Descriptive statistics
Statistical tests
Design of Experiment
Gage RR
Problem resolution
Control plan
User's Guide Menu

User Guide

Backward vs forward reasoning

As we have just seen, two types of reasoning can be used to solve a problem. Forward and backward reasoning: Though the first type of reasoning is very practical for solving "simple" problems for which we already know the solution, it does not apply if the solution is unknown. Like Sherlock Holmes, we have to be able to find clues that will allow us to find a problem's causes. This is the objective:
*“Once the impossible has been eliminated, whatever remains, however improbable, must be the truth” *
Backward reasoning, Gregg Young
The objective, therefore, is to find clues that will reveal the source of a problem. There are three types of contrasts:
  • Intra-unit: Within the same product certain contrasts exist, a portion of the product is good as long as another portion of the product is bad Unit to unit: One product is good while the next one is bad
  • Period to period: The process' performance varies over time. Sometimes 100% of products are good while, at other times, only 80% of products are good.
The multi-vari is generally the proper tool for revealing the main reason for performance variation.
To use the multi-vari, measure 3 to 5 consecutive parts regularly every hour several times a day over a few days. Generally this generates enough data to reveal the main contrasts, even if it requires more parts at times.
Once the main contrast is identified, the study can be carried out using the following tools:
  • Main contrast = intra-unit -> Defect Concentration Diagram
  • Main contrast = unit-to-unit -> Inversion or Pairwise Comparison Test
  • Main contrast = period-to-period -> Pairwise Comparison