MFLOPS measures million floating point operations per second. Most apps you use every day don't touch floating point, so that benchmark is less relevant if you are looking for a realistic benchmarks.
On the other hand, Quadrant measures CPU, Memory, IO, 2D and 3D speeds. Granted, the weight between those 5 categories may not always reflect how you use your phone. But they are valid to certain extent. XDA lag fix speeds up IO - if you are using an app that happens to rely a lot on IO, it will be faster. Otherwise, it won't. If you are playing lots of 3D games, then 3D component is important. If you just want general smoothness out of your phone, CPU score is important. etc.
I think at the end, you need to know what the benchmark score means and how to interpret it. IMO, just ignoring it is being ignorant.
If you look at the order of what you see here, I don't think it's totally out of realism:
Smartphone Benchmarks - Compare and submit performance benchmark speeds of HTC, Motorola, Samsung, LG, Sony Ericsson and other smartphones
SF scores high with those patches because of IO and 3D. If you use an app that relies on IO and 3D (like 3D games), it will play better than other phones.
Perhaps the best way to publish these benchmarks is to show its 5 components individually so that you can better weigh each component to how you use your phone.
