Final review of the game

Options
2

Comments

  • Quantius
    Quantius Posts: 228 Tile Toppler
    Options
    1. RNG is RNG. And yes, the AI cascades constantly. However, the AI is still dumb and easy to beat. The issue is that there needs to be *some* feeling that you're trying to overcome an opponent rather than just matching gems to tick down hp. I think of it like mana screw in paper mtg, I'm going to straight up lose some games because of RNG. I bet you win most of your games in mtgpq though.

    2. Client does crash sometimes, but I think this is more prominent on older devices. I've had one crash this month.

    3. I agree that there are some balance issues, but they're mostly legacy issues and the rest of what you said applies to paper MTG as well. Everything that gets put on the board starts the clock to your death. Razia will kill most PWs in MTGPQ in about 10-11 swings. A vanilla 3/3 in paper MTG will kill in 7 swings. So yes, anything that hits the board MUST be dealt with, that's literally the game.

    4. It's your decision how much time to spend playing this game or any other. You do not actually have to complete or participate in all events. If you're feeling FOMO or have an internal need to "get every possible point" then no one is going to be able to solve it. It's far better to have lots of events and options for people than to have it like it used to be where you do your training ground and then have nothing but story mode grinding for 250 runes.

    5. I just closed out the client and booted it up. 19 seconds. I have an iPhone X. I really think this is more to do with older devices than anything (the X is now 2 years old).

    6. Sounds right.

    For all its faults, while MTGPQ is not a 1:1 clone of MTG, it definitely brings a level of similarity that one wouldn't expect a match-3 game to have. Maybe some of you have never played MTG or played it in a competitive setting, but it's a fast paced game where a single misplay can cost you the match and RNG plays a significant role in how things play out. Nature of the beast.
  • Mburn7
    Mburn7 Posts: 3,427 Chairperson of the Boards
    Options
    Quantius said:
    1. RNG is RNG. And yes, the AI cascades constantly. However, the AI is still dumb and easy to beat. The issue is that there needs to be *some* feeling that you're trying to overcome an opponent rather than just matching gems to tick down hp. I think of it like mana screw in paper mtg, I'm going to straight up lose some games because of RNG. I bet you win most of your games in mtgpq though.


    5. I just closed out the client and booted it up. 19 seconds. I have an iPhone X. I really think this is more to do with older devices than anything (the X is now 2 years old).

    I find the rest of what you said very interesting and I mostly agree with it, but I want to address these comments.

    1.  The AI's intelligence has been a source of a lot of debate over the years.  When I first started it was pretty decent, but its slowly gotten worse as time goes by.  For a little bit during the Ixalan block the AI was supercharged and so many people complained about how difficult it was they nerfed it waaaay down so that you can win almost every game you play.  I also agree it should be harder.

    5.  I'm on an LG G6.  That's also 2.5 years old now.  It takes well over a minute to start up.  I don't know if its the device or my connection or my other apps or what, but that's definitely not good and definitely not just an issue with older phones (that being said, I also rarely have crashes from phone calls and stuff that others do, so it may be a combination of issues).

    But thanks for a very well thought out and well written comment on these issues!
  • Quantius
    Quantius Posts: 228 Tile Toppler
    Options
    Yeah I do remember when they gave us smart greg! The problem wasn't smart greg by itself, it was that they didn't take away the cascades so it was just brutal to play against. I'm all for a smarter greg if they tone down the silly level of cascades for AI.
  • Ampersand
    Ampersand Posts: 206 Tile Toppler
    Options
    It does seem to me like Greg performs a little differently depending on event and PW. I've thought for a while that some PW's made "smarter" gem matches and plays than others.
  • Machine
    Machine Posts: 789 Critical Contributor
    Options
    Ampersand said:
    It does seem to me like Greg performs a little differently depending on event and PW. I've thought for a while that some PW's made "smarter" gem matches and plays than others.
    I absolutely feel the same way. A great example is the Terror event. Sometimes the AI will only match off color matches and loyalty matches, while in the next fight he makes the best on color move possible (and still cascades like crazy).

