Requests for AI improvements
Report wrong Card behavior to get it fixed.
PLEASE ADD SAVEGAMES TO YOUR TOPIC !
PLEASE ADD SAVEGAMES TO YOUR TOPIC !
Moderators: BAgate, drool66, Aswan jaguar, gmzombie, stassy, CCGHQ Admins
Re: Requests for AI improvements
by stassy » 29 Nov 2013, 10:57
AI doesn't activate Scourglass at all
- stassy
- Moderator
- Posts: 5274
- Joined: 25 Feb 2009, 07:06
- Has thanked: 471 times
- Been thanked: 337 times
Re: Requests for AI improvements
by HarlequinCasts » 08 Dec 2013, 21:53
The AI is totally irresponsible with Claws of Gix . It will just sac lands willy nilly until it happens to sac the Glaws of Gix itself. Like: Turn 1 Claws of Gix, play land, sac land. Turn 2 play land sac land etc.
I'm sure this is much like any AI who has ever cast Arcbound Ravager and sacc'ed its entire battlefield immediately.
I'm sure this is much like any AI who has ever cast Arcbound Ravager and sacc'ed its entire battlefield immediately.
-
HarlequinCasts - Posts: 922
- Joined: 07 May 2013, 14:33
- Has thanked: 68 times
- Been thanked: 30 times
Re: Requests for AI improvements
by Von Sighkick » 15 Dec 2013, 11:58
I trust it's the 110% general consensus here that the AI doesn't give 2 cent about what you, the opponent, have out on the table that could kill it on YOUR NEXT TURN! This is normally seen in the AI's ignorance of the killing potential from your normal visible creatures (but sadly also everything else i.e. lands + artifacts that can be turned into creatures, lands + artifacts + enchantments with direct damage abilities etc. etc., especially when the AI considers whether to attack itself... and with how many creatures!!!
Now I likewise trust that those of you who bought the original game when it game out in 1997 (like myself) do remember that this was NOT, repeat NOT, the case in the original game up to and including the last patch version 1.25 for the original game! I.E. back when there were still separate executables as opposed to one merged one "Magic.exe" which was introduced with the first expansion from Microprose.
In other words, Microprose programming team flat out broke the AI's "sense of defense" but first and foremost basically just it's ability to smartly and properly defend itself - earlier having taken EVERYTHING visible on your part on the table into account (unsure whether it cheated too by scrying your hand) - when they made the 1st expansion and thus merged the duel.exe with the the deck.exe etc. and of course added some brand new code.
This Microprose Manalink Community here is forever grateful to MOK for being the pioneer so to speak to undertake the daunting task of "deciphering" and "reverse engineer" the Microprose MTG source code... but personally I never ever understood why he did not pick (or at least began with) the 1.25 version of duel.exe with it's surprisingly very decent & proper AI (except for a few specific cards)... instead of the later magic.exe 2.0+ where the AI ability to defend itself etc. is BROKEN as outlined above!
To sum it all up guys, this is the ONE reason I always end up shelving this game pretty quickly after drooling over your impressive patches and new cards for a week or 2... the AI's ability to properly defend itself is so abysmal - non-existing actually since it's coding is broken - that words really can't neither describe it nor the frustration it brings, when after hours and days having constructed your newest favorite "killer deck", for example the following happens:
The AI chose to completely ignore, game after game, Vreska The Unseens's 3 Assassin creatures tokens - with whom just ONE of them damage the AI just once, the AI loses the game! regardless of darn everything incl. how many life points the AI has accumulated etc. - especially when the AI considers whether to attack with it's own creatures and how many of them...in other words just another Deja Vu of the classic broken AI defense behavior.
Not sure if Vreska and her assassins have been mentioned in the Bug Forum before, but I do understand that this is not quite the proper form to use to describe a card bug/AI request... but this entire post was far from meant as such!
This post was basically meant as a sad reminder to the hard working development team here: That no matter how many lovely card fixes, new cards and card sets you impressively add to this game... the AI will in all probability NEVER EVER be able to remotely defend it self in any remotely sound way since , as mention now several times, it has been broken since patch 1.25!!! So unless you simply "stumble upon" a way (or a new MOK!) to transfer all of old MOK's and those contributors (many unnamed) who initially stepped in his footsteps's great work somehow over to the duel.exe from patch version 1.25 and then have the present day's development team to their thang with the new C coding and other new stuff and enhancements.. then using this program as anything other than a very nice and very sentimental (but up-to-date card wise) DECK-BUILDING utility is IMHO not likely, since the major AI part of the code in magic.exe, and thus the broken part, is gonna remain more or less impossible to locate and decipher!
