Stuart, we have a Honeywell thermostat with AIR (Adaptive Intelligent Recovery) too. Honeywell basically states that AIR will do it's best to reach the set point BY the time you set as opposed to starting to reach the set point AT the time you set. I have found nothing in Honeywell literature that states that AIR will reach the set point during recovery no matter how long it takes as long as the system is "catching up" to the set point without using auxiliary heat (in order to save the homeowner money). This doesn't mean that this isn't the case, it just means that I have not found it. I also have not found anything that states that auxiliary heat WILL go on with a 1,2,3,etc. degree difference either.

How fast the system is catching up and the number of degrees the system needs to catch up may be factored into auxiliary heat engagement and part of the thermostat program. I assume thermostat manufacturers factor in both comfort (response time of setbacks in this case) as well as savings on auxiliary heating.

If you have found any detailed information from Honeywell on how AIR actually works please share it with us. Earlier this year, I contacted Honeywell to get more information as to when 2nd stage, as opposed to 1st stage, would engage. Honeywell would not share any information with me about the programming of when 2nd stage would be engaged (proprietary information). I assume Honeywell would not be any more forthcoming regarding auxiliary heat and AIR. Here is their phone number in case you would like to try.800.468.1502.

It would be interesting to see under what conditions your auxiliary heat would turn on (or not turn on) with a 10 degree setback using AIR. Someone that has more experience with monitoring would be better able to guide you. I assume some sort of meter with a hold function would be involved. Please let us know how you make out.

Dewayne and Bill both spent a good deal of time to benefit all of us with their "set back" research.
There are many caveats that need to be minded (location, temperature swing, auxilliary interuption, etc......), but we must concede that set backs save energy in some circumstances. 
The most notable to consider, is that in both cases the loop fields are oversized. Bill's (for heating) as he lives in a cooling dominated climate and Dewayne's because he was the installing contractor and was able to add loop with out significant expense.
Similar resuts may not occur with a tightly designed loop field.
j

Joe - you've hit a nail on the head for me. I'm basically suspicious of setbacks on any system with aux strips. Aux has its place, but daily recovery isn't it. Bill and Dewayne both being overlooped allows their systems to quickly recover from setbacks, and Bill is in a very mild winter climate.

Anyone north of I-40 with a heatpump, whether geo or not, should be very wary of purported savings from setbacks, IMO.

The temporary EWT depression and concomitant reduction of COP during a long interval of continuous operation necessitated by setback recovery may well negate the savings from the setback's temporary reduction of thermal gradient across the building envelope, even if aux is prevented from kicking in.

The more I think about systems in heating-dominated climates, the more I would recommend set-it-and-forget-it, except in cases of multiday absences from the home.

Not sure it applies here, but had a recent quote on a job where consumer has a 20 year old open loop CM (belt drive blower) heat pump. Aging well is suffering trouble and it's time to go closed or make expensive well modifications.
In reading the existing heat pump manual i discovered it was not suitable to most closed loop apps. (45* minimum). Hit the design calc software and discovered that 5 miles of loop were required (loosely $50,000) to avoid replacing the heat pump.
What's my point?
In geo, EWT has a huge influence on everything from being able to function at all -to recovery. In todays economy people want the best price, which curiously doesn't often come with the most loops.
A comment I really liked from the president of geo excel was "geo can do whatever you want as long as you design for it".
I often worry that a system designed for huge recoveries is not designed for best ROI.
j