American option price under Cox-Ross-Rubinstein

[alexandrebrilhante]

Binomial options pricing models like Cox-Ross-Rubinstein have been widely used since they are able to handle a variety of conditions for which other models cannot easily be applied proving themselves to be more accurate, particularly for longer-dated options on securities with dividend payments.

[neil-menghani]

This seems interesting. Can I tackle this issue?

[saxena-ashish-g]

Alexandre: Thanks for suggesting this issue and thanks Neil for volunteering to tackle it.

Binomial trees in general are relatively lower priority for us, partly because they don't generalize efficiently to the multi asset case. It is also not clear how efficiently they can be implemented in TF. However, CRR has the simplicity of recombining trees so the memory requirements should be reasonable for reasonable time steps and completeness of the toolbox would be nice.

If you are still happy to do it, Neil, I will assign the issue to you.

[saxena-ashish-g]

@neil-menghani
There has been no activity on this so I will be reassigning this to myself.

[saxena-ashish-g]

@brilhana

I have added a (admittedly not the most general) version of the CRR binomial tree for American (and European) options.
The main limitations I see (which we will remedy in the next few iterations) will be to support time dependent rates and volatility.
I will leave the bug open for a couple of days and then close it to give you a chance to comment if you give it a spin.

[saxena-ashish-g]

I am closing this issue now. Please create a new one if you spot any issues with the current implementation.