Imagine being the person who the mathematician comes back to because they've explored all their options and according to the algorithm found out you're the best one and you still decide to take them back
Aren't there a lot of assumptions in each of the algorithms suggested for the problems? Are these assumptions mentioned in the book? For example, the parking problem ... Is the assumption that the parking is unmanned? Could there be a different pattern for roadside parking vs parking in a mall vs parking in a business complex?
Is the 37% rule valid for unordered things? For example I've 10 candy samples and I've to select the best and I've time constraints. So randomly selecting 37% to evaluate and selecting the best Except those 37% group of sample
40:00 What about FOMO, or fear of missing out, where instead of dwelling on missed (unseized) opportunities in the past, you dwell on missed (unevaluated) opportunities in the future? With the cost to "play the slot machine," or message another potential date online, growing ever smaller in today's more connected society, FOMO seems to be on the rise. Susceptibility to regret versus FOMO likely varies person to person depending on wide ranging factors like prior luck, patience, faith in the system, and your personal utility function (i.e. is it best-or-nothing for you, or would you be just as happy with any outcome above a threshold value). I know that I personally am more prone to regret than FOMO, so like Bezos, this is what I tend to minimize for, but the same may not necessarily be good advice for others.
One problem with implementing Optimal stopping in real might probably be, it is not worth it go to the extent of 37% in terms of time and effort for the payoff of that particular problem.. and it is worth it in some other problems to go beyond because of it's payoff.. or I haven't understood the underlying mathematical structure properly..
Can this be classified as a religion? Or can religion be classified as an algorithm? This book seems like a very individual-focused religion in that case...
+jandroid33 maybe the better analogy would be 'A Rulebook to live by' like the bible or other religious rulebooks. A full blown religion usually has parts that have to be believed because they are not scientifically provable and focuses around holy objects.
Aren't all ideas quasi algorithm? If I have an established procedure given to me by religion for particular scenarios, that I don't stray from because that's the software programmed into me. How am I not acting out an algorithm?
Friend, Front end "exploration" / Back end "exploitation", seems to describe Mormon Batch Dating & single / drawn out Victorian Courting. In both cases presumably without sex.
How can you calculate the probability of being caught as a robber if your probability of getting caught is going to be 0 until you're caught? For example, if you have 3 successful robberies, then by looking at previous outcomes (3 out of 3 robberies being successful), you'll conclude that the probability of a successful robbery is 1.0 or 100%.
+Jon Wise I think you should start by learning to compose full and meaningful sentences in the English language. (EDIT: The comment has now been changed.)
I think the answer comes in terms of probabilities derived from other ways. Perhaps you look at the overall statistics, not just the statistics of your own instances. Perhaps you can calculate it by composed events of a robbery, which would be much more involved, but still separate from purely your own experiences, and valid. One component in the very long equation may be the probability of leaving your DNA at the scene * probability of them finding it. Another component would be the likelihood of getting away from cops should you end up in a high speed chase, given the probability that you end up in a high speed chase. The probability you can get them to cough up the money without them alerting authorities via the proverbial panic button under the desk. Combine all these, and you'd have a probability based on the average bank, or your specific target bank, which would be independent of your own runs. So the short answer is, use probabilities based on external data, or relevant and related enough data. And if all else fails, engage in the mathematically heavy endeavor of calculating probabilities up from elementary events or events broken down sufficiently enough.
Wow, this is bad... Sorry, but this is knowledge normal people get in high-school, y'know... when they START dating, (as opposed to after finishing their PhD in the nerd's case, I'm guessing?) No one should be married before finishing college, so that's 8 YEARS a normal person has dated. If you don't know what you want from relationships after 8 years, you NEVER will... Math won't help you.