vault backup: 2026-04-18 20:47:30

three-point-estimation.md

@@ -0,0 +1,23 @@
+---
+id:
+aliases: []
+title: Three-Point Estimation
+tags: []
+---
+# Three-Point Estimation
+
+* best-case or optimistic
+* worst-case or pessimistic
+* most-likely
+
+[Wikipedia](https://en.wikipedia.org/wiki/Three-point_estimation)
+
+[PMI](https://www.projectmanagement.com/wikis/368763/3-points-estimating)
+
+![](https://www.tutorialspoint.com/estimation_techniques/images/three_point_estimation.jpg)
+
+A [[risk]]-neutral party would set selling price at the most likely.
+The difference between the pessimistic estimate and selling price
+is the [[value-at-risk]].
+
+Three-point estimating is considered the bare minimum in most industries.

timestamped/2026-04-18_18-44-11.md

@@ -14,8 +14,8 @@ but I think often of [[construction-estimating|estimators]]
 like I once knew well,
 who somehow managed to be in the industry for decades
 while believing that it is poor estimating
-to accept any risk in a bid,
+to accept any [[risk]] in a bid,
-that all uncertainty must be clarified.
+that all [[uncertainty]] must be clarified.
 
 Of course,
 they didn't believe that with their actions,
@@ -25,4 +25,34 @@ like you do when you're a new estimator
 who barely knows what the word means,
 you're just happy to be out of the field
 where it's hot and no one is ever happy.
+But you've got a brain between your ears,
+so you ask the estimator---
+we'll call him Steve---
+you ask Steve what the limit is,
+because obviously some scope 
+Steve cranes his neck and squints at you, mouth half open.
+This is how he expresses incredulity.
+With that invitation, you continue:
 
+Your time costs your owners about fifty dollars an hour
+including salary and burden,
+and Steve's time is worth even more.
+Between the time spent
+* quadruple checking that the question isn't answered in the documents,
+* considering whether you should wait until you have more questions,
+* typing the RFI and sending the email,
+* parsing the engineer's inevitably unhelpful response,
+* consulting with Steve for his interpretation,
+* and so on and so on,
+it's a safe bet that the average cost
+to officially resolve uncertainty of any magnitude
+is in the hundreds of dollars.
+
+Assuming a contract win rate of fifty percent,
+any cost of [[risk-mitigation]]
+greater than half the reduction in [[value-at-risk]]
+represents a loss.
+
+Steve moves the goalposts,
+says of course he didn't mean "the small stuff",
+and also that you're overcomplicating it.