I used Topmeteo and Blipmaps this last summer at Ely, Nv. I also did my own analysis using skew-T diagrams. What I really like about Topmeteo was:
1. The predicted cross country distance
2. The hourly forecasts
3. The detailed cloud types
4. Wind predictions
What I didn't like was the near total lack of geographic references (towns, roads, etc.), making it more difficult to get a frame of reference. Blipmaps has the Univiewer that is much more detailed.
Generally, I thought the accuracy of Topmeteo was good, but missed overdevelopment at times. The accuracy of the predicted cross country distance was all over the map, and I couldn't rely on it. At best, it could be used as an upper limit. Sometimes it predicted Cu and it never appeared, and other times it was vice versa. At least it was better than nothing.
My most insightful observation is the inherent error in atmospheric models. If you compare model soundings to actual balloon soundings there is a significant loss of vertical resolution. This means that inversion layers are washed out dramatically. This can impact predicted thermal heights by 3,000 ft, which is not an insignificant difference (actual thermal heights were generally higher than predicted. In fact it means you can fly much further distances than predicted. This happens because model heat from higher altitudes gets moved down to lower altitudes, limiting usable thermal height, and is unavoidable due to the vertical resolution of the models. Examining the skew-Ts also gives me a better idea of thunder storm probability (if you know how to read them). This is a real issue flying in out west in the high mountain deserts.
Tom