There sure seems to be a lot of confusion by some people about what
the the "ground side" of the tuner should be connected to.

While it is true that you don't need a "true ground" for an end fed
antenna to work properly, you do need the other half of it or a
counterpoise instead.

A dipole works well if both ends are well above ground and isolated
from ground. If you place one leg of that dipole vertical and the
other leg horizontal you have a vertical with a counterpoise. Provided
the counterpoise leg is well above ground. It will work fine.

However if you have that counterpoise leg on or very near the ground
then it starts coupling to the ground. When it does that it greatly
detunes that counterpoise leg.
Making it a 1/4 wavelength when above ground works fine. As soon as
you place it near ground it is no longer a resonant leg. The
capacitance is much higher than in space. The velocity of propagation
slows way down. Those things make it no longer resonant. It is not
resonant and it is poorly coupled to ground. That is going to give
poor performance of the antenna.

For it to be effective it then needs to be well coupled to the ground.
When it is then it is no longer a resonant counterpoise. Making it a
quarter wave long does little good.

The point is that making a radial a resonant length that is to couple
to ground or is in close vicinity to ground is useless. Because it is
no longer resonant in those conditions.

Likewise something like a ground screen imbedded in the fly bridge of
a boat will not be resonant (maybe at one frequency if you are lucky)
and it is too far away from ground to successfully couple to it.
About the only thing it will do is provide a radiator to couple RF
into the surrounding wiring.

On a boat it is almost impossible to have the room for a resonant
radial that is independent of ground.

A radial needs to be resonant if not coupled to ground.

The alternative is to tightly couple to ground. To couple to ground
you need to get things down close to ground (ground in this case being
the sea)
whether you couple to ground with radials or directly couple with
metal in contact with the water.
Direct contact with the water is the most efficient as the
conductivity of salt water is so high. It makes a much better
conductor than any soil that you could encounter on land.
Because salt water is so high in conductivity you don't need much
surface area to obtain an efficient ground contact.

Once you do have a good ground system you want the ground side of your
antenna tuner to be as close to it as practical.
If you have a long ground lead from your ground system to your tuner,
it will still function as an antenna system.

But one of the problems that you run into first of all is that the
tuner has other things connected to it besides just the antenna lead
and the ground lead.
You have the coax from the transceiver and the control cables that run
the tuner. If the tuner is connected directly to ground then so will
the coax cable and the control cables be at ground as far as the RF is
concerned. But if you have several feet of ground lead between your
tuner and ground system then that puts your coax and control cable
above ground.

That wouldn't matter except that they are tied to the antenna tuner's
ground lead!
Remember that when the ground side of the tuner is not right at ground
it then becomes part of the antenna that radiates. When that happens,
because the coax cable and control cable are tied there they become
part of the antenna!

Your transmitter is being directly coupled to the outside shield of
the coax and to that control cable. These cables may run past many
other wires on the boat. That will induce RF into these other cables
too. It will also make your transceiver hot with RF. You have one big
mess! Also whatever is in proximity to the ground lead to the tuner,
is going to get RF induced into it too.

By keeping the ground lead short you eliminate those problems. You
will get much less RF induced into other wires than you will by
allowing DIRECT coupling to them with the ground floating. Then the
only problem you have left is RF getting induced into other wiring
from the radiation of the antenna itself. Keeping the antenna and lead
as far away from other wiring and objects as possible is the thing to
strive for. Still not 100% but much easier to deal with than ground
induced RF.

At the lower frequencies a moderate length ground lead may work ok or
provide a mild nuisance with stray RF. At higher frequencies where the
long ground lead may be a significant part of a wavelength (even
though it may not seem that long) you may not have any ground at all
at the tuner if it happens to approach a quarter wavelength. With a
lead a quarter wave length long if you ground one end the other end is
effectively an open circuit. In that case you get full power into your
"other leads connected to your tuner rather than a portion of it.

In the days of the 2 mhz AM ship radios, 10 to 15 foot ground leads
were common and worked ok. That is a rather small percentage of a wave
length at 2 mhz. But with the newer equipment operating on higher
frequencies this length becomes a much more significant part of a wave
length and places things much more above ground.

There must be some carryover from the old days as to how to install
radios these days. "If it worked then it should work now". And it
would if you were to limit yourself to 2 mhz!

Looking at steel ships and how it is done on them is irrelevant. On a
large steel ship the whole thing is ground as far as the antenna is
concerned. You can put the tuner anywhere you want to and you will not
be bothered with ground currents in other conductors.

The antenna starts at ground no matter where the tuner is mounted,
whether you like it or not.
Remember, If it isn't ground then it is antenna.

More than I was going to say. You are probably bored long ago.

Regards
Gary