Computer Simulations to Compare the Chess Greats Throughout History

Many wonder what would have happened had Fischer not vacated his title and faced Karpov in 1975.  Or how Fischer would have stacked up vs. Kasparov or Carlsen.  Or how about Morphy against the greats of modern day.

Has there ever been an attempt to come up with a computer simulated game/match between the greats of different eras?  Presumably this could be achieved through the analysis of playing styles and move tendencies just as basic parameters with the assistance of top notch chess engines and the computer technology of today. 

What else would have to be taken into account to make such a simulation possible?  Or would it even be possible?  And if possible how accurate would the simulation be?

Thanks in advance for all replies.

There are bots here that mimic that great players. It would be possible to pit them against each other (in a casual game) but I don't know if it would trip any red flags with the software.

Yes that is right about the bots.  I wonder why more is not discussed about matching them up against each other.  But as you allude to the software just might not be up to snuff to the point where the simulation would be anywhere close to realistic.  But I suppose since they already have these bots now in the not to distance future such a realistic simulation could occur.  

     The problem with this idea is that the programmer would have to be nearly as good a player as the person they were trying to imitate to be able to construct a program that could play just like the champion they are named after. Just sprinkling in enough minor mistakes to get the bots' practical rating to equal that of the subject at their prime doesn't do the trick. And of course, often the human plays above his rated strength, sometimes blunders and loses rating points.

Trying to simulate the play of actual people is really freaking hard, probably even impossible at this point in time.  One the other hand, people have attempted to compare players across different periods through various rating approaches, Jeff Sonas' work at chessmetrics.com probably being the most in-depth. IIRC, at one point he claimed his ratings were universally comparable ... that is, you could predict the results of a Morphy vs. Karpov match based on his ratings. Not sure if he still claims that or not.

Folks have also evaluated historic players by running their games against an engine to get accuracy scores (much as one can get on chess.com).

I have neither the chess nor statistical skills to evaluate the utility of these approaches, but I'm sure each has strong and weak points and can be interestingly played off against each other.

Interesting idea, but it won't work in practice.