I'm using a procedure as described in the interactive brokers article here (https://www.interactivebrokers.com/en/index.php?f=5910&ns=T) to compute a probability distribution from option (call) prices.
In essence you solve a very simple system of two linear equations at each strike.
The issue is I get negative probabilities coming out of it.
I'm looking for a simple and robust procedure to do this estimation. Thoughts?