Really interesting points! You might be tempted to ask, why not play two kick serves? Maybe you cannot make a point off more than 60% of your kickers, but if your consistency is 90%, that should beat the 70%/70% for the flat serve. But no, it doesn't: The win percentage of kick/kick would be 59%, which is less than 64% for flat/flat. So in general, the % points won off a serve can matter more than its consistency. In other words: if you have a very consistent, but tame serve, you are wasting all your chances to score, because once you place the serve in the box, it counts and you can't repeat. If you hit bombs, but you miss the first time, you have one additional try for an ace, which is more successful on the long run. The formula for overall win percentage is p1p2 + q1p3p4. p1 = likelihood of 1st serve in q1 = likelihood of 1st serve miss p2 = likelihood to win the point off a 1st serve p3 = likelihood of 2nd serve in p4 = likelihood to win the point off a 2nd serve. To avoid confusion, 1st and 2nd serve refer to the order of serves, not the style. But you can enter your own likelihoods, whether you want to serve flat/flat, flat/kick, kick/kick, or even kick/flat.
Interesting idea. Worth keeping in mind too that the differential between first and second serve success (which in turn will decide whether it is fruitful to go for two first serves or not) will depend on the level of your opponent. Against a mid-level club player, a pro will barely suffer more with a first serve than a second serve. Against the same player, a 2.5 or 3 will suffer against both. In both of these cases therefore, hitting two first serves (and more double faults) is just throwing away points. It's probably players of a similar level with whom the greatest differential is seen (which might explain why first and second serves are conventional at pretty much all levels). Definitely a case where the actual empirical data will make all the difference, if one is lucky enough to have it.
