Multivariate analysis - getting into shape
Maybe you had this moment in your life, you wake up in the morning crawl to the bathroom mirror and get terrified. You got fat. Nice job bucko, now you gotta fix yourself. But how to get about it? Should I spend top dollar on some magical pills that'll make me look like Tom Cruise? Or maybe I need a full-time gym membership? Is it worth looking for a brofessor?
Determine which strategy for getting into shape is the best given this set of assumptions:
- We are low on cash - the cheaper the better.
- Our time is cheap and we have a lot of it.
- We aren't afraid of needles, maybe a bit of juice is all you need.
- Although cash is low - a gym membership is at our disposal.
Multivariate analysis is an analytic tool that should in principle aid a person in the decision making process. So how does an MVA work? First you need a problem to solve. That problem can be anything, choosing a fitness routine, designing an interchange, trying to order the "best" pizza (best according to your subjective criterion). MVA is the tool for you. Of course sometimes in may be an overkill, however common sense can't be taught so let's focus on the tool. If you want to read more about this topic, I highly recommend this presentation by PhD Tadeusz Zieliński.
How to conduct a MVA
- Identify and describe in detail the problem.
- Find a domain expert.
- Pick the dimensions along which you'll analyse the problem.
- Set up a MVA table.
- Determine weight for the dimensions of analysis.
- Create solution variants for the MVA (you should have at least 3 variants = 2 edge cases and something in between).
- Score the solutions according to the weights.
- Reject solutions that don't meet the threshold.
- Pick the best solution.
Defining the problem
We aim at getting into shape. That means loosing fat, gaining muscle, generally improving our health. With the help of Mr. Piotr Węgierek, we were able to come up with five dimensions to analyse this problem.
- Time (speed) - how long does it take to get results.
- Health - how does the strategy affect our physical health (both in long and short term).
- Sanity - how does the strategy affect our sanity.
- Cash - how costly is the strategy.
- Time (required time) - how time consuming is the strategy.
- Diet + training - a moderate training routine 3 times at the gym per week + a personal diet created by a dietician based on our caloric and macro requirements, meals prepared personally using high-quality products.
- Juice + training - full cycle of anabolic steroid (AKA juice) + pumping as much iron as possible for greater gains, normal XXI century peasant diet consisting of home-made meals, fast food and weekend drinking.
- Training - no changes to the diet, just training at the gym.
- Personal trainer + training - a training routine designed by a PT, changes to the diet along PT's guidelines.
- Box diet - no training/exercise, we just get a balanced box diet for our specific needs delivered to our home everyday.
Now having dimensions of analysis and strategy variants, we use our domain expert to score each option based on their experience, domain knowledge and other wizardry. The weights were also assigned by the domain expert, interestingly sanity got the highest weight of 5. As a side note apparently a strict training routine + strict diet impacts the mind the strongest, so it's really easy to lose sanity while working on yourself. Maximum score = 150, min = 0, threshold = 100 everything below 100 isn't worth pursuing according to our initial assumptions.
|1||Time (speed)||1||slowly = 1
fast = 10
|2||Health||2||unhealthy = 1
healthy = 10
|3||Sanity||5||insane = 1
happy = 10
|4||Cash||3||expensive = 1
cheap = 10
|5||Time (required time)||4||lot of time = 1
little time = 10
Since I am not a PT, explaining why variant C scored more on dimension 3 then variant B is beyond my knowledge. That's why you need a good domain expert, that you can trust.
Things to look out for
With great power comes great responsibility. MVA is a double edged sword, by that I mean that it's easy to fool yourself by tinkering with the weights and scores. It's extremely easy to skew the weights or set up the whole analysis so that we get what we want. Unfortunately people are very susceptible to confirmation bias, and may skew the analysis unintentionally. Even worse, world is full of evil people that will do everything to further their agenda, that includes preparing dishonest MVAs' and using them as some divine wisdom. Don't be a sheep, think for yourself, be sceptical.
According to our MVA (Table 1.) the best strategy for getting into shape in case of our initial assumptions is a well balanced diet with physical exercises. Special thanks to Mr. Piotr Węgierek for his domain expertise.