PS: I sincerely hope you guys from the development team do not take this post the wrong way! I have great admiration for both your dedication and hard work over the years... I simply personally have reached the point (not for the 1st time in those 16 years which have passed since it 1st saw the light of day!)once and for all I truly believe, where I have simply become TOO fed up with this broken AI... and is shelving this game for good (there are other great Deck Builders out there nowadays for Android etc.)... at least until you publish a new patch!!!

Now I likewise trust that those of you who bought the original game when it game out in 1997 (like myself) do remember that this was NOT, repeat NOT, the case in the original game up to and including the last patch version 1.25 for the original game! I.E. back when there were still separate executables as opposed to one merged one "Magic.exe" which was introduced with the first expansion from Microprose.
In other words, Microprose programming team flat out broke the AI's "sense of defense" but first and foremost basically just it's ability to smartly and properly defend itself - earlier having taken EVERYTHING visible on your part on the table into account (unsure whether it cheated too by scrying your hand) - when they made the 1st expansion and thus merged the duel.exe with the the deck.exe etc. and of course added some brand new code.
This Microprose Manalink Community here is forever grateful to MOK for being the pioneer so to speak to undertake the daunting task of "deciphering" and "reverse engineer" the Microprose MTG source code... but personally I never ever understood why he did not pick (or at least began with) the 1.25 version of duel.exe with it's surprisingly very decent & proper AI (except for a few specific cards)... instead of the later magic.exe 2.0+ where the AI ability to defend itself etc. is BROKEN as outlined above!
To sum it all up guys, this is the ONE reason I always end up shelving this game pretty quickly after drooling over your impressive patches and new cards for a week or 2... the AI's ability to properly defend itself is so abysmal - non-existing actually since it's coding is broken - that words really can't neither describe it nor the frustration it brings, when after hours and days having constructed your newest favorite "killer deck", for example the following happens:
The AI chose to completely ignore, game after game, Vreska The Unseens's 3 Assassin creatures tokens - with whom just ONE of them damage the AI just once, the AI loses the game! regardless of darn everything incl. how many life points the AI has accumulated etc. - especially when the AI considers whether to attack with it's own creatures and how many of them...in other words just another Deja Vu of the classic broken AI defense behavior.
Not sure if Vreska and her assassins have been mentioned in the Bug Forum before, but I do understand that this is not quite the proper form to use to describe a card bug/AI request... but this entire post was far from meant as such!
This post was basically meant as a sad reminder to the hard working development team here: That no matter how many lovely card fixes, new cards and card sets you impressively add to this game... the AI will in all probability NEVER EVER be able to remotely defend it self in any remotely sound way since , as mention now several times, it has been broken since patch 1.25!!! So unless you simply "stumble upon" a way (or a new MOK!) to transfer all of old MOK's and those contributors (many unnamed) who initially stepped in his footsteps's great work somehow over to the duel.exe from patch version 1.25 and then have the present day's development team to their thang with the new C coding and other new stuff and enhancements.. then using this program as anything other than a very nice and very sentimental (but up-to-date card wise) DECK-BUILDING utility is IMHO not likely, since the major AI part of the code in magic.exe, and thus the broken part, is gonna remain more or less impossible to locate and decipher!
PS: I sincerely hope you guys from the development team do not take this post the wrong way! I have great admiration for both your dedication and hard work over the years... I simply personally have reached the point (not for the 1st time in those 16 years which have passed since it 1st saw the light of day!)once and for all I truly believe, where I have simply become TOO fed up with this broken AI... and is shelving this game for good (there are other great Deck Builders out there nowadays for Android etc.)... at least until you publish a new patch!!!




- Von Sighkick
- Posts: 5
- Joined: 09 Dec 2013, 03:59
- Has thanked: 0 time
- Been thanked: 0 time
Re: Requests for AI improvements
by stassy » 15 Dec 2013, 16:37
Even if the AI was good in 1.25, I don't know if it would be able to manage new cards, especially complicated ones like Planeswalker or Profane Command type spells in specific complicated games.
