Just did some very rough calcs based some some antique wind tunnel data, so hopefully I won't be shot for using it.
Winglet was found to generate a sideforce (i.e. inboard along the span) of approximately 2% of the total aircraft lift at moderate lift coefficients. The overall drag reduction attributed to the winglet was of the order of 0.2% of aircraft lift under the same conditions.
Some simple trig will reveal that if the winglet "lift" is considered to be angled 1 degree forward of true inboard, the forward acting component would be 2% * tan 1 = 0.03% of the overall lift, or just over 10% of the total winglet effect.
If one were to assume a rather generous 5 degree toe-out effect, then the forward component would be 2% * tan 5 = 0.18% of the overall lift, which would be some 90% of the overall winglet effect.
Now that latter number is clearly too high, and my expectation is that if I were to try to crunch the data to calculate the actual winglet force direction it'd be in the 1-2 degree forward range, which would imply that the forward acting force is worth some 25% of the overall winglet effect.
It would be a lot easier with access to some tip tank type data, but there's none to hand.
So, is it a huge effect? No. Even the overall winglet effect is not huge, after all. But the force effect is a significant contributor to the overall effect of the winglet on the aircraft performance.
....
And if your mysterious friend believes that only flapping aerodynamic surfaces can generate forces, I await with interest his explanation of how a (non flapping) aircraft wing works.
