just confirmation bias that makes us see things that way, even though the game is using a perfectly fair random number generator (RNG) to create tiles. There are just as many times that the player gets ridiculously good tiles generated in their favour, but hardly anyone complains about those.

I suggest that a fair RNG is not necessarily the the best thing for the game, with regards to the player's enjoyment. If you have the patience for it, look at my recent run on the 4* Deadpool's Daily Quest versus Quicksilver as an example. The first 7 minutes or so were horrible. I did everything I could to prevent QS from matching blue, but he always did it anyway due to new blue tiles randomly dropping in just for him. It was fair and random, but QS collected 27 blue before I could use IW's force bubbles even once. After I won that match, the boards became overwhelmingly in my favour. Match 5 again and again. It helped me beat the rest of DDQ very quickly. Again, it was fair and random, but I had the feeling this didn't really count as a true test of my speed running ability, because of the ridiculous cascades and match 5s for me.

What if the game had kept track of the player's luck? So if the AI got too many cascades at once, the game would shift the RNG to favour the player a bit more, and the other way around also. Obviously, it's going to be difficult to measure something like luck. But let's say it can be done, and the developers manage to derive the ideal RNG luck factor for player enjoyment and game balance. Would that be an improvement? I'm interested to read what others think. If it were you in that video, would you have preferred a more balanced luck factor? Or do the two extremes make it more exciting? Or is having a fully random RNG more important?

Comments

  • XandorXerxes
    XandorXerxes Posts: 340 Mover and Shaker
    There's no such thing as a fully random RNG. Enjoy: http://engineering.mit.edu/ask/can-comp ... dom-number

    tl;dr: Computers generate random numbers with an algorithm based on a changing seed number. If an error causes the seed number to occasionally not change or be generated improperly, you'll get mirrored results despite supposedly random generation.

    Have you done the math on how often a cascade should fall in like that? It's not very likely... at all. Just a 4 in a row is a 0.29% chance, or 3 times in every 1,000 drops. How many horizontal cascades (when the entire row comes in) have you seen? If you assumed between you and the computer that there were 100 horizontal row drops come in per round (that's not happening) you would see a match-4 pop in once every 3.33 rounds. Even at the much more modest 20 horizontal row drops per round, you would see it happen just once every 17 rounds.

    What I wish they would do instead of trying to "adjust" luck is just have a secondary board above the player board that drops the tiles in. That way all the match 4-5 silliness is worked out before the player even sees it. You'll still see the occasional cascade drop in, but not the stupid 6 in a rows we occasionally see now.
  • Malcrof
    Malcrof Posts: 5,971 Chairperson of the Boards
    Moved to Suggestions
  • simonsez
    simonsez Posts: 4,663 Chairperson of the Boards
    Infrared wrote:
    Or is having a fully random RNG more important?
    Yes, it is. Having the devs tinker with a system we already think is biased, in order to intentionally added biases, is a recipe for disaster.
  • Infrared
    Infrared Posts: 240 Tile Toppler
    For the record, even though I did use the word "suggest" in the original post, this was not intended as an actual suggestion for a change in the game. It was meant to put forth an idea for discussion.
    There's no such thing as a fully random RNG.
    Fair point. I guess by fully random, I just meant it not being weighted in any way.
    simonsez wrote:
    Having the devs tinker with a system we already think is biased, in order to intentionally added biases, is a recipe for disaster.
    I concede that it's very easy to tinykitty everything up by changing the RNG in the wrong way. But I am not convinced that there could not be a right way to change it that would result in a better game, however difficult that might be.
  • frostCoH
    frostCoH Posts: 71 Match Maker
    I agree having more "real" RNG would be superior to whatever has happened recently. Before the last simulator event it felt like the drops were weighted to the house which is fine, but now it feels as if I'm playing with a claw machine that purposefully screws with you. It *feels* obvious that they have rules governing the game's luck, but recently the luck (automatically dropping exactly what the computer needs to match away placed countdowns, such as Hawkguy's red arrow, suddenly filling the board with yellow when Astonishing Wolvie is in need of health during the current PVE, etc.) maybe this is confirmation bias but it sure as hell doesn't feel that way.
  • Dayv
    Dayv Posts: 4,449 Chairperson of the Boards
    frostCoH wrote:
    maybe this is confirmation bias
    We have a winner! icon_e_wink.gif