suggestion about player recognition

[Chiel ten Brinke]

Also, we have a suggestion. Perhaps it would be a nice option to be
able to turn off player (and human) recognition
I would like to play the very-hard niveau, and see if i can beat the
AI's for their strategie, not their ability to try killing Human.

and sometimes... im just curious which AI would win if other AI's
wouldnt have specific anti-options build in....

I agree that the AI's in TurboRisk 1.3 were not so strong that it was
hard for a human to win; Human (and other player) recognition were
therefore permitted. But yet the AI's are much stronger. It is not
necessary to have this option anymore. (In fact it is quite dishonest
nowadays :) )

What do you think?

[Nathan Scarbrough]

I also would like the ability to turn off player recognition in the
preferences, but I think it should be on by default. I haven't given
it too much thought but I don't think that turning off player
recognition would create any bugs.

I disagree however with Cheil's statement about player recognition
being dishonest.
Vexer targets human players because they are the biggest threat, not
simply because they are human. If another AI is created that is a
bigger threat then I would change Vexer to target the new AI instead
of the human player. Vexer's goal is to win, not just to kill the
human player. For example, if there are two or more human players
then Vexer doesn't target either of them because he thinks that they
will keep each other in check and not allow the other to get too
strong. Knowing your opponents is part of the game. When you sit
across from your opponents in the traditional board game you know
their personality and style just like I know that Australian is always
going to go for Australia.

I have thought about having Vexer, frank+ and struggler+ only target a
human player if the level is set to *Very Hard*. I'm also planning on
removing the targeting code from frank+ and struggler+ when cards are
based on combo. I won't get to it for a couple months though.

I also disagree that the AI's are so strong now that we no longer need
this option. They still aren't strong enough to give me much trouble
(I'm sure I would have a harder time beating Vexer if I hadn't written
it though).

That being said, I still would like the option to disable player
recognition in the preferences because it would be interesting and
because there are people who think it's not fair. I think that it
should be on by default though, otherwise Frank wouldn't do very well
in games with Australian. Maybe there could be two options, one to
disable human player recognition and the other to disable AI to AI
recognition.

[Australian.and. Descartes]

Well I was wondering when the issue of player recognition would come
up...first time around, I was kind of puritannical about not including
player recognition, despite the fact the frank player had player
recognition for australian.

But many years on, I would mostly agree with Nathan here.

I coded a lot more player recognition second time around.

Reason being is that the players try to defend against potential
attacks from other strong players, and they view the human as an equal
threat to other strong players (but not greater). If the player
recognition was not there they would have to view all the other
players as serious potential threats, and my new players would just
build up huge armies to defend against potential threats from weak
players which they don't actually need to worry about (and which as
nathan points out, a human player would know, giving him a BIG
advantage). So the games would become more drawn-out and long. Well
actually, by default, the new players will stop defending properly, so
will become a lot worse without the player recognition. But australian
+ will do better against the original computer players including frank
+, as vexer mentioned.

So in my view, please don't add the option at all. Took long enough to
write the AI already, without adding more options....and I don't have
more time to re-code them.

I should also add - my players don't try to kill a human player any
more than the other computer players, or than they try to kill each
other. I can't comment on other programs.

[Mattias]

First off i like the idea to be able to remove player recognition, and turn it on by default, since im doing it myself constantly, and want AI's to do it too, so putting it 'on' by default is a good thing.

Note that there might be a sort of "blind" option, so that Human doesnt get to see the AI's name either, just random picked colors, and he must guess who is australian, and who is zotob (which is very easy to do, but than its too late ;) )

 

still, it is kinda unfair, since its not the first time that I have, say, 300 armies in Siam (conquerored australia) and two AI's, constantly at India and China (minus card attacks), who both have 200 armies, suddenly one starts attacking me (sorry, but this is vexer most of the time), and the other finishes it off...

just because i happen to be a human player

i prefer AI's to defend better vs humans, but to kill them off like that... ;-)

 

Also, know that turning player recognition off will only result in some functions always returning false (or -1, whatever), but not throw in special exceptions, bugs or something like it.

Your AI's will work too if you change all the program names... (if not, dont publish it =P)

It wont spoil your AI either, unless it has no good strategie on himself, just an answer to other players, but thats the fault of the AI, not player recognition

 

also, its always possible for an AI to check if a certain other AI might have an australian tactic, and thus de dangerous, you dont specifically HAVE to know the program name

your AI would be way better if he could recognize australian, wahtever program name i give him....

 

2010/11/18 Australian.and. Descartes <australian.a...@gmail.com>

[Chiel ten Brinke]

I agree with Mattias there should be completely blind play (human AND AI do not know there enemies). I agree with vexer and descartes too, that AI might still be too weak to do without player recognition, but i suppose that an experienced human player is really confused when blindly playing a game against AI :). Yet I think that the possibility to turn off player recognition for AI only is not very relevant.

[Australian.and. Descartes]

mattias, chiel - which of the computer players have you found / do you
feel are specifically trying to kill human players, more than computer
players?

[Mario]

On 19 Nov, 12:31, Chiel ten Brinke <ctenbri...@gmail.com> wrote:
> I agree with Mattias there should be completely blind play (human AND AI do
> not know there enemies).

During a Risk game (one with actual dice and cards!), you usually know
your opponents, and for sure you consider their individual strenght
and style of playing in taking your decisions. I agree with Nathan
that knowing your opponents is part of the game.

Similarely, I consider fair that in a TR game every player knows its
opponents, and set a strategy according to this information. For this
reason, I wouldn't make the plauyers blind by default.

However, I consider the idea of a completely blind game quite
interesting, it could be a nice variation of the game. I'm not sure
wether offering just the option for a completely blind game, or - as
Nathan suggested - an option for blind game and another for blind AI.
A completely blind game is something different from Risk, but still
fair, while blinding just the AI players sounds to me like
cheating :-)

As for the implementation, I think I could make PHuman return always
false, PName and PProgram return always "?". This shouldn't creat
problems to the TRPs. Additionally, I could add and SBlindGame that
returns true or false according to the kind of game being played.

Mario