Statistics

• Statistics are a well founded way to do machine learning.
• If your boxing opponent has tried to hit you 5 times, and each time it is with his left hand, what do you think he'll do next?
• This is a really simple form of statistics.
• If he's tried to hit you 100 times, and 80 of those have been lefts, what do you think he'll do next?
• If of the 20 rights, 10 were succeded by rights, what will he do after a left?
• What if you include footing, or hooks vs. jabs in this calculation?
• Multivaritate statistics gets very complex, but it's well understood.
• However, the more variables you have, the more data you need.