The AI is still enough with anything aggro and autopilot decks, like RDW, Bant or at most Selesnya Aggro, I would be satisfied of it if it could at least be able to cast all spells when mana permit (some cards get randomly never played even if the mana requirement is met)
The AI is still enough with anything aggro and autopilot decks, like RDW, Bant or at most Selesnya Aggro, I would be satisfied of it if it could at least be able to cast all spells when mana permit (some cards get randomly never played even if the mana requirement is met)
- stassy
- Moderator
- Posts: 5274
- Joined: 25 Feb 2009, 07:06
- Has thanked: 471 times
- Been thanked: 337 times
Re: Requests for AI improvements
by Korath » 15 Dec 2013, 19:30
If the combat AI quality decreased between different releases of Microprose code, rather than between Microprose and our stuff - it's not clear from what you wrote - I might be able to figure out what they changed, given before- and after- executables. I sure haven't had much success so far working with what we currently have; and watching the AI attack with all its first-striking 1/1s at 4 life just because my Juggernaut is tapped is frustrating.
-
Korath - DEVELOPER
- Posts: 3708
- Joined: 02 Jun 2013, 05:57
- Has thanked: 496 times
- Been thanked: 1108 times
Re: Requests for AI improvements
by gmzombie » 15 Dec 2013, 22:28
im all honesty Von sighkick i was there when Mok jumped in and started to add new cards and modify the game. he chose the latest exe as it was the last one made by microprose which had alot of cards added. while i agree the ai does do some very stupid things sometimes i can also get my ass whooped by it too. the computer will never be as good as a human. Never and the reason is we dont have enough knowledge of the system as a whole yet. the Dev's past and present have done wonders for the game along with the people here who have all contributed during its life cycle. When Mok reversed it he didnt give anybody the source he had. unfortantly he lost all of that info and that is why he originally lost interest in updating the game anymore. Luckly I was able to get him (not knowing though until i returned to the forum as he originally said no) to come back and help us with the 2k fix. He always said if we got a full list of what routines to what and mapped out the program he would like to come back and help.Anyways in short mok started with the last exe which was if i recall duels of the planeswalkers or spells of the ancients whichever came last.
can I maze of ith your snowstorm?
http://home.comcast.net/~gmzombie/index.html old stuff in here. don't use this stuff right now till I get time to get back into it and readjust.
http://home.comcast.net/~gmzombie/index.html old stuff in here. don't use this stuff right now till I get time to get back into it and readjust.
- gmzombie
- Posts: 857
- Joined: 26 Feb 2009, 01:05
- Location: Wyoming, Mi
- Has thanked: 200 times
- Been thanked: 51 times
Re: Requests for AI improvements
by Aswan jaguar » 17 Dec 2013, 05:26
I have played the Original game,Unfortunately I don't have it any more neither the friend who lent it to me.I can't remember if it defended differently that it does now.
Anyway if anyone has an old PC with windows 95 or 98 or has a virtual machine with such windows and can test this please do so.If the original Microprose MTG is the problem I have an installation that will probably run I can zip it and send it.
I am trying to establish a virtul machine myself to test this(I hope I will manage to do it)as the result is interesting.
Anyway if anyone has an old PC with windows 95 or 98 or has a virtual machine with such windows and can test this please do so.If the original Microprose MTG is the problem I have an installation that will probably run I can zip it and send it.
I am trying to establish a virtul machine myself to test this(I hope I will manage to do it)as the result is interesting.
---
Trying to squash some bugs and playtesting.
Trying to squash some bugs and playtesting.
