Bug reports
by ubeefx
Moderators: ubeefx, beholder, melvin, ShawnieBoy, Lodici, CCGHQ Admins
Re: Bug reports
by ubeefx » 03 Jan 2011, 19:33
The persist triggered ability uses the stack.
When the Puppeteer Clique persist resolves first, the Safehold Elite is still in the graveyard.
So it can target that card and steal it.
Then the Safehold Elite triggered ability will no longer resolve because the card is not in graveyard anymore.
So this is basically a matter of ordering triggered abilities on stack.
As noted in rule differences, this is done automatically by game, in a consistent way.
Sometimes this will be in your advantage, sometimes not.
When the Puppeteer Clique persist resolves first, the Safehold Elite is still in the graveyard.
So it can target that card and steal it.
Then the Safehold Elite triggered ability will no longer resolve because the card is not in graveyard anymore.
So this is basically a matter of ordering triggered abilities on stack.
As noted in rule differences, this is done automatically by game, in a consistent way.
Sometimes this will be in your advantage, sometimes not.
Re: Bug reports
by Incantus » 03 Jan 2011, 21:44
While I understand why you automatically order triggered abilities that go on the stack (to reduce the search space), the particular ordering you implement shouldn't affect the AI. I haven't looked at the code, but at the moment the ordering seems random (but consistently random, so the game will always pick the same order when searching the minmax tree). Instead, you could follow APNAP order but within each player's set of triggers order them consistently randomly. Then the above rule violation wouldn't be possible (but you shouldn't affect the AI speed since it's not a new minmax branch point).
Re: Bug reports
by puxl » 04 Jan 2011, 00:56
Well, there would be more cases where this could matter, generally anything with both Persist and "come into play" effects. Question is, if this can be fixed so the play will be more according to "official" rules without adding too much complexity in the AI and related code ....
Murderous Redcap is another example, it have both Persist but it targets living creatures instead.
In theory with "good" timing you can kill something with Murderous Redcap returning to game and then pick that killed card one step later by the puppeteers.
Question is, if this can be fixed without bringing too much complexity to the code. One way could be grouping such effects, so first all interested parties will pick targets for all effects that have targets and then one step later these effects are executed - optimally "at once", but it could be sequential. It would not be perfect, but it should prevent things like combo of Murderous Redcap and Puppeteers that would not be possible in the real game.
Murderous Redcap is another example, it have both Persist but it targets living creatures instead.
In theory with "good" timing you can kill something with Murderous Redcap returning to game and then pick that killed card one step later by the puppeteers.
Question is, if this can be fixed without bringing too much complexity to the code. One way could be grouping such effects, so first all interested parties will pick targets for all effects that have targets and then one step later these effects are executed - optimally "at once", but it could be sequential. It would not be perfect, but it should prevent things like combo of Murderous Redcap and Puppeteers that would not be possible in the real game.
- puxl
- Posts: 11
- Joined: 23 Apr 2010, 18:20
- Has thanked: 0 time
- Been thanked: 0 time
Re: Bug reports
by ubeefx » 04 Jan 2011, 11:57
Implementing APNAP order as Incantus suggests is doable and does indeed not make AI do more evaluation.
But it is also clear to me that Magarena will appeal mostly to casual players like me.
But it is also clear to me that Magarena will appeal mostly to casual players like me.
Re: Bug reports
by genesis » 09 Jan 2011, 17:05
1. Pongify doesn't seem to let you target your own creatures
2. On a number of occasions where a Rite of Replication produces lots of tokens (especially when it copied creatures that themselves produced tokens when coming into play), the AI basically froze. The UI still responded but then game never proceeded to the next step, as the AI was "thinking".
3. Less of a bug but still an odd decision for an AI to make. I had a Gloomwidow, which was enchanted with Pacifism, and another creature (can't recall which one) which was freely attacking and blocking. The AI played the Grasp of Darkness targeting the Gloomwidow
2. On a number of occasions where a Rite of Replication produces lots of tokens (especially when it copied creatures that themselves produced tokens when coming into play), the AI basically froze. The UI still responded but then game never proceeded to the next step, as the AI was "thinking".
3. Less of a bug but still an odd decision for an AI to make. I had a Gloomwidow, which was enchanted with Pacifism, and another creature (can't recall which one) which was freely attacking and blocking. The AI played the Grasp of Darkness targeting the Gloomwidow
- genesis
- Posts: 6
- Joined: 09 Jan 2011, 16:51
- Has thanked: 0 time
- Been thanked: 0 time
Re: Bug reports
by ubeefx » 09 Jan 2011, 18:02
1. Indeed, with the target filtering enabled Pongify could not target your own creatures.
Because this sometimes makes sense, I changed it for next release.
2. Was this on level 8? Too many creatures can be devastating for AI, even on level 6. The combat options explode and then even the clever approach becomes slow.
3. Yes, the AI has no clue about effect of Pacifism or Narcolepsy for the moment when scoring permanents.
Because this sometimes makes sense, I changed it for next release.
2. Was this on level 8? Too many creatures can be devastating for AI, even on level 6. The combat options explode and then even the clever approach becomes slow.
