In this video we perform a step-by-step multivariable optimization for a teaching example called Statapult (DoE). In the end, we challenge you to outperform us. Please upload your attempts, questions, etc.
Perfect explanation, the best so far in RU-vid, I will use it for my 6Sigma trainings. Thank you very much! My questions: 1) Why the slope on the system hardware to get a better curve of the shots? To get the parts into the basket? 2) In base of what judgment was the selection of the fix factors and why you decide not to analyse them? For example the ball cup height or band routing? 3) How was the selection of factor levels? Just testing to see the maximum and minimum achievable results (0-300cms)? In the reality that cost money. I would like to understand why not cero could be a good operating window? 4) Why the not the use of replicates in the DOE? Is there a reason why just 16 runs without replication? 5) Does measurement system (in this case, the person telling the results) was confirmed before the experiment? Thank you once again, best regards.
Thank you for your questions! 1) Or else you cannot get a variation in the height of flight which is intuitively important. 2) In this example it was engineering judgment basically 3) It might. This is the simplest instruction video we could think of. 4) see previous answer. 5) no, also same reason