about projects similar to the project
we were working on, formed a solid
basis for our HVAC design workflow.
We supplemented that “global”
research and workflow by attending
local ASHRAE chapter meetings
and talking with our peers (engineers
and contractors) to understand the
local capabilities and costs of various systems being considered for
application. That basic workflow still
works well today.
Next, when we met with an architect,
and rarely the owner, for the first time,
our workflow went something like this:
1. The architect had previously met
with the owner and determined the
type of building, its site, orientation,
structure, and materials and typically
had developed a basic floor plan of
the building based on the owner’s
program (sometimes that program
included the type of HVAC system
that was expected to be designed).
2. Using the research we had
done, we would quickly make some
back-of-napkin calculations of the
expected HVAC loads based on local
rules of thumb and experience. For
example, we often used 400 ft2/ton,
400 cfm/ton, 55°F cooling unit
discharge air temperature, 3 gpm/ton,
and 1 cfm average airflow per square
foot of occupied floor space, etc.
3. Then we would roll out the architect’s drawings and tape a layer of
flimsy vellum over them and begin to
sketch duct layouts and draw boxes
for air-handling units, VAV terminal
units, cooling towers, chillers, boilers, etc., on the vellum.
4. Next, we’d assign some sizes to
the duct and equipment based on
the rules of thumb we were using
for the project and quickly check to
see if the duct and equipment would
fit within the structure and space
allowed by the architect.
5. Finally, we’d take that information
back to the office, make some single
line drawings, develop basic equip-
ment schedules, write a narrative of
the HVAC systems (I hesitate to call it
a basis of design [BOD] given today’s
standards), and submit it along with
the other disciplines to the owner and
to a local contractor for pricing as a
schematic design (SD) package.
Things were kept simple, and
high precision was not demanded.
Owners were usually willing to carry
reasonable contingencies in their
construction budgets, and contrac-
tors adjusted their pricing based on
some level of uncertainty in the level
of detail in the SD package. People
also smoked in buildings, and no one
had given much thought to VOCs and
their impact on IEQ in those days.
Energy was considered abundant
and inexpensive, so much so that
the chairman of the regional power
company where I worked once
proclaimed that nuclear power was
going to be so abundant and inex-
pensive that there would be no need
for anyone to have a utility meter on
Then came the energy crisis of
the late 1970s and early 1980s.
Suddenly, owners, architects, engineers, policy makers, and regulators
became very interested in energy.
Using HVAC energy and economic
I use the terms accuracy, precision,
prediction, and forecast in specific and
calculated ways throughout this article.
As an engineer and building analyst, I
think we often confuse accuracy with
precision when performing BPA and
reporting the results to our clients and
In early phase modeling, accuracy of
the results when comparing alternatives
is more important for allowing the owner
and design team to make more informed
decisions than is the precision of the
forecasted performance to what may be
the actual performance. A North Carolina
State University website (http://tinyurl.
“A good analogy for understanding
accuracy and precision is to imagine a
basketball player shooting baskets. If the
player shoots with accuracy, his aim will
always take the ball close to or into the
basket. If the player shoots with precision,
his aim will always take the ball to the
same location, which may or may not be
close to the basket. A good player will be
both accurate and precise by shooting the
ball the same way each time and each
time making it in the basket.”
If our early phase comparisons use
consistent workflows and methodologies
and trusted simulation tools and provide
results that are “close enough” to known
values from prior experience or from public
databases of building performance, then
the results should be accurate enough to
inform better decisions without the rigor of
performing expensive, detailed modeling
so early in the process.
We do not need to try to forecast pre-
cise numbers such as absolute energy
consumption or indoor CO2 levels or work
surface illumination at this stage in the
design process. Most of the information
would be too soon obsolete to be useful.
I also use the term forecast instead of
the word predict when referring to esti-
mates of future building performance.
While the two terms are often seen today
as synonyms, their historical meanings
are quite different. The word predict is
something the town soothsayer might tell
you during a palm reading—a pure guess
backed up by little science.
The word forecast was more often used
when some form of science or formal ana-
lytical process was used to estimate future
performance, such as the rigorous models
the National Weather Service produces to
forecast future weather events. A forecast
is typically delivered with some statement
about the uncertainty embodied in the
results of the analysis. Since we cannot
predict the future, as engineers we should
begin reporting the results of our BPA as
forecasts and explain to our clients the
potential range of possible outcomes.
Accuracy vs. Precision, Prediction vs. Forecast