Here's my input, having flown VFR to pretty far bits of Europe and having had to get the weather "right".
Today's weather is (according to something I read years ago) likely to be substantially same as today's, 47% of the time, so this is really useless.
Also there are as many "weather" sites as there are fronts on their way from the USA at any one time! And a lot of them have duff data. And often it isn't obvious. I've seen MSLP charts which were 1-2 days older than the current ones from the UK MO or from Avbrief. The only way you spot that the chart is older than one can get elsewhere is by working out the preparation time by subtracting its forward time from the time/date it was prepared for; not something people normally check.
Also I've had TAFs over a day old, and ones which were several hours older than one could get elsewhere, from some sites which can do it via SMS. So beware, even with TAFs.
So there is zero point in going to some of these sites just to get stuff which is available FOR FREE from the reputable providers such as UKMO or Avbrief. This is basically TAFs, METARs, MSLP charts.
I believe there is NO point in playing at amateur weather forecasting. Lots of people say they have a JAA ATPL which gives them a superb understanding of how weather works. I'd suggest they get themselves employed as consultants for the UKMO and help all of us! In fact anybody who can do a much better job, and can support their expertise statistically, can make money out of it, because much of industry/commerce pays money for this data.
The TAFs are produced by people that do it all day long, and they have access to data which nobody else gets - certainly which nobody else gets unless they pay serious money. So, I doubt that, statistically, anybody is going to beat a TAF for reliability.
The problem is further ahead, say 2-3 days. The free MO data is only the MSLP charts, and they are wide open to interpretation. So people use other stuff, and typically this is based on GFS. One of the many graphical gateways into GFS is this one
http://www.arl.noaa.gov/ready/cmet.html
which I've been using for a few years. The temp/dewpoint graph is particularly handy in indicating the likely cloudbase.
Some experts have said to me that the UKMO model is more accurate than the GFS model for the UK, so one can use GFS for the rough idea, and use the UKMO MSLP charts for timing (as seen from the positions of the fronts).
I've used the Sembach site for its "cloud progs" but usually I find it wide of the mark.
Other than the above, any forecast past a few days is likely to be imagination!