For higher accuracy it's best to use a period of 2 months when there was significant heat loading, and no cooling loads. Unless yours is a superinsulated house or purpose-designed for high passive solar gain, use 65F as the presumptive heating/cooling balance point for heating degree days. Degreedays.net is a reasonable source for nearby weather stations with good HDD data.
When you end up with a therms per HDD ratio, first convert that to output-BTU/HDD using the nameplate efficiency of the boiler (depending on vintage P204 will be anywhere from 78% to 82%. If you simply divide the D.0.E. output by the BTU-input numbers on the nameplate, that's the efficiency to use.)
With that number in hand, divide by the number of hours in a day (usually 24
) to end up with a BTU per degree-hour number.
Estimate your 99% outside design temperature based on that of nearby
listed cities, and comparing mid-winter daily low temps using
Weatherspark or similar weather history sites with those of nearby listed cities.
Subtract your 99% outside design temperature from the 65F heating base temperature for the number of heating-degrees.
Multiply your BTU/degree-hour constant by the heating degrees number- that's going to be within 10% of your actual heat load (usually within 5%.)
This method fails if you are using auxilliary heating, leave the house unoccupied for days/weeks with the thermostats set to 50F, etc. but works pretty well for occupied buildings that are heated to 68-72F during at least good fraction of the day, and keeps it of 60F or greater the rest of the time.
eg: Say your therms/HDD number works out to 0.3 therms per HDD, and you determine than your 99% outside design temp is +7F.
In an 80% boiler that's 0.8 x 100,000 BTU/therm x 0.3 therms = 24,000 BTU/HDD, or 24,000/24 hours= 1000 BTU/degree-hour.
You have 65F-7F= 58F heating degrees.
Your heat load is then about 1000 BTU/degree-hour x 58F degrees = 58,000 BTU/hr.
If your other heat load calculations used a design temperature different from the 99% outside design temp, use that to compare apples-to-apples, but using the 99% number is really the only "proper" temp to use for sizing the boiler, since the load rarely comes out EXACTLY at the output figure for the boiler, and the next size up delivers plenty of margin.