Discuss the wiki-tag on this page. Here is the place to ask questions and propose changes.
New Comment
1 comment, sorted by

Imagine that you are trying to find a doctor in a particular specialty. You are able to think of 12 possible reasons the doctor might refuse to see you. Some are more probable than others and some are easier to minimize/solve than others. You have 5 or 6 doctors to choose from and the 12 failure modes apply to each of them differently. For instance, Dr. A may have a 25% chance of saying "no new patients" whereas Dr. B might be "50%" and Dr. C may be "80%".  What would be the recommended way to reduce the likelihood of failure without spending an inordinate amount of time mitigating things?


1 solution would be to identify the probabilities for all 12 for 2 doctors given what is currently known (for instance, if problem X might derail things 80% of the time but you can think of something that will drive it down to 50%, leave it at 80% for now) and then make a decision tree to figure out how many options should be evaluated and which of the 12 things actually need to be mitigated.

For instance, maybe 8 of the 12 things for Dr. A have low probabilities (1%-5%) and the remaining 4 are 25%-50%. Similarly, for Dr. B, perhaps 11 are 1%-5% and the 12th one is 10%. Then you know that even with this worst case scenario, the odds of both A and B turning you down are very low (e.g. 50% * 10% = .5%) and you don't have to even look at the other doctors. Then you can go back to the list of 12  things and figure out which items will increase your odds for the least amount of effort.

At that point, you can pick the desired probability of success and work at each of the mitigations in turn until you reach the desired number.