Radio Spies - new problem 251
Back to General discussions forum
Friends and Colleagues!
Unwilling to end this year with "too round" number of tasks - so here is yet another interactive puzzle, based upon the paper-and-pencil game we played in university.
According to logs our dear colleague Mathias aka gardengnome is already testing
it bombing the server with requests :)
Just wanted to note - attempts limit for it was set quite arbitrarily, upon average from several my manual runs. If you feel it is too prohibitive for automated solver - let's discuss. I myself haven't yet tried to solve it programmatically (it was partly my goal to create a task on which I can compete too).
I think the current attempt limit of 24 is fine. My heuristic - non-trivial but also not super advanced - probably finds a solution in 3 out of 4 runs within that limit.
I had to resist the temptation to invoke Z3 :).
Happy New Year everybody!
Mathias, Hi! Wow, that was fast :) I thought it will hang unsolved for a couple days (judging by my own ideas how I'd try solving it)! Thanks for test/proving this!
temptation to invoke Z3 :)
Wow once more :) I have just learned something new. By quick glance it's something reminding of good old Prolog, but my knowledge of the topic is very limited. Probably I should try to figure out how this works!
One more general observation for all interactive games: I would allow to play these games to the end without any time restrictions - for manual exploration. However, to accept them as solved they should be completed within very tight time limits. Otherwise it is possible to solve them manually and get access to the solutions.
I agree with gardengnome on a strict timelimit for an interactive game. However, when I first looked at this one I found it almost impossible to play in manual mode. If I took more than 2 seconds for a move I lost the response. Fortunately I only wanted a very quick look at the data format in order to write an automatic routine. If anyone can actually solve this problem in manual mode I have to take my hat off to them. I certainly couldn't do it. As well as having rapid reactions you would have to think at a phenomenal rate too!
This was a really good way to end the old year. Happy 2022 to one and all!
Surprising, but it just recently occurred to me that time limit could be imposed on the whole game length, rather than for single move. I'm a slowpoke. Limiting the move time was not quite convenient for manual play... Also server may impose unpredictable delays (I guess especially when connection is not reused, not quite sure). So if the approach taken in this spy game is convenient (?) let's keep with it :)
you would have to think at a phenomenal rate too
For me it took about 15-20 minutes to solve manually, with making drawings etc...
P.S. it's curious to watch these very different "approaches" - gardengnome still continue resubmitting problem, seemingly
experimenting/improving/polishing solution - while CSFPython (it looks so) decided to write solution without trying
the game first - with the aim of passing in one shot (and succeeded) :)
Indeed. Three reasons:
- I like to be happy with the program itself - structure, algorithm, naming, etc - so keep on polishing it after the initial successful solution. Done that for many of the problems.
- I've recently started to program in python (it has grown on me), and #251 is the first larger problem for which I have only submitted a python solution (#247 was only three lines of code). Lots to learn and improve.
- My solution for #251 includes a coloured representation of the board after every move, and I quite enjoy seeing the spies (they are red, but that's purely coincidental :) ) appearing during the game.
Ah, really, I noticed you switched to Python recently (though the meaning of "it grown on you" is tricky for me to catch). Perhaps we should try to add interactive demos to problems using some python-on-javascript for sake of those who wants to tackle with "something bigger" :)
And since "spies" in this game do not move, perhaps I'd try to prepare a couple games with moving opponents (like unix robots, which should be funnier to visualize. Though probably they tend to be more of "arcade" rather than "logic" style, if "arcade" could be mentioned along with turn-based games...
As for colors, I remember amusing thing - in soviet / russian etc military maps "ours" are depicted with red and "theirs" with blue. It is said also that most or all other militaries use opposite coloring (red means danger, right?) The custom is allegedly not principally related to bolsheviks/communists, but rather to the usage of captured / stolen maps either at WWI or the following civil war - though not sure how much truth is in this story :)