3. Yes, the AI has no clue about effect of Pacifism or Narcolepsy for the moment when scoring permanents.
Re: Bug reports
by puxl » 09 Jan 2011, 20:47
Perhaps the targeting could be modified so the recommended targets (which would be enemy creatures for Pongify) are shown in green and other legal targets (those that usually does not make sense, like killing your own creature) in yellowubeefx wrote:1. Indeed, with the target filtering enabled Pongify could not target your own creatures.
Because this sometimes makes sense, I changed it for next release.
For example, if you have 3 creatures and opponent plays Defense of the Heart, sometimes it may be better to kill one of your creatures instead of letting your opponent put two creatures (which would be two strongest creatures your opponent have in library) directly into play.
- puxl
- Posts: 11
- Joined: 23 Apr 2010, 18:20
- Has thanked: 0 time
- Been thanked: 0 time
Re: Bug reports
by ubeefx » 10 Jan 2011, 08:01
Magarena supports legal targets filtering (see preferences) for the player.
This will limit the targets to what usually is best (for instance creatures of opponent for dealing damage) and allows the game to advance faster because in some cases there will be no targets due to the filtering.
When this behavior is not wanted, you can turn it off in preferences. So you can do what you want.
For some cards I can adjust the filtering behavior, when the effect can be beneficial for player too, like Pongify.
This will limit the targets to what usually is best (for instance creatures of opponent for dealing damage) and allows the game to advance faster because in some cases there will be no targets due to the filtering.
When this behavior is not wanted, you can turn it off in preferences. So you can do what you want.
For some cards I can adjust the filtering behavior, when the effect can be beneficial for player too, like Pongify.
Re: Bug reports
by MageKing17 » 14 Jan 2011, 12:18
Except Pongify wouldn't actually help there because Defense of the Heart would still see 3 creatures when its trigger resolved, due to the token.puxl wrote:For example, if you have 3 creatures and opponent plays Defense of the Heart, sometimes it may be better to kill one of your creatures instead of letting your opponent put two creatures (which would be two strongest creatures your opponent have in library) directly into play.

-
MageKing17 - Programmer
- Posts: 473
- Joined: 12 Jun 2008, 20:40
- Has thanked: 5 times
- Been thanked: 9 times
Re: Bug reports
by Aswan jaguar » 16 Jan 2011, 18:56
Is there a problem of how trable damage is assigned?AI attacks with Charging Troll 3\3 with Behemoth Sledge and Gorgon Flail totally 10/10 without double strike I block with Wall of Denial 1/9 the damage should be 9 to wall and the 1 left damage it should be assigned to me.AI was attacking with 10/10 troll and 2/3 Merfolk Seastalkers I had 9 or 10 life and after this attack -1.It seems damage was assigned 1 to wall and 9 to me?
---
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: Bug reports
by ubeefx » 16 Jan 2011, 19:46
Starting with M11, you only need to assign a single point of damage from a creature with deathtouch to each blocker, which makes deathtouch a potent combination with trample. A 4/4 with trample and deathtouch can assign one damage to a blocking 8/8 and the other 3 damage to the defending player. I think this example answers your question.
Re: Bug reports
by Aswan jaguar » 16 Jan 2011, 19:50
Thanks wasn't aware of this.Trample and deathtouch a deadly compination then.
---
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: Bug reports
by geckoteck » 20 Jan 2011, 12:54
Noble Hierarch was targetted by Into the Roil and I wanted to tap it to float mana, but I couldn't use it without first specifying what other card or ability I wanted to use it for, which was impossible since I wanted the mana to re-cast the Hierarch once it was returned to my hand.
Another bug I noticed was that when my turn ended I had 8 cards in my hand and wasn't forced to discard down to 7.
Another bug I noticed was that when my turn ended I had 8 cards in my hand and wasn't forced to discard down to 7.
- geckoteck
- Posts: 3
- Joined: 20 Jan 2011, 12:52
- Has thanked: 0 time
- Been thanked: 0 time
Re: Bug reports
by geckoteck » 20 Jan 2011, 14:05
There's some crazy error where if you save a duel, then load a duel, it messes up your deck (I had a 60 card deck with 24 land, but it got reduced to a 40 card deck with only my first 4 land) and also changes how many cards you start with (I've seen 7, 8 and 10 so far). This also happened to the opponent's deck. If you exit the program, then after you start it again you once again select "load duel", the problem persists. I first noticed this when I start a new game, didn't like my hand and selected "load duel" again (to approximate a mulligan basically).
- geckoteck
- Posts: 3
- Joined: 20 Jan 2011, 12:52
- Has thanked: 0 time
- Been thanked: 0 time
Re: Bug reports
by MageKing17 » 20 Jan 2011, 20:41
I believe Magarena disallows floating mana to simplify the AI computations.geckoteck wrote:Noble Hierarch was targetted by Into the Roil and I wanted to tap it to float mana, but I couldn't use it without first specifying what other card or ability I wanted to use it for, which was impossible since I wanted the mana to re-cast the Hierarch once it was returned to my hand.
-
MageKing17 - Programmer
- Posts: 473
- Joined: 12 Jun 2008, 20:40
- Has thanked: 5 times
- Been thanked: 9 times
Who is online
Users browsing this forum: No registered users and 42 guests