If you're talking about mountain-forecast.com I think they're pretty good for detailed info on mountain peaks. From what I gather, they combine regional computer forecast models with historic data and local sensor information on the peaks they forecast for. I use them a lot in the winter for backcountry and frontcountry resorts. I've found them to be useful for getting a sense of temperature variation with elevation, when an inversion layer might be happening (warming high on the mountain), the freezing elevation etc. They post precise numbers, but everyone should realize there's a range of error associated with any forecast. NOAA's pinpoint map algorithm seems to be pretty much interpolation for elevation and a few other geographic effects with a lot less detail than the mountain-forecast.com.
These forecasts all struggle when the computer models diverge, which often happens when a storm is approaching or moving out or generally unsettled conditions. For a good technical discussion of Eastern Sierra computer models, trends, and forecasts, try
Howard Sheckter in Mammoth. He doesn't post every day, but he gives you a big picture sense of what's being predicted and why. He's usually really good about picking up monsoonal trends and other Eastern Sierra local effects that you should be aware of before heading out.
Personally, I try to poll several different sources that use different models. Watch out for forecasts that just regurgitate the NWS - the consistency might impress you, but its really the same source. A good forecaster will interpret and discuss the variation in computer models and which ones do the best under the circumstances.