-
Aswan jaguar - Super Tester Elite
- Posts: 8131
- Joined: 13 May 2010, 12:17
- Has thanked: 748 times
- Been thanked: 479 times
Re: Requests for AI improvements
by stassy » 17 Dec 2013, 07:46
Ciroth mediafire archives seems still up btw
- stassy
- Moderator
- Posts: 5274
- Joined: 25 Feb 2009, 07:06
- Has thanked: 471 times
- Been thanked: 337 times
Re: Requests for AI improvements
by HarlequinCasts » 05 Jan 2014, 01:41
I am perhaps embarrassed to say the I really enjoy playing Battle of Wits decks. It would blow my freaking mind if we could make the AI interact with Battle of Wits at all 
As far as I can tell, no player ever plays a Battle of Wits unless they are about to win the game with it. Is there any way to increase the priority of that card from the AI's perspective? I have many many times seen the AI actually have a Disenchant or a Boomerang or a Counterspell the turn that I drop a Battle of Wits and it simply will never ever choose it as a target (often it will Disenchant or Oblivion Ring a mana artifact instead). I would love it if the AI would interact with Battle of Wits someday

As far as I can tell, no player ever plays a Battle of Wits unless they are about to win the game with it. Is there any way to increase the priority of that card from the AI's perspective? I have many many times seen the AI actually have a Disenchant or a Boomerang or a Counterspell the turn that I drop a Battle of Wits and it simply will never ever choose it as a target (often it will Disenchant or Oblivion Ring a mana artifact instead). I would love it if the AI would interact with Battle of Wits someday

-
HarlequinCasts - Posts: 922
- Joined: 07 May 2013, 14:33
- Has thanked: 68 times
- Been thanked: 30 times
Re: Requests for AI improvements
by stassy » 05 Jan 2014, 09:41
Currently its AI base value is set to 40 (aka lowest priority for enchantment), so let's try to set it to 100 to see if it does anything...
- stassy
- Moderator
- Posts: 5274
- Joined: 25 Feb 2009, 07:06
- Has thanked: 471 times
- Been thanked: 337 times
Re: Requests for AI improvements
by HarlequinCasts » 08 Jan 2014, 00:12
Hmm so far I've seen an AI not Counterspell a Battle of Wits , AI Karn Liberated avoid it, and just recently I saw an AI Disenchant its own Kismet end of turn rather than my game-winning Battle of Wits. I'll keep testing but so far I don't see the AI treating it with any more actual priority.stassy wrote:Currently its AI base value is set to 40 (aka lowest priority for enchantment), so let's try to set it to 100 to see if it does anything...
-
HarlequinCasts - Posts: 922
- Joined: 07 May 2013, 14:33
- Has thanked: 68 times
- Been thanked: 30 times
Re: Requests for AI improvements
by stassy » 08 Jan 2014, 08:53
Unfortunately I didn't have time to send the new modified AI base value to Sonic so from PTT to CIA nothing changed =/
- stassy
- Moderator
- Posts: 5274
- Joined: 25 Feb 2009, 07:06
- Has thanked: 471 times
- Been thanked: 337 times
Re: Requests for AI improvements
by HarlequinCasts » 08 Jan 2014, 15:05
No worries, I'll test it next time ^_^
-
HarlequinCasts - Posts: 922
- Joined: 07 May 2013, 14:33
- Has thanked: 68 times
- Been thanked: 30 times
Thank you Korath for AI using Deathtouch better
by Aswan jaguar » 11 Jan 2014, 10:24
Instead of a request I just want to thank you Korath so,so SO much for the help you gave to AI concerning Deathtouch and (other similar similar abilities probably) AI makes a lot of times the correct choices when attacking and defending with or against them.
I was playing with and against Typhoid Rats and I was happily surprised and while I tried to take advantage of AI playing poorly against Deathtouch AI outsmarted me and won the game in his turn
(which wouldn't have happened as AI played before).
EDIT It would be great if it is possible add a check so AI recognises that his first strikers and double strikers can deal properly with Deathtouch creatures (at least without first strike themselves)of the same or less power it would be great.That's the only big bad play from AI that I noticed.


I was playing with and against Typhoid Rats and I was happily surprised and while I tried to take advantage of AI playing poorly against Deathtouch AI outsmarted me and won the game in his turn

EDIT It would be great if it is possible add a check so AI recognises that his first strikers and double strikers can deal properly with Deathtouch creatures (at least without first strike themselves)of the same or less power it would be great.That's the only big bad play from AI that I noticed.
Last edited by Aswan jaguar on 11 Jan 2014, 10:30, edited 2 times in total.
Reason: EDIT
Reason: EDIT
---
Trying to squash some bugs and playtesting.
Trying to squash some bugs and playtesting.
-
Aswan jaguar - Super Tester Elite
- Posts: 8131
- Joined: 13 May 2010, 12:17
- Has thanked: 748 times
- Been thanked: 479 times
Who is online
Users browsing this forum: No registered users and 52 guests