Board Thread:Wiki Running/@comment-3225604-20150806111049/@comment-26235098-20150807171237

Wildoneshelper wrote:

Counterargument: Ok. I was under the impression the poll did have a limited number of entries. I was confused about that I guess. If the poll has no cap, then I agree on point 1 - makes 100% sense to me.
 * Point 1: [...] The poll has no cap in the number of levels, as long as it has votes.

Wildoneshelper wrote:

Catinthedark wrote: Not sure how all of this came out of my objection. All I was saying is that 100% comment requirement seems excessive to me, when comments are still subjective, even if they do provide a means of dialog. I will not compare to a poll here because if I were to do it via poll, the poll would not be asking people to vote for a label -- I agree that would be WORSE than the comment system. Wildoneshelper wrote:
 * Against point 3 - Probability, stats, and the random factor, i.e. some lucky sod is going to throw the rating by passing in 10 goes and saying it's hard.
 * Point 3: According to your claim, we can simply eliminate the poll too, because you could just click which difficulty you prefer and then submit your vote, without giving reasons. Now, most insanely hard levels are based on the comments rather than the poll. Comments are somehow better than poll because some votes contain reasons which convince people why the level is insanely hard. If we do according to your point, then why a difficulty is there on the first place? Why should we grade levels?

My point on probability and stats is that we simply do not have enough sample data and no logarithmic calculation can make up for that. If data is skewed, it is skewed. A small data sample means probabilties may not be accurately reflected. The larger the sample, the more accurately probabilities are reflected in fact (or reported experiences here).
 * The accuracy: Logarithmically, the difficulty could fit into the system. For example, let's say 30 tries is very hard and 100 tries is insanely hard, then 50 tries could be extremely hard. Or let's say 100 tries is insanely hard, 1000 tries is nowhere near "insanely hard" and it's "impossible", making 100 tries to be "extremely hard" only. Similarly, the logarithmic thinking could apply to probability of passing the level and the average number of special candies or combos produced in a single level.

We already get pretty specific, with very/somewhat ratings, and yes we could add more, but... IDK that it would really be meaningful, given that we cannot assign them in a controlled, experimental, scientific way.

The more I think about it, and the more I think our current ratings are more than adequate.

My only reason to support this would be that SOOO many people seem to want an extra difficulty on the high-end...