Amount of variables and how many categories in each is entirely up to you. You are a player, you decide.

To realise( filter) important factors need to perform factor analysis.

You got to understand something. Everything matters. There are relationships between variables as well, not only between variable and outcome. Sometimes variables that you wouldn't consider important have more affect on outcome then these you use for prediction.

Many focus on functional relationship between vb prediction and outcome or ball/rotor speed with scatter... well, sometimes it even works.

Any variable could be potentially used as a predictor. Variables do affect other variables. Variables in conjunction do affect result. If variable does affect other variables that affect result, we can say that it affects result indirectly.

Result(target) is not only observed distances ( predictions - outcome), it can be any...numbers for example, groups of numbers, map of interferencial numbers distribution, reduced variance in skips distribution.... any really.