Re: Explaining in the simplest mannar the power curve.
So the formula works for your description. However, in practical flying, the IAS should never change on an ILS approach and rate of descent should be directly controlled by the elevator to correct for glideslope deviations caused by changes in groundspeed, turbulence, etc. This now leaves slight power changes to keep the airspeed within tolerances.
I left out part of the explanation to address your exact question. You asked about staying on the glideslope, rather than airspeed.
My training goals are always zero deviation. Perfect airspeed, perfect localizer, perfect glideslope. I never *intentionally* accept a deviation. That said, some of the GA training materials find small airspeed variations perfectly acceptable. One of the most commonly used is by Peter Dogan, the Instrument Flight Training Manual. He recommends accepting small deviations, making power changes only when they become close to PTS limits. Generally, absent sharp wind gradients, this is possible to do, well within PTS limts.
Still, I disagree with him.
Let me step back and go over what's happening when you try to maintain proper airspeed too.
If you're below glideslope, you need to increase your excess power in order to return and stay on the glideslope. You can do this by decreasing your drag by pulling back on the yoke slightly, or you can increase your power setting.
Using the yoke is preferable, because it's quicker. However, pulling back on the yoke results in a reduction in airspeed, eventually. Adding power would fix the problem, without changing the airspeed, but it's slow.
The best technique is a combination of both. Pull back on the yoke and add a touch of power. As the power kicks in, the yoke is returned to the original position. We think about this as "pitching for glideslope and powering for airspeed", but the aerodynamics is the reverse.
The reason I originally said "airspeed reductions" to remain on glideslope, I was merely explaining why it worked, not making a recommendation.