I think the problem with creating a formula for - what is, in essence - art, is that we can only arrive at proximates or averages.
A formula can be created for measurable matters, such as for maths or physics, because in that field it's not unclear what exactly is being perceived, counted and how.
With art or humanities in general, you can claim that x amount of certain colour in a picture evokes a response, based on a poll or study of 1,000 or 1,000,000 people, but you'd still only get the average. That same colour might not evoke any reaction at all in 10% of the population, and still in mathematical sense the formula would be seen as functioning, because it predicts the average or most likely answer.
Hence, I think that counting or defining elements that a tarot deck needs for readability will never succeed. UNLESS we try to find what type of decks are readable for "an average" reader, based on, for example, the sales stats of popular decks. This won't still ever cover all readers out there and it only gives an indication of what's the flavour in reading of that era.