A comparison of models of the antennas I use for SOTA using 4nec2.
All low mounted inverted-V dipoles will show a pattern with the greatest gain perpendicular to the ground. It does not matter if the ground is absorbing or imperfect there will be an image of the antenna which will reflect the rf. This is perfect for NVIS on the lower bands where the F-layer will reflect it back down at high (elevation) angle.
A vertical is ideal for DX but ideally it needs to be elevated above the ground (so does a dipole if it is not to be affected by the ground). In that case the highest gain is horizontal. If it is near the ground then it will be lossy and the maximum gain will be at a higher (elevation) angle.
You can see the above in the figure below. A 40 metre dipole has highest gain upwards but the quarter-wave vertical will out-perform it at lower angles even though it is inefficient with a gain of -4.5 dBd (i.e. 4.5 dB less than a dipole in free-space). Note, it is hard to raise this 40 metre antenna for SOTA use so that the feed is appreciably above the ground.
A 40 metre inverted-V dipole with top at 5 metres (blue) and a quarter-wave 40 metre ground mounted vertical with three resonant radials (red), both over a finite ground (13+i0.002)
For 17 metres I use a 5/8 wave vertical because I can load a 40 metre vertical to work on that band without having to dismantle everything (I just add a stub). As before, in the figure below a low slung dipole shows lots of gain vertically but it is not so clear cut that the 5/8 wave is any better; it is more about what is suitable for the location. An elevated quarter-wave ground-plane is more efficient at low angles and it is possible to raise it high enough using a 10 metre pole.
A 17 metre inverted-V dipole with top at 5 metres (blue), a stub-matched 5/8-wave 17 metre ground mounted vertical with three radials (red) and a 17 metre quarter-wave ground-plane with the feed point at 5 metres (green), all a over a finite ground (13+i0.002)
While the ground is flat in the model it should not change the overall pattern to any great extent other than to tilt the lobe relative to the average ground tilt.
The difference between portable verticals and dipoles is the order of a few dB for DX; the patterns for other HF bands will be similar. It comes down to preference and convenience but all the antennas I have described can be made easily (without the need for a tuner) and I have built and used them.
I did a write about a set of fan dipoles that I built a long time ago which I have used since: