It is currently 02 Nov 2025, 00:47
   
Text Size

Bug/issue discussion thread

A version of WOTC's game by telengard

Moderators: telengard, CCGHQ Admins

Re: Bug/issue discussion thread

Postby telengard » 22 May 2010, 19:27

Twayblade wrote:The phase depth determines aggression though right? Is that capped at 3?
Phase depth just determines how far ahead in phases the AI will look. There is no cap but which-ever is reached first the tree depth or phase depth, will cause the AI to stop going deeper.

There are some flags for AI style, but I haven't exposed those through the GUI yet. You can use them from the command line (see doc/README for more info).

Also, I posted the thread on the random warband generator, it is here:

http://www.slightlymagic.net/forum/viewtopic.php?f=51&t=2640

~telengard
Author of Dreamblade:
viewtopic.php?f=51&t=1215
User avatar
telengard
DEVELOPER
 
Posts: 383
Joined: 23 May 2009, 23:04
Has thanked: 5 times
Been thanked: 30 times

Re: Bug/issue discussion thread

Postby Twayblade » 23 May 2010, 08:45

So is there an ideal hardest setting for the standard modern computer user?
Twayblade
 
Posts: 14
Joined: 21 May 2010, 02:12
Has thanked: 0 time
Been thanked: 0 time

Re: Bug/issue discussion thread

Postby telengard » 23 May 2010, 16:36

Twayblade wrote:So is there an ideal hardest setting for the standard modern computer user?
I usually play with the defaults, but if you are a good player you may find the AI (depending on the warband) a little lacking.

The settings I mentioned *could* equate to stronger play, but I can't say I've given them a lot of play time. The other 2 settings are "aggressive" and "defensive". Depending on a lot of conditions these could make better decisions. They are just changes in expectations of dice rolls.

I'm hoping that the new multi-core AI will allow for much better play. Next release will have my first pass at that, and then lots of tweaking. :)

~telengard
Author of Dreamblade:
viewtopic.php?f=51&t=1215
User avatar
telengard
DEVELOPER
 
Posts: 383
Joined: 23 May 2009, 23:04
Has thanked: 5 times
Been thanked: 30 times

Re: Bug/issue discussion thread

Postby Twayblade » 24 May 2010, 00:00

I am definitely noticing that the ai never chooses to destroy and even seems to pass on taking the initiative to attack by moving pieces backwards then forwards instead. This is on the "aggressive" phase depth of 4. Also in situations where we are tied and ai goes second, the ai could choose to destroy a piece instead of disrupt to win the round but does not. I haven't had a piece killed by the ai in several games. I never lose any pieces or feel threatened by an imminent "attack". It would be cool to have a killer ai option, maxing all opportunities for attack and destruction even over other strategic choices. I'd be curious to see which mode was more challenging.
Twayblade
 
Posts: 14
Joined: 21 May 2010, 02:12
Has thanked: 0 time
Been thanked: 0 time

Re: Bug/issue discussion thread

Postby telengard » 24 May 2010, 00:38

Twayblade wrote:I am definitely noticing that the ai never chooses to destroy and even seems to pass on taking the initiative to attack by moving pieces backwards then forwards instead. This is on the "aggressive" phase depth of 4. Also in situations where we are tied and ai goes second, the ai could choose to destroy a piece instead of disrupt to win the round but does not. I haven't had a piece killed by the ai in several games. I never lose any pieces or feel threatened by an imminent "attack". It would be cool to have a killer ai option, maxing all opportunities for attack and destruction even over other strategic choices. I'd be curious to see which mode was more challenging.
My guess for the scenario where destroying would have won the turn, I imagine that the AI couldn't see far enough. Keep in mind, the phase depth won't come into play if the tree depth gets hit first. Looking ahead 4 phases means a quite significant # of actions to look forward.

Did you try out the AI style options via the command line?

~telengard
Author of Dreamblade:
viewtopic.php?f=51&t=1215
User avatar
telengard
DEVELOPER
 
Posts: 383
Joined: 23 May 2009, 23:04
Has thanked: 5 times
Been thanked: 30 times

Re: Bug/issue discussion thread

Postby Twayblade » 24 May 2010, 02:40

I read the readme and forum comments but making sense of it and using the command line taxes my computer savvy and enthusiasm greatly.
Twayblade
 
Posts: 14
Joined: 21 May 2010, 02:12
Has thanked: 0 time
Been thanked: 0 time

Re: Bug/issue discussion thread

Postby telengard » 24 May 2010, 03:30

Twayblade wrote:I read the readme and forum comments but making sense of it and using the command line taxes my computer savvy and enthusiasm greatly.
I'll expose them (and a few others, like random seed) via the GUI for the next release...

~telengard
Author of Dreamblade:
viewtopic.php?f=51&t=1215
User avatar
telengard
DEVELOPER
 
Posts: 383
Joined: 23 May 2009, 23:04
Has thanked: 5 times
Been thanked: 30 times

Re: Bug/issue discussion thread

Postby Anborn » 04 Jun 2010, 20:53

oh how i love(d) dreamblade.

i recently found and downloaded this program and have played a hand full of games.

I think i've found 2 pieces which don't seem to work properly.

The IRP (immortals resting place) seems to be working for both players, not just allies.

also the zungar winddancer... his ability to move during the spawn phase when you spawn a 3 power creature, isn't firing.

i have only played a couple of games, and haven't fully tested the above, but you may want to take a look. everything else seems to be working like a champ.

kudos telengard to everything you have done and continue to do for the game.
Anborn
 
Posts: 56
Joined: 04 Jun 2010, 18:43
Has thanked: 0 time
Been thanked: 1 time

Re: Bug/issue discussion thread

Postby telengard » 04 Jun 2010, 22:17

Anborn wrote:oh how i love(d) dreamblade.

i recently found and downloaded this program and have played a hand full of games.
I still love it too, probably my favorite (analog) game of all time.

Anborn wrote:I think i've found 2 pieces which don't seem to work properly.

The IRP (immortals resting place) seems to be working for both players, not just allies.

also the zungar winddancer... his ability to move during the spawn phase when you spawn a 3 power creature, isn't firing.
Nice catches. I just eyeball'ed IRP and see an issue with it (which might be the reason it triggered for both). I'll have to look up the rules for allies in regards to locations. Currently, nothing is an ally to a location. So I'm surprised it worked for either player. :)

I'll investigate the Zungar Winddancer too. If either piece is fixed just by modifying the XML defs I'll post 'em here so you don't have to wait for the next release.

Anborn wrote:i have only played a couple of games, and haven't fully tested the above, but you may want to take a look. everything else seems to be working like a champ.

kudos telengard to everything you have done and continue to do for the game.
Thanks for the kind words and bug reports, glad you are enjoying the game. Anything you find as far as bugs definitely post it here and I'll fix them. Also, having the game.log for games exhibiting weird behavior can be helpful too.

~telengard
Author of Dreamblade:
viewtopic.php?f=51&t=1215
User avatar
telengard
DEVELOPER
 
Posts: 383
Joined: 23 May 2009, 23:04
Has thanked: 5 times
Been thanked: 30 times

Re: Bug/issue discussion thread

Postby telengard » 04 Jun 2010, 23:54

I can't seem to repro the Etherdance 3 issue. Are you sure the Zungar Winddancer was unengaged and the piece you were spawning had a power of exactly 3? For some reason I thought it was spawn cost 3 and was testing that and it didn't trigger and I re-read the text and it said power. I spawned a Hazardin Grenadier and it triggered.

As for the Immortals' Resting Place, after testing it, I understand the issue. The same thing happens with any ability that involves negating/modifying damage assignment. Another one such ability is Bodyguard. Here's the explanation from the README.

Code: Select all
Assigning damage is a little different than may be expected.  For instance, if you have a piece with Bodyguard along with others who can be assigned
damage, they will all show up as eligible for having damage assigned to them.  However, damage will not be *applied* to them unless the Bodyguard requirement is met.  An in effect Trample, or Grand Melee also does this.
       
Technically speaking, it is correct as far as the rules but may be a little confusing.  It is a long and complicated story as to why it is like this but mostly has to do with applying damage from a single source all at once to each mini so that modifiers can be applied.
Sorry for the confusion on that one. I spent quite a long time trying to make it more intuitive, but the rules + code won't let it happen. If Dreamblade had a 2 stage damage process I could have (i.e. assign and then apply). I did test this and the deathblow damage never made it to the creatures under the effect of the IRP. I've updated the README to mention IRP too.

If you see the Etherdance issue come up again, or any other issue just let me know. :)

~telengard
Author of Dreamblade:
viewtopic.php?f=51&t=1215
User avatar
telengard
DEVELOPER
 
Posts: 383
Joined: 23 May 2009, 23:04
Has thanked: 5 times
Been thanked: 30 times

Re: Bug/issue discussion thread

Postby Anborn » 15 Jun 2010, 19:54

telengard - gotcha, on looking back at IRP closer, i see that damage is rolled just never applied. Good times there.

same with etherdance, it seems to be working properly.

in other news... kendra vale is listed as a madness piece. a simple adjustment of that file's data field fixes that though.

kudos again on the program/game and keeping db alive.
Anborn
 
Posts: 56
Joined: 04 Jun 2010, 18:43
Has thanked: 0 time
Been thanked: 1 time

Re: Bug/issue discussion thread

Postby telengard » 15 Jun 2010, 22:35

Anborn wrote:telengard - gotcha, on looking back at IRP closer, i see that damage is rolled just never applied. Good times there.
I really did try to make that part of the game play like real life, but it literally is impossible w/ the rules and having an AI player.

Anborn wrote:same with etherdance, it seems to be working properly.

in other news... kendra vale is listed as a madness piece. a simple adjustment of that file's data field fixes that though.

kudos again on the program/game and keeping db alive.
Excellent catch, I will fix that piece. If you find yourself for some reason making some customs using the XML files, please share. :)

You're very welcome, glad you are enjoying it!

~telengard
Author of Dreamblade:
viewtopic.php?f=51&t=1215
User avatar
telengard
DEVELOPER
 
Posts: 383
Joined: 23 May 2009, 23:04
Has thanked: 5 times
Been thanked: 30 times

Re: Bug/issue discussion thread

Postby Anborn » 21 Jun 2010, 19:49

no customs here.

That said i've got a hand full of warbands i've created for fun and the ones i used back in the day for competitive play (a VP I used back in the day that destroys at will, some MV warp/charge bands, a mono fear, a almost mono passion). if others have warbands to share all the better... I'll try and type them up sometime soon for inclusion. I've yet to find a warband that the AI playes well.

The only "wish" for the program i'ld make is the ability to play online. the AI while ok, is no substitute for live play. regardless of the warband I play, or it plays, so far i'm batting 1.000 against it. i imagine it to be a fairly large undertaking though and probably not worth the effort given the lack of playerbase. :/
Anborn
 
Posts: 56
Joined: 04 Jun 2010, 18:43
Has thanked: 0 time
Been thanked: 1 time

Re: Bug/issue discussion thread

Postby telengard » 21 Jun 2010, 23:48

Anborn wrote:no customs here.

That said i've got a hand full of warbands i've created for fun and the ones i used back in the day for competitive play (a VP I used back in the day that destroys at will, some MV warp/charge bands, a mono fear, a almost mono passion). if others have warbands to share all the better... I'll try and type them up sometime soon for inclusion. I've yet to find a warband that the AI playes well.

The only "wish" for the program i'ld make is the ability to play online. the AI while ok, is no substitute for live play. regardless of the warband I play, or it plays, so far i'm batting 1.000 against it. i imagine it to be a fairly large undertaking though and probably not worth the effort given the lack of playerbase. :/
I'll gladly include any warbands you have. The warband I play against a lot (with random warbands for me) is chessmaster, and I do lose. But I'm just an average player and was only in a few tournaments.

At some point I will add net-play. The hard work is mostly done (the engine and rules enforcement). It is some work but it is on my TODO list. The last big TODO was multi-core support which is in there now. I'm actually surprised how many hits and downloads I get (I keep track w/ statcounter). There are more players kicking around than we think. :)

Have you tried bumping up the number of AI threads and depth? The AI may play a little better tactically than it does with the defaults.

If during your plays you find any other bugs, just let me know and I'll fix them.

~telengard
Author of Dreamblade:
viewtopic.php?f=51&t=1215
User avatar
telengard
DEVELOPER
 
Posts: 383
Joined: 23 May 2009, 23:04
Has thanked: 5 times
Been thanked: 30 times

Re: Bug/issue discussion thread

Postby Anborn » 24 Jun 2010, 13:49

Telengard,

Good advice on upping the tree/phase depth numbers... that helps the AI tremenously.

In my last game I found 2 bugs and one AI behavior that probably should be looked at. The bugs were similar reguarding the Locus Drive on the Blademiller, and the AI behavior was regarding the choice to gambit.

During the 2nd half of Turn4, I had disrupted the Blademiller to A4. Immediately following during the Spawn phase of Turn5, the AI had moved the Blademiller to C2.

Later during the 1st half of Turn8, I had disrupted the Blademiller to A5. The AI moved it to A4 during a shift. Then during the spawn phase of Turn9, the Blademiller jumped to C3.

Then on the spawn phase of Turn9, I was ahead 5-2, when the AI spawned the Angel of Haste, and choose to use the Gambit, which caused the AI to immediately lose. Maybe put in a check where if the opponent has 5 victory turns, not to activate any gambit abilities?

I unfortunately do not have the game log for this game. The settings on my AI were tree depth 7, phase depth 4, AI threads 2.

I'll try and do some more testing having the AI play the same band/pieces again, and if duplicated I will provide a game log.

-Anborn
Anborn
 
Posts: 56
Joined: 04 Jun 2010, 18:43
Has thanked: 0 time
Been thanked: 1 time

PreviousNext

Return to Dreamblade

Who is online

Users browsing this forum: No registered users and 2 guests

Main Menu

User Menu

Our Partners


Who is online

In total there are 2 users online :: 0 registered, 0 hidden and 2 guests (based on users active over the past 10 minutes)
Most users ever online was 9298 on 10 Oct 2025, 12:54

Users browsing this forum: No registered users and 2 guests

Login Form