TLDR: When you are wrong on estimating TIME in a phase of a construction project....how off are you?   x1.5?  x3?  x6? x 10?  And yes I know the technical answer is "IT DEPENDS" so I will ask that we skip the micro as this is a question about MACRO for the purpose of understanding risk.

Detailed:

So I am a big fan of PERT method for estimating time of work.  PERT works with an optimistic number, expected, and pessimistic number gives you a planned amount based on a bell curve (which you can skew if you choose).   I come from the software world and when you have a feature you have to estimate a distribution that has worked well for me is this.

Expected Time of feature is 1 week.

• Values a function of typical load.
• Optimistic value: .8 of expected (When its better than expected its really only 20% better)
• Expected value: 1 of expected time
• Pessimistic value: x 6 of expected time.  (When we are wrong...its not a little wrong its usually off by a lot from things we cannot predict.)

Using the Normal Distribution (Normal Bell Curve)

Normal PERT

• = (Optimistic Value + Expected Value*4 + Pessimist Value) / 6 
• = (.8 + 1*4 + 6)/6 = 10.8/6 
• = Planned Value of 1.8 Weeks
• *I use this if residual risk is low

Negatively Skewed PERT  

• = (Optimistic Value + Expected Value*3 + Pessimist Value*2) / 6
• = (.8 + 1*3 + 6*2)/6 = 15.8/6
• = Planned Value of 2.63
• *I use this if if the residual risk is medium

Very Negatively Skewed PERT

• = (Optimistic Value + Expected Value*2 + Pessimist Value*3) / 6
• = (.8 + 1*2 + 6*3)/6 = 20.8/6
• = Planned Value of 3.46
• *I use this if if the residual risk is medium

In my professional life this approach has served me very well.  So when you are wrong on estimating TIME in a phase of a construction project, in general....how off are you? x1.5? x3? x6? x 10?  And yes I know the technical answer is "IT DEPENDS" so I will ask that we skip the micro as this is a question about MACRO for the purpose of understanding risk.

Thanks I appreciate your help!