You all need to look at ensemble models way more. The operational ECMWF, GFS, and NAM models are just one determination for each. However, if you change the parameters slightly in each one, you will come up with a larger collection of solutions where more will prefer one forecast over the other. Those operational models can be outliers at times, which I have seen happen a fair bit.
Overall, the ECMWF is the best long-term model. The ECMWF can be too slow, but the GFS can be too fast and aggressive. Lately, I find that going in the middle of them is the way to go. The NAM is best for convection and the first 24 hours, but it quickly loses value when you go out a few days. Lastly, there are times that you must go against all of the models, because they can't capture all the small-scale features in your area. Personal experience and knowledge of local weather patterns simply can't be understated when forecasting.