Current Known Bugs list

[Rob Cashwalker]

The computer gave Soltari Emissary Shadow repeatedly in the same turn.

Limiting the computer to using the KPump command to once per turn may take care of this problem.

I ran into that sorta issue as well.... I think it can be done a little smarter than once per turn - if the creature already has the particular keyword, don't add it again.... There may be another creature in play with a targeted keyword pump. No way to know which ability will be chosen first, so they both have to check if the keyword is already there. Wish we could make it even smarter-- given a choice of a targeted pump and a self-pump, the computer should prefer the self-pump.

[Chris H.]

In play testing Valor with the newest version I see Valor in the cards.txt file stating:

Valor
3 W
Creature Incarnation
As long as Valor is in your graveyard, creatures you control have vigilance.
2/2
First Strike

When Valor went to the grave yard there was this statement on the stack: "1. Valor - Human creatures have Vigilence." The Puppeteer that is in play correctly states that it has first strike.

I took a peak at the code in CardFactory.java and found the following:

ability.setStackDescription(card.getName() +" - " +card.getOwner() +" creatures have Vigilance.");

I think that both CardFactory.java and cards.txt needs to be changed.

Valor correctly has first strike and is giving first strike to other creatures when Valor goes to the grave yard.

[Chris H.]

The computer put an Elvish Farmer into play. My Garruk Wildspeaker is showing Loyaity and Spore counters. The Elvish Farmer is apparently placing spore counters on both cards.

The AI was down to 4 life and it had at least 3 Saprolings in play. It failed to "Sacrifice a Saproling: You gain 2 life." I imagine that writing the AI code to handle this could be burdensome.

[GandoTheBard]

The computer put an Elvish Farmer into play. My Garruk Wildspeaker is showing Loyaity and Spore counters. The Elvish Farmer is apparently placing spore counters on both cards.

In play testing Valor with the newest version I see Valor in the cards.txt file stating:

Valor
3 W
Creature Incarnation
As long as Valor is in your graveyard, creatures you control have vigilance.
2/2
First Strike

When Valor went to the grave yard there was this statement on the stack: "1. Valor - Human creatures have Vigilence." The Puppeteer that is in play correctly states that it has first strike.

I took a peak at the code in CardFactory.java and found the following:

ability.setStackDescription(card.getName() +" - " +card.getOwner() +" creatures have Vigilance.");

I think that both CardFactory.java and cards.txt needs to be changed.

Valor correctly has first strike and is giving first strike to other creatures when Valor goes to the grave yard.

Yes both of these are already on the list.

The AI was down to 4 life and it had at least 3 Saprolings in play. It failed to "Sacrifice a Saproling: You gain 2 life." I imagine that writing the AI code to handle this could be burdensome.

This one is new.

[DennisBergkamp]

Ugh, some interesting bugs here... Valor should be an easy fix, not sure what's going on with Elvish Farmer / Garruk Wildspeaker.

[Chris H.]

-- Hooded Kavu is given fear twice, where it doesnt have a cumulative effect.

This appears to happen to other creatures that are using the KPump keyword.

--Flash is still apparently not fixed on permanents in play.

In a past version there was only three flash creatures that were having this problem. I think I remember retesting this in the following version and found that these 3 creatures were fixed.

-- There seems to be an error with some of the new pump creatures. Im not sure if its merely a text error or an actual bug.

There was several pump creatures that still had the Pump keyword when the code base changed this over to PTPump. This problem has been resolved with the current version.

-- Viashino Slasher gets +1/+1 instead of +1/-1.

This problem is now corrected.

-- Thunder Wall is missing the keyword "Flying"

-- Wall of Fire does not have the keyword Defender.

Thunder Wall now has flying and Wall of Fire now has defender.

[Almost_Clever]

Onulet casting cost should be 3 instead of 2.

[DennisBergkamp]

I just fixed Onulet, Planeswalker loyalty counter text and I even added something to Elvish Farmer and Mycologist where the AI will sac Saprolings if his life total becomes 5 or less.

Good job on the bug reports guys, keep em coming

[Almost_Clever]

I thought I had run into a problem with exalted, but in trying to isolate it, I found more evidence that first strike does not work properly. Three more examples:

I attack with Rhox Charger (3/3) boosted to 6/6 with three exalted creatures. Computer blocks with two Youthful Knights and a Steadfast Guard and I cast Elvish Fury on R.C. (now 8/8). First strike damage is assigned to R.C. (8/8 with 4 assigned damage). Regular damage is assigned and now R.C. is 8/8 with 4 damage and 6 assigned damage so it dies at end of combat.

Computer pumps Capashen Knight twice and attacks with it (3/1 first strike). I block with Akrasan Squire and pump it twice with Elvish Fury (5/5). First strike damage is assigned (A.S. is 5/5 with 3 assigned damage). Regular damage is assigned and now C.K. is 3/1 with 0 damage (and 0 assigned damage) and A.S. is 5/5 with 3 damage (and 0 assigned damage).

Computer attacks with Knight of Meadowgrain (2/2). I block with A.S. and pump it with Elvish Fury (3/3). First strike is assigned, A.S. is 3/3 with 2 assigned damage. Regular damage is assigned (but not really), Computer gains two life from K.o.M. End of combat, Computer gains two life (again) from K.o.M.

[GandoTheBard]

I thought I had run into a problem with exalted, but in trying to isolate it, I found more evidence that first strike does not work properly. Three more examples:

I attack with Rhox Charger (3/3) boosted to 6/6 with three exalted creatures. Computer blocks with two Youthful Knights and a Steadfast Guard and I cast Elvish Fury on R.C. (now 8/8). First strike damage is assigned to R.C. (8/8 with 4 assigned damage). Regular damage is assigned and now R.C. is 8/8 with 4 damage and 6 assigned damage so it dies at end of combat.

Computer pumps Capashen Knight twice and attacks with it (3/1 first strike). I block with Akrasan Squire and pump it twice with Elvish Fury (5/5). First strike damage is assigned (A.S. is 5/5 with 3 assigned damage). Regular damage is assigned and now C.K. is 3/1 with 0 damage (and 0 assigned damage) and A.S. is 5/5 with 3 damage (and 0 assigned damage).

Computer attacks with Knight of Meadowgrain (2/2). I block with A.S. and pump it with Elvish Fury (3/3). First strike is assigned, A.S. is 3/3 with 2 assigned damage. Regular damage is assigned (but not really), Computer gains two life from K.o.M. End of combat, Computer gains two life (again) from K.o.M.

Ok so it sounds like the first bug is this:
The computer cheats. wait no I mean... First Strike damage is clearly being assigned in addition to regular combat damage as if it were Double Strike.

The second bug is that sometimes after first strike damage is assigned regular damage is nulled somehow but effects that count damage do so twice anyway (see first bug). I'm guessing in an attempt to stop first strikers from dealing damage during the regular damage assignment, the program is nulling all regular damage instead of just not assigning damage from First Strikers during the regular damage assignment step.

Yes that is quite convoluted sounding but I think it is logical without having even peaked at the code. If You AC agree these are the bugs you are reporting I will put them on the first post listing.

[DennisBergkamp]

Hmmm, I'm not sure what's going on here exactly.. maybe jpb can shed more light on this matter? I will look into this though, since these seem like pretty serious bugs.

[Joku]

I've encountered a few problems with first strike that I don't think have been covered yet. First is the first strike and lifelink combo. A creature with both first strike and lifelink (like Knight of Meadowgrain, or any creature given these abilities) when attacking will trigger lifelink in both the first strike combat phase and the normal combat phase. The attacker does the correct damage, but ends up receiving twice the life back.

I also did some experiments with first strike creatures vs normal creatures. I pit a basic creature with first strike (Coiled Tinviper) against a stronger normal creature (Trained Armodon). I attacked with both types and defended with both types. In all instances, the Trained Armodon should survive and the Coiled Tinviper should be destroyed, but this did not occur in any of the cases. Whenever the Trained Armodon is the attacker, both creatures are destroyed. When I attack with the Coiled Tinviper, both creatures are destroyed. When the AI attacks with the Coiled Tinviper, both creatures survive.

[DennisBergkamp]

I've encountered a few problems with first strike that I don't think have been covered yet. First is the first strike and lifelink combo. A creature with both first strike and lifelink (like Knight of Meadowgrain, or any creature given these abilities) when attacking will trigger lifelink in both the first strike combat phase and the normal combat phase. The attacker does the correct damage, but ends up receiving twice the life back.

I have fixed this. (I'm not sure about the other first strike bugs just yet).
Looking at the code, another thing I'm pretty sure about is that if say, an unblocked Shadowmage Infiltrator gains First Strike, its controller will draw 2 cards.

[Almost_Clever]

More evidence of poor AI combat decisions.
Computer controls lots of 3/1 (Goblin Deathraiders) and 4/4 creatures (Obsidian Giant) that I hold off with a single 0/5 creature (Thallid Shell-Dweller) (this problem has been previously documented). The AI will not block my attacking Saproling tokens until I whittle its life to 7 or 8. At that point, the AI will block if I attack multiple Saprolings, but will preferentially block with Goblin Deathraiders rather than Obsidian Giants. If I attack with one Saproling token, the computer won't block until its life is down to 4, and then will always block with any available Goblin Deathraiders before it will block with an Obsidian Giant.
Given that the never anticipates any combat tricks (even those sitting on the board, such as Timberwatch Elf or Sorceress Queen), the AI should always block if it can assign lethal damage to an attacker without losing a blocker (well, except for weird cases like the incarnations), and should preferentially block without losing a creature over trading or chump blocking.

[Chris H.]

The computer had multiple Disciple of Kangee and Serra's Blessing in play.

The Disciple of Kangee had the word Vigilence displayed 4 times on their card. Once for being a Disciple of Kangee and 3 times for the 3 Serra's Blessing in play.