Preventing Mistakes from Happening in Calculations and Estimations
Everyone makes mistakes, in fact there is an old adage in management that you could tell a person’s worth by the number of mistakes he or she makes.
Only those that are doing something make mistakes, whereas those who do nothing don’t make mistakes either.
So, although there will always be mistakes, the issue is how do you minimize the number of mistakes.
Often mistakes are so simple, yet so costly. Take this project for example, the Mars probe built by NASA at a cost of US$125 million which was intended to study the climate on the planet Mars; as well as act as a relay for the a separate Mars Polar Lander.
The Jet Propulsion Laboratory (JPL) used the metric system of millimetres in its calculations, while Lockheed Martin who designed and built the spacecraft built the craft, used Imperial/English units of feet and inches.
So, Lockheed Martin supplied crucial acceleration data in English units of feet and pounds; which was not converted by JPL into the metric system.
Thus, when the spacecraft was deaccelerating onto Mars the wrong amount of fuel was burned causing the spacecraft to disintegrate into the Martian atmosphere.
A study of what went wrong concluded that the major contributing factor lay with the engineering function, who’s role was to track and double check all interconnected aspects, was not robust enough.
The problem with Checking Calculations and Estimates
The JPL report concluded that the testing should have been more robust. That is easy for one to say.
On the other hand, they did not provide a solution on how to make the checking more robust.
The report did not highlight the biggest problem you can encounter when checking your own, or someone else’s work.
The problem you face is that the checker tends to simply repeat the same mistakes.
This is because the checker is usually biased and one-sided and simply follows what was previously done.
So How Then Do You Prevent the Mistakes in Calculations or Estimates?
The answer is for the checker to take a completely different approach that is not reliant, or uses, the original calculations or estimates.
Now with the power of computer software, NASA should have engaged with a software company that was completely independent of JPL or Lockheed Martin to write a computer simulation package that was a computer model of the spacecraft.
Then when JPL entered the amount of fuel to burn into this computer simulated model, the results would immediately show that the spacecraft was entering the Mars atmosphere too fast and it would be destroyed.
Such simulation packages are not cheap to write but the costs of writing such a package are insignificant compared with the $125 million loss of the Mars spacecraft.
How Can This Approach be Applied to Normal Engineering Calculations and Estimates?
The same rule applies, that however the estimate is done in the first place, the checker needs to have a different approach than simply re-checking the original calculations and so potentially repeating the same mistakes.
To achieve this different approach, modern estimating packages such as produces a 3D model of the Roof; and a 3D model of Reinforcing Steel that is being estimated.
That way the checker can overview a 3D model and instantly see if parts of the roof is missing, or the reinforcing steel has gaps in it.
While such software can’t prevent all mistakes, it certainly takes the estimate to a new level of accuracy as the 3D model immediately shows what would have been the simplest mistakes such as the estimate missing a complete section out of the estimate.