This article requires some knowledge of elementary statistics from the reader. But I hope that even those who haven’t had any statistical background will understand this. Anyway I hope it will be a good read!
Suppose we can summarize some characteristics of a person when he is sober or simply not drunk by a normal distribution with mean Ms (standing for mean sober) and variance Vs( standing for Variance sober). For example assume that characteristic is number of jokes told on a date. Actually jokes are bad example because joke is a discreet random variable so better to say speed with which one drives a car. So if a person is sober he drives a car at 60kph with some probability. Such a characteristic will be represented by a following graph:
So in the above graph X is just a speed of car when a person is sober. And if on average he drives at 60kph when he is sober then 60=Ms. Of course sometimes his mood is up and he drives faster whereas sometimes his mood is down and he drives slower or the other way around. Note that if a person’s distribution is close to y-axis his distribution will be truncated because it is probably not a good idea to assume negative values for characteristics. For example in the case of jokes: somebody can’t tell negative amount of jokes or can’t drive a car at –60kph. Of course this model is technically and graphically clumsy but my aim is not precision but rather explanation. So we can come up with many different characteristics that can be graphed like above. Consider talking, then X can be either amount of words spoken when sober on a typical day or words per minute which is just a speed when sober with which a person usually talks. Now we can come to the interesting part of the model: how does distribution changes when a person is drunk? (Important point to know is that the area under the curve is equal to one, which means that the person doesn’t do anything which is out of the curve range) (Also technical notes: Mean is just an average, while variance can be paraphrased as “differences in day to day behaviour”)
Consider a picture below
(D is the speed of the car when a person is drunk). Now when the person got drunk variance of his characteristic increased Vd (Stands for Variance drunk). So that Vd>Vs. Note however that his mean is still the same as Md=Ms (or in other words he is still the same person). In the case speed he still drives at the same speed on average even when drunk. But now with higher variance he is more likely to do some stuff he was less likely to do when he was sober. For example drunk people do things they regret, or do things they thought they never would do. Of course distribution could change in other way
In the above “upper” picture after a person got drunk his distribution got skewed to the left. So that his mean changed and now Md>Ms. You sometimes notice that you cannot recognize a person when he is drunk, it is like looking on some other person. Well no big deal, not only his Variance increased so that Vd>Vs but also he is more likely drive fast on average because his mean also changed. So this person gets really excited and risk-loving when he gets alcohol but also he gets really unstable. In the above “lower” picture the person has Md<Ms which means in case of jokes that he is kind of getting depressed when he is drunk. For example if you go to the bar you can see two or three people absolutely without a mood, but probably there are not so when they are sober, and importantly not all of them got some bad news today. So that only fraction of people who are depressed are depressed because of some bad news while other fraction is simply depressed because of the effect of the alcohol. Also important that you frequently observe drunk people mood changes from depressed to absolutely “high” in a blink of an eye that is caught by higher variance.
Now let us try to use our model and come up with some recommendations for people.
We can in general categorize each characteristic into a “bad” one and a “good” one. Of course such categorization will be highly subjective but suppose we can do that. For example getting into fight when you are drunk can be categorized as “bad” characteristic while getting a better mood while not always good at least harmless so can be categorized as “good” characteristic. So assume that bad is something that everybody or at least most people evade, while good is something that most people enjoy and find entertaining.
So the recommendation will be as follows: if your Md for bad characteristic is higher than Ms or your Vd is higher than Vs or even both, then it is probably not a good idea for you to drink much. If on the other hand your Md for bad characteristic is smaller than Ms and at the same time your Vd is smaller than Vs (because if you fight much when you are sober only having a smaller variance does not help much, and maybe even hurts the situation) probably drinking doesn’t turn you into a monster so you can possibly go on.
The same can be said in the other direction: if your Md for good characteristic is larger than Ms then people probably love to drink with you.
How you can get this statistics for yourself? Simply ask your friends and I am sure you will get a great, detailed feedback especially if your characteristics are mainly “bad” ones.
Good Luck, Hope it helps)