Forge Release 03/10/2019 ver 1.6.22

[BloodReyvyn]

The AI tried to use the activated ability of Ensouled Scimitar while targeting its own creature with it's equip ability, which caused Flash to crash. (I had to manually close Flash through the Task Manager)

[BloodReyvyn]

The opposite happens too. I can't remember how many games were lost because I couldn't draw any terrain, and next game I would only draw terrains...
I know this situation happens in real life, but I'm usually making sure my terrains are evenly split in my deck before shuffling, being mana screwed rarely happens to me in real life.

You do realize evenly splitting your land before shuffling is either cheating or a waste of time right? Getting lots of runs of no land or all land is called variance. It has nothing to do with the shuffler being wrong.

This is absolutely right, but no one shuffling their deck is going to be this accurate when shuffling a deck either, so the deck is going to be psuedo-random anyways.

I feel like there was a similar discussion around another game (a match-3 game, probably PuzzleQuest or Gems of War or something, anyway..) about this and confirmation bias making their completely random number generator being so random that it genuinely felt like the computer was too damn lucky as, thusly, many people cried "AI CHEATS!!!" but of course, it didn't. They eventually fixed it to make the drops actually be biased in some way, but I don't actually recall exactly *how,* and the general consensus was that the game "felt" more fair and random, despite the opposite being true. Weird phenomenon.

None of this is a complaint, by the way, just an interesting parallel.

[nmg1]

One bug i noticed while playing a Mono Blue Artifacts deck (standard):
when i equip a [[Blackblade Reforged]] onto my [[Sai, Master Thopterist]] and enter combat, the game hangs 'waiting for AI'. I'm not even sure if it enters the blockers phase...

[BloodReyvyn]

The AI had Sprout Swarm, and the game just got slower and slower with each passing phase until the end of their turn where it got stuck on "Waiting for opponent..." I said "screw it" and tried to concede, but crashed and got this error.

Code: Select all

java.util.ConcurrentModificationException
   at java.util.LinkedList$ListItr.checkForComodification(Unknown Source)
   at java.util.LinkedList$ListItr.next(Unknown Source)
   at forge.ai.ComputerUtil.canRegenerate(ComputerUtil.java:876)
   at forge.ai.ComputerUtilCombat.canDestroyAttacker(ComputerUtilCombat.java:1848)
   at forge.ai.AiBlockController.getKillingBlockers(AiBlockController.java:106)
   at forge.ai.AiBlockController.makeGoodBlocks(AiBlockController.java:234)
   at forge.ai.AiBlockController.assignBlockers(AiBlockController.java:1023)
   at forge.ai.AiBlockController.assignBlockersForCombat(AiBlockController.java:950)
   at forge.ai.ComputerUtilCard.getLikelyBlockers(ComputerUtilCard.java:540)
   at forge.ai.PlayerControllerAi.chooseCardsForConvokeOrImprovise(PlayerControllerAi.java:1042)
   at forge.game.cost.CostAdjustment.adjustCostByConvokeOrImprovise(CostAdjustment.java:255)
   at forge.game.cost.CostAdjustment.adjust(CostAdjustment.java:233)
   at forge.ai.ComputerUtilMana.calculateManaCost(ComputerUtilMana.java:1144)
   at forge.ai.ComputerUtilMana.payManaCost(ComputerUtilMana.java:87)
   at forge.ai.ComputerUtilMana.canPayManaCost(ComputerUtilMana.java:51)
   at forge.ai.ComputerUtilCost.canPayCost(ComputerUtilCost.java:538)
   at forge.ai.PlayerControllerAi.chooseOptionalCosts(PlayerControllerAi.java:1185)
   at forge.ai.ComputerUtilAbility.getOriginalAndAltCostAbilities(ComputerUtilAbility.java:137)
   at forge.ai.AiController.chooseSpellAbilityToPlayFromList(AiController.java:1569)
   at forge.ai.AiController.getSpellAbilityToPlay(AiController.java:1558)
   at forge.ai.AiController.chooseSpellAbilityToPlay(AiController.java:1408)
   at forge.ai.PlayerControllerAi.chooseSpellAbilityToPlay(PlayerControllerAi.java:561)
   at forge.game.phase.PhaseHandler.startFirstTurn(PhaseHandler.java:981)
   at forge.game.GameAction.startGame(GameAction.java:1636)
   at forge.game.Match.startGame(Match.java:87)
   at forge.match.HostedMatch$2.run(HostedMatch.java:235)
   at java.util.concurrent.ThreadPoolExecutor.runWorker(Unknown Source)
   at java.util.concurrent.ThreadPoolExecutor$Worker.run(Unknown Source)
   at java.lang.Thread.run(Unknown Source)


[gos]

I cast Abnormal Endurance on my Thorn Elemental as a response to the AI targeting it with Contract Killing. It died, but did not return to the battlefield.

