I'm interested in measuring the effectiveness of my unique radiant system in terms of  Kwh per degree days.   Not necessarily  comparing it to different designs/sizes etc.     I expected it to be a linear function  but my rough estimates indicate otherwise.
I would appreciate any comments.      

Oh boy. Data. Post it.

Yes, data is always good…please post it. While building heat loss and utility costs are proportional to delta temp (i.e., difference between indoor/outdoor temps) and HDD respectively, true thermal efficiency (i.e., how well fuel is converted to delivered heat) is more complicated and likely would not be.

Unless you are metering it separately, it's going to be tough to make any numerical model of what's going on.

Also in high-R houses using base 65F can be pretty far off, and even in "typical" houses the temperature you keep the house and the other electric loads can make the average heating/cooling balance point stray pretty far from base-65. In any house there can be month-by-month variations in solar gain that would skew the kwh/HDD number by high-single digit percentages too.

If you are using a heat pump it ain't gonna be linear. The colder it gets, the harder the little fella has to work to push the same amount of heat uphill.

Another thing that can skew things is wind. Even if the temperatures are the same, a night with wind can scrub the warm off the home and sometimes, you get air exchange issues as the wind strikes different aspects.

Hmmmm...don't know if HDD data takes wind into account.

I would say that HDD data takes everything into account...solar, wind, etc. HDD just uses historical temp data that actually occurred during some specified time period relative to some defined temp datum to determine the monthly/annual HDD data. So, whatever caused this historical temp data was taken into account. However, forecasting building BTU heat gain/loss as affected by solar, wind, etc. is a function of the building construction characteristics (e.g., fenestration location/size, building assembly R-values, infiltration sealing, etc.) and the indoor/outdoor design temps used in the analysis or model, which may or may not be the same temp data as the HDD temp data.

HDD numbers are strictly based on the mean of the binned hourly temperature data, and remain agnostic on the state of wind/sun/rain over that time period. But it takes a LOT of wind or a very high-solar-gain house to skew heat loss against HDD numbers.

Clearly there is a difference in the instantaneous load between a high-wind 30F night and a calm 30F day, but the average energy use will still hew closely to a linear difference between the daily mean-temperature and the approximated heating/cooling balance point.

The amount of built-in presumption of heat load tools for wind-washing and infiltration are generous in the extreme, and should be discounted heavily for high-R air-tight houses. For a house that leaks less than 1000cfm/50 the true difference in heat load due to the wind is "in the noise" of all the other inherent error in heat load calculations, and isn't easily measured. But in a plank sheathed 1950s house with R9 econobatts in the cavities (or less) wind is a bigger issue. The typical average heating/cooling balance point of that 1950s with a 70F interior proved statisically to be about 65F outdoor temp, but that's back in the day when a TV was a 500watt space heater. Some of the plug loads have improved in efficiency since then (CFL/LED light bulbs, hot water heaters) which would raise the balance point some, but others (refrigerators) have gotten bigger and use more power now per household than back then, which lowers the balance point. The ratio of humans per square foot has also gone down, which raises the balance point, but air tightness, windows, and insulation levels have all improved, which lowers the balance point.

Bottom line, 65F is still a good enough presumptive balance point to use as the HDD base for most existing homes, but not necessarily YOUR home. The tighter and higher-R you make it the lower the balance point, but it takes a super-insulated building to average in the 50s, at which point solar gains start to become a large day-to-day error, even on homes with modest amounts of glazed area.

A really very well written and informative HDD discussion Dana...thank you! I would say that 60F is likely the presumptive balance point for typical new construction these days. And YES, the low to mid 50F range and properly accounting for solar heat gain is where it is at once you get into well insulated passive solar building design.

Where's the data?

Was the OP just "data teasing"?

Come on ICF, give this old retired ME from 1958 but still trying to keep up with you young techs a brake! As i was going to say I am most appreciated that you three very knowledgeable posters responded. I've been lurking around here for over 3 years and have a pretty good sense who is who. It is my pleasure to converse with you all.
Let me explain; my building site is a "retirement hobby". I like to design, build, test and analyze. I treat the building sort of like a "pilot plant". It is now utilized as a weekend retreat with less than 10% occupancy during the heating season. Small (500 sqft), one floor with 2 lofts (224sqft) and gambrel roof.
The radiant floor consists of 2-~230 ft long ,1/2 diam loops, one zone with 4.5kw Electro mini-boiler, above subfloor sandwich construction.5 small windows (~45sqft) and one entrance door make the building quite tight. Design heat load ~9000.
One disclaimer, I am not doing a rigorous statistical analysis. My data was "mined" in the sense I have selected random time periods over the last 3 winters (92 days) when there was no occupancy and I am confident the power consumption is for heating purposes.
This winter which has been much colder than the previous two has seen my consumption increase from .58kwh/DD to .67kwh/DD; a 16% increase. Significant or not? Comments? Note the average daily temps this winter is 19*F versus 27*F previously. I will admit that I did NOT hold other variables constant such as boiler water temps, flow velocity or floor slab temp settings.
Question/comments positive or negative are welcome.

If your kwh/DD is increasing with colder weather, then you need to bump up your balance point. Not surprising if you are using measurements from unoccupied space.