Yep, the figure you gave are nonsense because you're mixing air, wind and ground speeds up in a salad of confusion. None of your numbers add up to anything that makes sense.

You said-"There is a 20kt wind from left to right and I start from the left at 20kts "

That implies that you are approching the square loop with a tailwind of 20 kts plus a forward airspeed of 20 knots because you're flying downwind for a total ground speed of 40 kts.

You then said- "(I) pull straight up directly iin front of me. I maintain a straight line and slow to 20 kts at the top where I turn into the wind"

Well, since you were already doing 20 kts AIRSPEED when you were flying downwind to enter into the square loop you can't very well slow down in the vetical line to 20 kts. This is the start of the confusion. But forget that for the moment. So the airplane flew along at 20kt airspeed and pulled vertical and maintained a 20kt airspeed up the line (perhaps thanks to adding power?). That's fine. As it did so the wind hitting the top of the model along with it's momentum from the 40kt ground speed kept it drifting belly first downwind at 20 kts all while the model was heading up at 20kts.

Now you turn the next corner to inverted to fly back into the wind where you said- "...at the top where I turn into the wind, lose 5 kts and continue at 15 kts plus the added wind speed of 20kt for an air speed of 35kts."

So why is this corner costing you 5 kts where the other one didn't? Let's skip any drag related speed losses in the corners for now even though they certainly occur. So the model flips inverted from the 20kt upline to flying at 20 kts of AIRSPEED into the 20kt headwind that is still coming from the left as per your original statement where both the model AND the wind are coming from the left at 20kts each. So you see where our and your confusion is coming from? So anyway, our model has just rotated to inverted and heading back the way it entered and facing the 20 kt wind from that direction. So it sits in the air upside down at zero ground speed apparently defying gravity but is actually very happy because it is still plowing through the passing air at its happy speed of 20 knts.

So from here you pull the last corner for the downline where you said- "I proceed to the down ward turn and dive straight down while chopping the throttle to a speed of 20kts. At the bottom of the leg I pull level, lose 0kts because gravity helped in keeping the speed up in the turn and proceed at 20kts minus the wind speed of 20kts coming from behind for a net air speed of 0kts and hope I can accelerate to a flying air speed before I hit the ground."

Nope, not the case at all. The model does come down at 20 kts like you say but it also is now being pushed by the drift so by the time you get to the bottom line where you pull back to level the model is being pushed canopy first "upwards" at 20 kts. Also it maintains it's speed through the corners because we're ignoring the drag losses in the turns so it pulls out with both 20 kts of airspeed PLUS the 20kts of wind speed from the tail and is immediately flying at 40 kts with out any sign of sag towards the ground.