Jun

2

Reading through Linchpin by Seth Godin advocates getting away from the quantified.

We measure the quantified because we can. But we should create the unquantified because it's so rare. If you can quantify it, then probably someone before you figured out a why to grind it out. And if you can grind it out, someone can grind it out cheaper than you can. On the other hand, the really valuable stuff, the stuff we pay a lot for, is unquantified. Things like creating joy or security or happiness. No easy measurements for those, thus they are art, and art is always worth more than the predicted.

Henry Gifford comments:

Similar situation with how much energy a boiler takes to heat a buildng. The quantifiable is the least important– the % efficiency while the boiler is operating (% of energy in fuel not going up the chimney, but going into heating the building).

The second most important thing is quantifiable enough to have % numbers on it, but those numbers are hopelessly inaccurate– the annual % of the energy in the fuel not going up the chimney (different from above because of the heat going up the chimney from a boiler that is hot, but not firing). Most of this loss can be avoided, but nobody bothers because the % ratings don't change with improvement.

The most important thing is so hard to quantify there is not even a unit for it– how unevenly the building is heated. The chair helped me put some units on this once, but otherwise, I have never even heard it discussed in my field. Uneven heating means overheating the whole building to satisfy the coldest room, with the overheating costing much more energy than the factors mentioned above.

So, in my field, everyone is running around measuring the least important factor, while the most important factors are not even discussed.

One solution is to focus on solving the problems, (short chimney, thermostat in every room) without being able to predict the resulting saving. But, without prediction, it's a very hard sell. Good for one's own portfolio, harder with other people's portfolios.

Peter Grieve adds:

When people are making purchasing decisions in areas in which they are not expert, there is a strong tendency to reduce things to a few numbers. These numbers are easy to compare, but present a grossly simplified picture. The simplest machine has myriad degrees of freedom, and so does the operating environment. The assessment of the interaction of these two cannot be crammed into four of five numbers.

This is especially true in government procurement, for example weapons systems. The M4A1 carbine is supposed to have an effective range of 300m. I have done a lot of shooting with the civilian version of this weapon, and I can assure you that this one number does not capture much of the essence of the range of the machine. On a hot or cold day? Shot by a master or a bumbler? Standing or prone? Against armored or unarmored targets? Shooting from sunlight or shade?

But this number is easy to compare with corresponding single numbers of other weapons.
Or take chess ratings. Chess strength is a very complicated thing, which cannot be reduced to one number. There are cases in top level chess in which A does well against B, B does well against C, and C does well against A (Nigel D. can help with examples here). Single real numbers are linearly ordered, and no three of them can satisfy A>B>C>A. Single real numbers are not adequate to encapsulate chess strength in its entirety.
 


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search