Should a star system be included in the guide? Currently there isn’t one. The idea behind one would be to direct people to the best problems quickly and easily. The disadvantage for me is that its another – after grading – attempt to absoultely quantify the unquantifiable. As in grading democracy doesn’t tend to work, unless you have a hugely popular area like Font and a website like bleau.info with a system to capture votes on grades or quality, the sample size is always going to be too small and easily distorted. So the reality is that a star system would largely represent the opinion of the author which is fine but I think it would missell so manyproblems in some peoples eyes that it might do more hard than good. The approach I favour is that of mentioning in an introduction to an area some of the standout classics and also saying in the description of a problem if it is of particularly high or low quality.
I also have plans for a list of the top 100 or 200 problems.
Assuming one would use a system like this:

*** very very nice problem one of the best in the country
** very nice problem, one of the best in the area
* worth climbing
0 probably not worth climbing

What would people think should be the spread between the ratings? Say in Glendo – or choose elsewhere – what percentage of problems should be in each category