Inverse probability weighting

Inverse probability weighting is a way of finding out how likely something is to happen, even if it hasn't happened yet.

Imagine you're playing with blocks and you want to know how likely it is for a blue block to be picked out of a pile of blocks. First, you need to know how many blue blocks there are and how many blocks there are in total. Let's say there are 10 blocks and 2 of them are blue.

Now, you can figure out the probability of picking a blue block by dividing the number of blue blocks by the total number of blocks: 2/10 = 0.2 or 20%.

But what if you don't have all the blocks in front of you yet? Maybe there are more blocks that haven't been made yet or maybe some of the blue blocks haven't been painted yet. Inverse probability weighting can help us estimate the probability of picking a blue block in this situation.

We can use some information we do have to estimate how many blue blocks there will be in the future. For example, if we know that blue blocks are made 20% of the time, we can assume that there will be 20% blue blocks in the future.

Then, we can use inverse probability weighting to weight our estimates based on how likely they are to be accurate. If we have more information that supports our estimate, like knowing that there will be more blue blocks made in the future, we can weight our estimate more heavily.

Overall, inverse probability weighting helps us make predictions about the future based on the information we have now.