[AStrangerIsHere]

Wrong thread.

[jrrkein]

Forge gives me some weird error. When Cradle of Vitality activates before paying its cost and after selecting a target it shows: Targetting Error and it didnt show *insert card name been targetted* here (No shroud/hexproof).

[Kilvoctu]

I cast Final Parting and the opponent was able to see what card I put into my hand. It doesn't say on the card that this should be happening.

[RumbleBBU]

I feel like there was a similar discussion around another game (a match-3 game, probably PuzzleQuest or Gems of War or something, anyway..) about this and confirmation bias making their completely random number generator being so random that it genuinely felt like the computer was too damn lucky as, thusly, many people cried "AI CHEATS!!!" but of course, it didn't. They eventually fixed it to make the drops actually be biased in some way, but I don't actually recall exactly *how,* and the general consensus was that the game "felt" more fair and random, despite the opposite being true. Weird phenomenon.

This is getting heavily off topic but anyway...

I don't think it is even possible to generate genuinely random numbers with software alone, without the help of specialized hardware (the photon-based device is the one I am familiar with, but apparently there are other quantum technology solutions available as well). Software-based solutions are only capable of generating pseudorandom numbers. Some of the better ones (like Mersenne Twister) may pass a number of statistical tests and seem random enough but at the end of the day, when all is said and done, they are still only pseudorandom.

Both pseudorandom and genuinely random RNGs ought to be capable of creating lucky/unlucky bouts, i.e., seemingly unlikely extended sequences of lucky/unlucky random numbers. Because if something is randomly possible no matter how unlikely, it may randomly sometimes happen. With some of the not-so-good RNGs, though, these sequences may repeat themselves over and over enough so that even the user notices. You know, the "not again!" moment...

[serrasmurf]

I'm unable to return cards from exile with Gustha's Scepter
Bugged?

[psly4mne]

Both pseudorandom and genuinely random RNGs ought to be capable of creating lucky/unlucky bouts, i.e., seemingly unlikely extended sequences of lucky/unlucky random numbers. Because if something is randomly possible no matter how unlikely, it may randomly sometimes happen. With some of the not-so-good RNGs, though, these sequences may repeat themselves over and over enough so that even the user notices. You know, the "not again!" moment...

I'm sorry, you cannot tell the difference between true RNG and any PRNG in use just by playing a few thousand games. You can, however, tell the difference between biased shuffling (e.g. manaweaving followed by insufficient randomization) and proper (P)RNG shuffling. If you're accustomed to intentionally bad shuffling, a proper shuffle will feel off, but it's your experience that is wrong.

[Stampedo]

The opposite happens too. I can't remember how many games were lost because I couldn't draw any terrain, and next game I would only draw terrains...
I know this situation happens in real life, but I'm usually making sure my terrains are evenly split in my deck before shuffling, being mana screwed rarely happens to me in real life.

You do realize evenly splitting your land before shuffling is either cheating or a waste of time right? Getting lots of runs of no land or all land is called variance. It has nothing to do with the shuffler being wrong.

I'm sorry if my comment was offensive. Splitting lands before shuffling is cheating? If you shuffle with all your lands packed together, it's very likely that you'll draw them subsequently.

[friarsol]

I'm sorry if my comment was offensive. Splitting lands before shuffling is cheating? If you shuffle with all your lands packed together, it's very likely that you'll draw them subsequently.

https://www.reddit.com/r/magicTCG/comme ... ual_issue/

[RumbleBBU]

Both pseudorandom and genuinely random RNGs ought to be capable of creating lucky/unlucky bouts, i.e., seemingly unlikely extended sequences of lucky/unlucky random numbers. Because if something is randomly possible no matter how unlikely, it may randomly sometimes happen. With some of the not-so-good RNGs, though, these sequences may repeat themselves over and over enough so that even the user notices. You know, the "not again!" moment...

I'm sorry, you cannot tell the difference between true RNG and any PRNG in use just by playing a few thousand games.

You are basically right, but you will notice that I was specifically talking about bad RNG implementations in that paragraph. With RNGs that are not statistically strong enough, predictability actually becomes an issue and can, e.g., open doors to prediction attacks in cryptography. But then, of course, the requirements for a cryptographically strong RNG are higher and more specific than for a general-purpose RNG. Even the Mersenne Twister RNG, which is statistically pretty strong, cannot (should not) be used for cryptography without certain modifications.

[Stampedo]

I'm sorry if my comment was offensive. Splitting lands before shuffling is cheating? If you shuffle with all your lands packed together, it's very likely that you'll draw them subsequently.

https://www.reddit.com/r/magicTCG/comme ... ual_issue/

Thanks but it doesn't really answer my question: how do you guys make sure your lands aren't all at the same place? Even shuffling can't prevent large groups of lands.
Just curious.