Welcome to RCU......

Your thinking of the Cl's is reversed. To obtain an elliptical lift distribution you need to have a progressivly lower Cl towards the tips. Not higher.

As mentioned washout can be used to delay stalling but the optimum washout twist only works at one angle of attack. Higher angles will produce the desired delay in the stall but lower angles of attack, as when trying to penetrate to get away from sink, can reduce the angles to where the tips are doing nothing or are even below the zero lift angle and are adding drag by being pulled effictively "upside down" through the air by the rest of the glider.

If you look at the polars for that family of airfoils I think you'll find that the thinner versions have a lower best Cd/Cl but that the sideways U shape is narrower than the thicker versions. This means you have an airfoil that has a much lower tolerable angle of attack range and will stall sooner than the thicker options.

If you wish to stick with the HQ series and still taper the wing towards the tips I would suggest that a more suitable setup would be the 2.5-9 out to the first break, then taper to a 2.5-8 at the next break with perhaps 1 degree of washout through that panel and finally the tip panel would taper to a 3.0-7.5 or 3.0-7 at the tip with no further washout or only slight washout. The added camber would provide the aerodynamic equivalent of washout.

My options will need some fine tuning as it's more concept than calculation as I write this off the top of my head but the idea is to have a wing with a minimum of physical washout so that the overall efficiency is not sensitive to the angle of attack like a more twisted wing would be.

I think that the idea of an elliptical lift distribution is a bit flawed. It may be perfect for aircraft that fly at one speed or as an aid to structural design but I think it's equally as important to match the wing segments to the work being done.

For example in your glider I would consider the high cruise speed typically used for exploring as much ground per foot lost as possible. Usually considered to be the best L/D but in truth I suspect most glider fliers tune for a slightly faster speed because they are working with visual cues and probably overestimate the speed needed. Anyway, at that speed I would aim for the inner 50-69% of the wing to do the major portion of the work, the next panel would taper and twist such that at the end of it the Cl was either 0 or very slightly above that value and finally the tip panel would provide no lift under these conditions, being set at the zero lift angle. My thinking being that with no or little lift there is a greatly reduced tip votex formation and thus lower drag.

Then as the angle of attack approaches the stall you need to look at how the airfoils react to the stall onset. If the tip wants to stall first then you can add camber or thickness or both to act like the aerodynamic equivalent of washout to delay the tip stalling.

Obviously this idea would require a bit of playing with something like Xfoil but hopefully you get the idea that I think it's important to study the wing's operation at the extremes of the desired speed range and juggle the factors to best satisfy that range.