"Triple-One-Hits" Too Often?
I have some comments about the "randomness" of the hit/shield counter, and I would like to ask you guys to start paying attention to the "streaks" of one-hits in a row. I noticed what seemed like thr...
Forum Discussion
I've never noticed anything outstanding in the block shield counter.
What I've noticed is that in all online CCGs, people frequently write questioning the RNG used for the game after any outstanding repeated cases of bad luck. Have you looked into the concept of bias in the perception of probabilities? It's quite fascinating.
EDIT below:
Apologies for sounding patronizing. I've seen this question asked soooooo many times it has become tiresome, specially when you're on the other side of the playing field and people are questioning the systems you create for them.
I always recommend people who see patterns in elements of video games that should be random to consider a few questions:
- Is it difficult to implement an RNG that is essentially indistinguishible from true randomness? (the answer is: no)
- Is there any benefit at all in siding with one player in a zero sum game? Will either player be incentivized to spend more money if they lose or win from what seems to be luck in the game?
- Is there a drawback to negatively affecting the RNG in a game? Will people feel cheated and want to leave the game if they repeatedly get bad luck?
If you take a close look at the answers to these questions and compare that with the existing research on cognitive bias and how the human brain perceives elements of luck, you'll see that its not worth digging further. People tend to see patterns, specially when faced with bad luck and there is no benefit for companies to deliberately force these patterns, nor is it difficult to avoid them.
Really the question that I ask myself is: why don't developers just transparently normalize luck altogether? that is, nudge the results so that these clumps of seemingly improbable patterns (which are totally normal in real randomness) never happen so players stop questioning the RNGs. But then, I know the answer. They do it in certain things, but for card games, most designers want raw randomness.
Did not feel it was patronizing at all. As I proved in the test app that I developed in a very short amount of time, YES, it is easy to create close to randomness in code. So I don't doubt the devs are competent. However, in my 30+ years of developing software, I have seen some "unintended consequences" creep into code too many times to not think it can't happen to the best devs on what are seemingly the easiest parts.
I have read about some of the topics you mention, and I agree it is quite fascinating. My degree is actually in Human Centered Design & Engineering, so I have studied quite a bit about perceptions and expectations.
The reason I even posted it was to get others watching and maybe get some feedback. Especially on the "looping edge cases" like Heartichoke/Flytrap. I am fully willing to accept that I just NOTICE it more there because you just sit there watching it helplessly. That said, I have seen a very high percentage of triple and even quadruple single hit streaks in that scenario ... well above the max 7% it should happen.
On a side note: Randomness in code is dependent on the quality of the seed. Predictable patterns exist in all systems, so even the seeding can get into patterns. So our company uses video recordings of live lava lamps we have on site, convert that video stream (color, contrast, movement direction etc) into a ever-changing number, and use those numbers as our random seed. Pretty cool stuff. Not that a CCG needs this level of randomness however. :-)
