Sportsmen and women at high level seem to obsess about every gram.
. . . (but it should not apply so much for most tourers, see more details below)
For rando racers (and for those many touring skiers who cannot resist competing with their partners) that obsession might make sense: concern about “every gram” which is attached to the _foot_. That’s because unlike pedaling a bicycle, climbing on skis does not have “gears”.
How fast you climb is the product of: (a) step length; (b) one separate step each for the number of legs; (c) turnover (frequency of complete stroke cycles with one step each by both legs). Since (a) is limited to some reasonable distance (not more than 1 meter), and (b) is fixed for humans (though dogs get double the number)—the main thing a racer can work on for higher climbing speed is (c) turnover frequency.
I think most touring skiers seldom achieve a frequency more than 40 per minute (of complete cycles with one step by each of the two legs), but racers can sustain 55 or more per minute. (Experienced bicyclists easily sustain 70-80).
The problem with frequency in ski-climbing is that torque load on the specific hip-flexion and knee-extension muscles for moving the ski forward depends on _more_ factors other than just ski+boot+binding weight. The specific load is also proportional to (a) the skier’s overall speed and (b) the _square_ of the turnover frequency. So ski-rando racers really are pushing the limits of torque load on those specific hip-flexion and knee-extension muscles. Anything they can do to reduce that load will yield something resembling a proportional improvement in sustainable climbing speed.
Touring skiers can avoid the problem of extra load from ski-boot-binding weight by playing this principle the opposite way: _reduce_ the turnover frequency. Suppose a skier can climb 400 meters per hour at a frequency of 40 per minute. Climbing with zig-zag switchback “conversions” at an average upward angle of 16 degrees, gives a step length of 30cm (not much, about a foot) and vertical gain per step of 8.3cm (not much, about three inches). If the same skier decreases their striding frequency to 30 per minute, but increases their step length to 36cm with 10cm vertical gain, they are climbing at 360 meters per hour.
. . . which is 10% slower than 400 meters, but the torque load on the muscles is only _half_ of what it was at 400 meters per hour (because of the much lower frequency).
Touring skiers have other strategies available to reduce load on those specific muscles _much_ more than by trimming weight off the binding.
. . . (Even if using a Diamir binding rather than Dynafit would tend to increase muscular load by as much as 10%, natural compensating motion adjustments can make the impact much much smaller.)
Also as the season progresses, each time skier does another tour, those hip-flexion and knee-extension muscles develop more strength and endurance for their special task of moving the weight of the ski forward.