Tests performed Saturday night at around 2300 local suggested that NRK Longyearbyen 1485 (in the desired signal path and in fact an indicator station) was around 5 dB stronger on the QDFA than on the beverage. At the same time, European interference was down by a few dB up to 10 dB, giving a net signal to noise improvement of 6-7 up to 15 dB over the beverage. Listen to this audio clip to hear the difference (beverage first, after 5-6 seconds you hear the QDFA).
However, at half an hour prior to sunset Sunday morning, we were unable to detect significant effects of the improved back-to-front ratio. It was also quite evident that the QDFA is less sensitive than the beverage up to around 1100 kHz. So in cases of low signal levels (say from the last hour before sunset and onwards) a beverage will outperform the QDFA with its gain alone, except perhaps in the highest part of the MW band (1400 kHz and up).
This is part of the equation though. We did not expect the QDFA to outperform the beverage at any other instances than during periods of high signal levels. If it in fact does that, and to what extent, can not be determined by one morning's listening.
The QDFA lobe is wide. Too wide to separate stations with close bearings, say 20-30 degrees, which a well designed and long beverage can do. We knew that. It was interesting to hear Northern Asian stations like Japan with similar signal strength as the North Asia beverage. This is not necessarily an advantage if you are pursuing daytime North America DX as Asian stations will tend to interfere. But it means that DX-ing Japan in the evening may be possible with less interference from European and Middle East stations than is possible with a beverage. Further tests are needed to confirm this.
We also don't know which effects different kinds of propagation has on the QDFA performance. It is plausible that in some instances, the attenuation of signals off the back may be significant, in other instances it will be undetectable. Again, one or two nights of listening is not enough to draw any conclusions.
A quick visit to the 60 meter and 49 meter bands revealed that the QDFA has a significantly higher noise level than the beverage and is much less suited for DX-ing than the beverage. Up to and including the 90 meter band the QDFA may actually be better, while they are roughly equal on the 75 meter band. The noise level is still marginally higher. Don't waste your time on the higher Shortwave bands with the QDFA.
The two illustrations shows the difference in gain in two ways; one over the MW spectrum where the QDFA gain "rolls off" towards 500 kHz, and one compares a small spectrum on low MW one hour after sunrise. The NA signals on the beverage weren't very strong either with VOWR-800 as the strongest one.