![]() ![]() ![]() Depending on the confidence level you select (95% for many applications), each parameter is expressed as. There is some statistical uncertainty in the fitting parameters, in this case, for and for the Weibull distribution. The problem with predicting extreme events by extrapolating mundane data is that the prediction is extremely sensitive to the shape of the tail of the distribution (the shape at large values). I could do better things with that time, like maybe finish my PhD. The number of attempts I will need to make in order to be 95% certain of success isĮxtrapolating the two-parameter Weibull distribution to find and assuming that each game takes 9.6 seconds each leads a prediction of 2 billion years. The probability that I will succeed after tries is Third, in spite of its outlier status, human nature makes us concoct explanations for its occurrence after the fact, making it explainable and predictable.īut how hard could it be to predict? If we trust that we have modeled the underlying distribution properly, why not just extrapolate out to 60 seconds and calculate the probability of success? Let’s call the probability that the survival time exceeds 60 seconds. How do we predict the probability of occurrence of an event that we’ve never witnessed? Nassim Nicholas Taleb refers to this kind of event as a Black Swan, especially if it has a disproportionate impact, like a market crash or extreme flood.įirst, is an outlier, as it lies outside the realm of regular expectations, because nothing in the past can convincingly point to its possibility. One of my central assumptions will be that my performance from match to match is stationary, that is, my score on match n is independent of my score on previous matches and does not depend on how many matches I have played.Ī few attempts at fitting the cumulative distribution accomplish nothing but to convince me of the futility of this exercise. Additionally, I played a few warm-up matches before recording statistics. I have already been playing this game off and on for about a month, so there’s reason to believe that my learning curve is relatively flat. I played 150 consecutive games on “hexagonest” mode and recorded my score (survival time) for each. First I generated some data on my gameplay. The goal of Super Hexagon is to dodge an endless stream of collapsing line segments and survive as long as possible. So given that a typical game for me lasts about 8 seconds, how many times do I have to play until I surmount the final barrier? There is no natural end, just a series of checkpoints, the last of which (“Hexagon”) comes at 60 seconds. A single game can cost less than 5 seconds before you crash into one of the infinite, endlessly collapsing geometric shapes. ![]() The feeling is akin to what I imagine slot machine play must feel like to a gambling addict. The gameplay is addictive, to say the least. Terry Cavanagh’s Super Hexagon is the hardest game I’ve ever played. ![]()
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |