dkates
  • dkates
  • Advanced Member Topic Starter
2016-01-11T15:46:24Z
I'm currently using the latest release version of the AI script. (It's listed on the AI updates thread as experimental version 0.32e, but unless I'm mistaken, that version was promoted to stable with the last YGOPro update.) So, here come some various notes and whatnot that I've noticed. I'm not going to mention certain issues that I know have already been reported and acknowledged. I've got one general note, and I'm commenting on each deck even if the comment amounts to "no current issues". Going from memory, so some of these may be vague or partially incorrect.

General:
  • The logic for using stat-modifiers during the Damage Step seems to sometimes be a bit off when the player's monster is in Defense Position. It generally seems like the AI is checking the ATK of the player's monster in these cases, when it should be checking the DEF.


Battlin' Boxers:
  • No major gameplay issues that I can recall, but the decklist seems like it could use an update. For example, I think it might be wise to drop Fiendish Chain and Jolt Counter, and to add in Solemn Notice.


Blackwing:
  • Again, no major gameplay issues. Decklist looks fine. Any plans to implement Blackwing - Death the Mourning Wind, Assault Blackwing - Chidori the Light Rain, and/or Blackwing - Harmattan the Sandstorm for this?


Bujin:
  • No major gameplay issues specific to this deck, though I think you have previously mentioned that general consensus is that it works better with more of an Xyz focus, which would require some significant updates to both script and decklist. I don't play Bujins myself, so I wouldn't know.


Burning Abyss:
  • No major issues. Any plans to implement Fiendish Rhino Warrior for this?


Chaos Dragons:
  • No major gameplay issues, but potential may exist for some decklist updates. Has there been any significant discussion on whether the new Felgrand cards (Felgrand the Great Divine Dragon, Paladin of Felgrand, Gospel of Revival, and/or Ruins of the Great Divine Dragon) would be effective in this?


Constellar:
  • No major gameplay issues. For my own use, I've done some experimenting with a build including Sheratan and/or Leonis, which would also incorporate Rank 3 in the Extra, but the jury's still out on how this compares to the current build in effectiveness. For now, decklist should probably stay more or less as it is.


Dark World:
  • Yet again, no major gameplay issues. The monster count seems surprisingly low; might it be helpful to drop the Upstart Goblins to add additional monsters?


Fire Fist:
  • No major gameplay issues. I kinda question this deck using Maxx "C", as this runs no other hand traps. I don't play Fire Fists myself, but I do find myself wondering whether this could incorporate Rekindling and/or Ties of the Brethren.


Frognarch:
  • Will sometimes go into Ghostrick Dullahan at times when it probably shouldn't, then Tribute Dullahan for a Tribute Summon. This is somewhat rare, and the play is justifiable if there is a Jackfrost in the Graveyard, but the AI doesn't seem to consider that.
  • Will sometimes leave itself unnecessarily open when doing plays involving Swap Frog. Given the deck's use of Battle Fader and Jackfrost, there are times when leaving the field clear instead of leaving Swap Frog on the field is the right play, but the AI does not seem to consider whether it has Fader/Jackfrost when making this decision.


Gadget:
  • No major issues. There may be potential for a decklist update. I find myself questioning the use of Call of the Haunted here. (Not an issue of it not having reason to be included, as it does have such reason, more an issue of whether it's still effective enough.) Also, could Electromagnetic Turtle be implemented for this?


Harpie:
  • No major issues. Like the Fire Fist, I somewhat question the use of Maxx "C" since the deck runs no other hand traps, but that's a minor thing at best. Maybe have those become Summoner Monk? Also like Fire Fists, I find myself wondering whether Ties of the Brethren could be effective here, though given that this deck is more dependent on its Extra Deck than the Fire Fists, that may not be advisable.


HAT:
  • Yet again, no major issues. The only thing I can even suggest is a one-card change to the decklist -- dropping the Wiretap for a Time-Space Trap Hole. (I assume Time-Space Trap Hole is not yet scripted, though.)


Heraldic Beast:
  • No major gameplay issues. Has already been mentioned as being a candidate for a decklist update, so I won't go into that in depth here.


Kozmo:
  • No major issues, though may benefit from a decklist update to include some of the Kozmo cards from after this was implemented. Again, I don't really play Kozmo, so I can't give much detail on what could be changed in that regard.


Masked Hero:
  • Chaining Mask Change to something that would remove a HERO from the field is a good play... in most cases. But when the effect in question is something like Raigeki, Dark Hole, or Torrential Tribute, and the HERO is not Bubbleman, the AI should not chain Mask Change. With Bubbleman, it should, but only if the player has at least one S/T that could be destroyed with the effect of Masked HERO Acid. (If Masked HERO Vapor is ever added, the AI should Mask Change Bubbleman to Vapor in instances like this.)
  • Earlier today, I had a situation where the AI had Mask Change face-down and a compatible HERO (Shadow Mist, which had been Summoned on the previous turn) on the field, and I ended up using Twin Twister targeting two cards (one of which was the Mask Change) in chain to the AI activating a Trap (Bottomless, I think). The AI should have chained Mask Change, but did not. I wish I could be more informative here.


MegaMonarch:
  • No major gameplay issues, and the decklist looks fine. At best, it could get some minor tweaks. I could see both versions of Thestalos being dropped, but that would be a matter of preference. The AI handles them well enough, though the fact that it can't necessarily evaluate the cards in the player's hand all that well can be a problem for Thestalos Mega. Limitation of the system, I know. But given that limitation, I wonder if the AI might be more effective with the Thestalos pair instead being appropriate versions of Mobius, Raiza, and/or additional Caius copies.
  • As another possible decklist change, this deck could potentially use Twin Twister in place of Mystical Space Typhoon. The highest priorities for the discard, not necessarily in order, would be Idea, Eidos, Pandeity, The Original Monarch, and Erebus.


Mermail:
  • Still occasionally brings out the wrong Mermail when chaining Abyss-Sphere to something that would remove Sphere from the field. Usually gets it right (meaning it gets Abysslinde), but sometimes doesn't even though it should be able to. That said, the Level 7's are the right backup option for this case, given the deck's use of Abyssgunde; however, Megalo should be higher priority than Leed, and I'm not sure this is the case.


Noble Knight:
  • Occasionally equips two copies of Excaliburn to the same monster, which is somewhat pointless. Otherwise, no gameplay issues come to mind, or at least not ones I haven't already reported.


Qliphort:
  • Nothing specific to this deck. Slightly surprising that it only runs two copies of Wavering Eyes instead of three, but that's a minor issues at best.


Tellarknight:
  • No major issues. Forbidden Chalice and Breakthrough Skill may be un-needed, but they're not hurting anything by being there. Some of those slots could be used for appropriate Trap Hole cards if Traptrix Rafflesia were to be added to this, I suppose.
Snarky
2016-01-13T17:22:35Z
Originally Posted by: dkates 

I'm currently using the latest release version of the AI script. (It's listed on the AI updates thread as experimental version 0.32e, but unless I'm mistaken, that version was promoted to stable with the last YGOPro update.)


You are correct, apparently I forgot to update the thread. Fixed that.

Quote:

The logic for using stat-modifiers during the Damage Step seems to sometimes be a bit off when the player's monster is in Defense Position. It generally seems like the AI is checking the ATK of the player's monster in these cases, when it should be checking the DEF.


This is true for some cards, I remember specifically Forbidden Chalice being handled incorrectly. Most of them should work correctly, though.

Quote:

Battlin' Boxers:

  • No major gameplay issues that I can recall, but the decklist seems like it could use an update. For example, I think it might be wise to drop Fiendish Chain and Jolt Counter, and to add in Solemn Notice.


Swapping out backrow is something I probably will let players handle on their end. If you prefer them having Notice in the deck, feel free to swap it out on your end.
Quote:


Blackwing:

  • Again, no major gameplay issues. Decklist looks fine. Any plans to implement Blackwing - Death the Mourning Wind, Assault Blackwing - Chidori the Light Rain, and/or Blackwing - Harmattan the Sandstorm for this?


Not yet, as far as I am aware, it is not even clear, how good these cards are exactly, and if they will find a way into the deck. Most of the initial impressions suggested, they were kind of mediocre, situational at best. I'll wait on that, and if some of them establish their position in the deck, I'll add them.
Quote:


Bujin:

  • No major gameplay issues specific to this deck, though I think you have previously mentioned that general consensus is that it works better with more of an Xyz focus, which would require some significant updates to both script and decklist. I don't play Bujins myself, so I wouldn't know.


I tried the XYZ focused style, and it seemed weaker to me personally. Maybe I'm playing it wrong. No changes planned so far.
Quote:


Burning Abyss:

  • No major issues. Any plans to implement Fiendish Rhino Warrior for this?


Not sure yet. Doesn't seem necessary, but all the OCG builds seem to play it.

Quote:


Chaos Dragons:

  • No major gameplay issues, but potential may exist for some decklist updates. Has there been any significant discussion on whether the new Felgrand cards (Felgrand the Great Divine Dragon, Paladin of Felgrand, Gospel of Revival, and/or Ruins of the Great Divine Dragon) would be effective in this?


No idea, but I feel like those should be handled in a dedicated deck, not shoehorned into a Chaos Dragon decklist.

Quote:


Constellar:

  • No major gameplay issues. For my own use, I've done some experimenting with a build including Sheratan and/or Leonis, which would also incorporate Rank 3 in the Extra, but the jury's still out on how this compares to the current build in effectiveness. For now, decklist should probably stay more or less as it is.


I'll just have to code Siat to be used properly, I guess :D
Quote:


Dark World:

  • Yet again, no major gameplay issues. The monster count seems surprisingly low; might it be helpful to drop the Upstart Goblins to add additional monsters?


It is some sort of turbo build, the monster count is low on purpose. The AI handles it a little incorrectly, it gets rid of its last DW monster in the hand too freely, losing access to its Graphas in the grave a lot of the time. I'll try to improve on that.

Quote:


Fire Fist:

  • No major gameplay issues. I kinda question this deck using Maxx "C", as this runs no other hand traps. I don't play Fire Fists myself, but I do find myself wondering whether this could incorporate Rekindling and/or Ties of the Brethren.


Rekindling will be used somewhat properly, I believe, feel free to use it on your end. Maxx "C" is pretty much the same situation as Solemn Notice, feel free to replace it on your end with other backrow or hand traps.
Quote:


Frognarch:

  • Will sometimes go into Ghostrick Dullahan at times when it probably shouldn't, then Tribute Dullahan for a Tribute Summon. This is somewhat rare, and the play is justifiable if there is a Jackfrost in the Graveyard, but the AI doesn't seem to consider that.
  • Will sometimes leave itself unnecessarily open when doing plays involving Swap Frog. Given the deck's use of Battle Fader and Jackfrost, there are times when leaving the field clear instead of leaving Swap Frog on the field is the right play, but the AI does not seem to consider whether it has Fader/Jackfrost when making this decision.


I'll take a look at those.
Quote:


Gadget:

  • No major issues. There may be potential for a decklist update. I find myself questioning the use of Call of the Haunted here. (Not an issue of it not having reason to be included, as it does have such reason, more an issue of whether it's still effective enough.) Also, could Electromagnetic Turtle be implemented for this?


Again, feel free to swap out CotH. Turtle seems better in a mill-based deck, it doesn't seem all that great in Gadgets. I guess, it makes a decent discard for Fortress?

I see more potential in Speedroid Menko, lvl 4 machine handtrap that summons itself for XYZ plays. Could replace Swift Scarecrow, it does essentially the same thing but better (in most situations) and has more synergy.

Quote:


Harpie:

  • No major issues. Like the Fire Fist, I somewhat question the use of Maxx "C" since the deck runs no other hand traps, but that's a minor thing at best. Maybe have those become Summoner Monk? Also like Fire Fists, I find myself wondering whether Ties of the Brethren could be effective here, though given that this deck is more dependent on its Extra Deck than the Fire Fists, that may not be advisable.



I don't feel like Monk in Harpies is very good. It was a necessary evil, back when we didn't have enough Harpie monsters, but with the arrival of Harpist, he is just not needed anymore. And he obviously works great with Sign, but if you don't have Sign, you usually don't want to lose your valuable spells. He makes good hands slightly better, while making bad hands significantly worse, imo.

I don't see the point of Brethren in Harpies. You want to be attacking and/or XYZ summoning in Harpies a lot. Brethren prevents both, while not really fixing any of the deck's issues, like the reliance on a normal summon (still need a Harpie on the field) or the consistency issues (dead draw by itself). Yes, it pulls Harpies out of the deck, but that doesn't help at all, if you cannot use them for the turn.
Quote:


HAT:

  • Yet again, no major issues. The only thing I can even suggest is a one-card change to the decklist -- dropping the Wiretap for a Time-Space Trap Hole. (I assume Time-Space Trap Hole is not yet scripted, though.)


I am not a fan of Time-Space, but I guess, I could script it at some point.
Quote:


Heraldic Beast:

  • No major gameplay issues. Has already been mentioned as being a candidate for a decklist update, so I won't go into that in depth here.


Not much progress on that front yet, currently working on Nekroz.
Quote:


Kozmo:

  • No major issues, though may benefit from a decklist update to include some of the Kozmo cards from after this was implemented. Again, I don't really play Kozmo, so I can't give much detail on what could be changed in that regard.


I agree. I noticed some problems as well, mainly the AI not using Strawman properly. Not sure on the ratios for the new cards yet, I'll wait on that. But Tincan as a consistency boost should definitely be in there.
Quote:


Masked Hero:

  • Chaining Mask Change to something that would remove a HERO from the field is a good play... in most cases. But when the effect in question is something like Raigeki, Dark Hole, or Torrential Tribute, and the HERO is not Bubbleman, the AI should not chain Mask Change. With Bubbleman, it should, but only if the player has at least one S/T that could be destroyed with the effect of Masked HERO Acid. (If Masked HERO Vapor is ever added, the AI should Mask Change Bubbleman to Vapor in instances like this.)
  • Earlier today, I had a situation where the AI had Mask Change face-down and a compatible HERO (Shadow Mist, which had been Summoned on the previous turn) on the field, and I ended up using Twin Twister targeting two cards (one of which was the Mask Change) in chain to the AI activating a Trap (Bottomless, I think). The AI should have chained Mask Change, but did not. I wish I could be more informative here.


Lots of stuff still not working correctly in that deck. Vapor is bad and doesn't deserve a spot just for that situation. I tried to cover a lot of different situations where Mask Change could be used, but some of them simply don't work at all, others need improvements like you describe.
Quote:


MegaMonarch:

  • No major gameplay issues, and the decklist looks fine. At best, it could get some minor tweaks. I could see both versions of Thestalos being dropped, but that would be a matter of preference. The AI handles them well enough, though the fact that it can't necessarily evaluate the cards in the player's hand all that well can be a problem for Thestalos Mega. Limitation of the system, I know. But given that limitation, I wonder if the AI might be more effective with the Thestalos pair instead being appropriate versions of Mobius, Raiza, and/or additional Caius copies.
  • As another possible decklist change, this deck could potentially use Twin Twister in place of Mystical Space Typhoon. The highest priorities for the discard, not necessarily in order, would be Idea, Eidos, Pandeity, The Original Monarch, and Erebus.


Thestalos' main nice is his ability to be played on turn 1 to establish a field and the Dominion lock. Cannot really do that with Mobius, Raiza or Caius. Granted, you have Aither and Erebus for that, too. Anyway, feel free to swap the Monarch ratios around a bit, they are used in Frog Monarchs, so the AI should be able to handle them mostly. Same with Twin Twister, once I implement it properly (probably next experimental update)
Quote:


Mermail:

  • Still occasionally brings out the wrong Mermail when chaining Abyss-Sphere to something that would remove Sphere from the field. Usually gets it right (meaning it gets Abysslinde), but sometimes doesn't even though it should be able to. That said, the Level 7's are the right backup option for this case, given the deck's use of Abyssgunde; however, Megalo should be higher priority than Leed, and I'm not sure this is the case.


I'd rather get Leed in the grave and draw Megalo, though, instead of the other way around. I investigated the issue with the wrong monster being summoned, but couldn't get it to happen in a controlled environment yet.
Quote:


Noble Knight:

  • Occasionally equips two copies of Excaliburn to the same monster, which is somewhat pointless. Otherwise, no gameplay issues come to mind, or at least not ones I haven't already reported.


A big King with 2 Excaliburns was significantly harder to kill than 1. I guess with twin Twister being a thing, that isn't necessarily the case anymore, but still, forcing the use of multiple removal effects on the equips before being able to target him doesn't seem too bad.
Quote:


Qliphort:

  • Nothing specific to this deck. Slightly surprising that it only runs two copies of Wavering Eyes instead of three, but that's a minor issues at best.


Feel free to add a 3rd on your end :)
Quote:


Tellarknight:

  • No major issues. Forbidden Chalice and Breakthrough Skill may be un-needed, but they're not hurting anything by being there. Some of those slots could be used for appropriate Trap Hole cards if Traptrix Rafflesia were to be added to this, I suppose.


And again, feel free to change it on your end. Rafflesia should be handled properly in any deck using a couple of trap holes.


Thanks for the notes. As mentioned, I am currently working on updating Nekroz. The previous ritual summoning logic didn't really work out and leaves much to be desired. So I am pretty much creating a new logic from scratch. Might take a while.
Similar Topics
Users browsing this topic