I did something similar, but I don't know which solution is more readable actually

no it doesn't

One tiny bit of code optimization that would still be possible here: You could leave out the tuple() constructor in the first map() call.

Truly magnificent! It uses exactly the same functional approach as my solution, only that yours is even more terse due to the use of nested lambdas as well as tuple assignment and map function. Especially the idea of employing map here is what I like most. My deepest respect!

magnificent!

Nice.
I have a question. Why cur_pos = initial_position does not work?! It will create a nested tuple actually. If initial_position = (0, 0), the cur_pos will be "((0, 0)),". I am looking for the reason of it.

good stuff

very nice

Simple

i traced the code, what i observed was that u are ignoring the negtive elements while addinf and assigning it to max found
but according to the description we can't do that

Yes it should and it is an obvious bug. But the tests surprisingly don't catch that.

Efficient! Should it not be .pop(0) to start from the fron of the tactics list? It might make a difference depending on the tactic set, or am I missing something?

Very nice. I note the pop() using and the switch at the end of the while loop

Impressively simple! I hopefully have learned something here!