Open LBlacklink opened 4 years ago
This is what I've been thinking about, too. I want to know what my expected value is if I do things optimally based off the current probabilities. If that expected value is higher, I should wait, otherwise I should sell. I've been playing around with this in my own project and I think I have something that is accurate. When I have some free time I can work on getting it into a branch here.
This would be the kicker -- in general, I would like the site to show EV[MAX[turnip_prices]]
, meaning what is the expected value of the maximum that turnip prices will be at any point after the current moment.
Then this value would answer the question 'should I sell now or later'. If EV[MAX[turnip_prices]] > current_turnip_price
, then wait. Otherwise, sell.
There is talk of adding violin plots and other methods of viewing more in-depth distributions of your patterns but I believe in addition to that it could be very efficient and effective to calculate the expected value of the probability set and show that.
If you're not familiar with the definition of expected value, it pretty much is the value that if you were to repeat this situation infinitely many times with the same conditions the average would be that value. It's very convenient for analysis of things like gambling games and it would be perfect for this situation; an expected value higher than your buy price will mean that if this scenario would be repeated infinitely you would make a profit. It is the ultimate compact number telling you wether to buy or sell.
I don't have too much knowledge about how you're currently programming this so I don't know how easy it is to implement this but if you're already able to plot violin or box plots that would indicate that you have enough data to calculate an expected value pretty easily.
The wikipedia article on this is pretty good if you want more info.