    While the cascades can be annoying, I consider that as a harder level for the AI.
  • Azerack
    Azerack Posts: 501 Critical Contributor
    Options
    Machine said:
    Ampersand said:
    It does seem to me like Greg performs a little differently depending on event and PW. I've thought for a while that some PW's made "smarter" gem matches and plays than others.
    I absolutely feel the same way. A great example is the Terror event. Sometimes the AI will only match off color matches and loyalty matches, while in the next fight he makes the best on color move possible (and still cascades like crazy).

    While the cascades can be annoying, I consider that as a harder level for the AI.
    Agreed. He seems a LOT easier on the player in, say, Story mode (except maybe for Chandra, who seems to get a LOT more cascades and such if you're trying out some crazy new deck) then, say, some of the PvE events where the 3rd node is just ruthless.  
  • Mburn7
    Mburn7 Posts: 3,427 Chairperson of the Boards
    Options
    starfall said:
    I've noticed Greg taking more match5's in this FIRF than he does in most games, but I put it down to confirmation bias. Why would Greg play better in FIRF anyway?

    I think Greg takes random moves on the board, and we, being humans, look for patterns in his randomness.
    I have a weird memory that Greg is supposed to get smarter the better you do in an event.  So he is dumb early in an event and then when the perfects start piling up he gets better.

    I have no idea if that's actually the case and how long ago that post was, but maybe that explains it?
  • OmegaLolrus
    OmegaLolrus Posts: 253 Mover and Shaker
    Options
    starfall said:
    In the latest TDW I've seen Greg both take match5s and leave match5's on the board.

    The ones that weird me out are when it's like... Back to back...take the match5, then ignore the other match5.
  • Volrak
    Volrak Posts: 732 Critical Contributor
    Options
    AI loyalty ability usage appears consistent with the following.  Immediately prior to each of the AI's gem matches:
    • Choose one of the 3 PW abilities, with equal chance for each.
    • Try to use that ability.  It of course does nothing in case of no valid targets, insufficient loyalty, or an ability already used this turn.
    So although I can advance no speculation on whether it's "sheer", it does appear random to me.  Under a game-value-per-development-effort kind of metric, it's not a bad approach.

    While we're at it, in most game modes, AI gem match decisions appear consistent with the following:
    • Compute the match value for each possible match (e.g. based on total mana+loyalty, opponent mana deprived, and special gems, for the immediate swap)
    • Identify the best value match
    • If the match value of the best match is above a certain threshold, then there's a chance that match will be rejected and some different match (possibly random) is taken instead.
    It appears that Oktagon have not also artificially forced the board to never cascade when the AI goes into "dumb move" mode, and any move, dumb or not, can and will sometimes cause cascades by chance.  Seeing a dumb move cause a cascade makes some folk feel certain that the AI is not only smart, but "cheating".  This is deeply ironic, since the whole point of the "dumb move" feature appears to be to slightly cripple the AI, and reduce negative sentiment resulting from fortune smiling upon the AI.
  • Volrak
    Volrak Posts: 732 Critical Contributor
    Options
    I'd like to say I can see where you're coming from, but I can't.

    Unpredictability makes a game harder; both in general, and in this particular instance.  When it's at the expense of optimality, perhaps not, but for most PWs (counter-examples exist), the optimal play does not involve spamming your first loyalty ability as soon as you can.

    Not knowing if or when Greg will use any particular ability not only adds tactical depth, it also increases the likelihood he'll use his ult.  That was far rarer when he would blow all his loyalty at the first opportunity.  Facing a PW with powerful ult and knowing they'd never use it wasn't my idea of challenging.  Not only that, but increased use of later abilities also adds variety to the game.

    The added challenge and variety of the new system make it better game value, in my view.
  • Stormcrow
    Stormcrow Posts: 462 Mover and Shaker
    Options
    starfall said:

    The more recent system in practice makes Greg an easier opponent than the older system did, and experienced players (one might argue, the majority of players?) do not want an easy opponent.

    I feel like this is only true for Nicol Bolas 1, Elspeth, and maaaybe Koth or Vraska 1 in certain matchups. Definitely isn't the case for, to name a few examples, Ob Nix, Karn, Saheeli, any version of Tezzeret, or Brokhan.
  • TomB
    TomB Posts: 269 Mover and Shaker
    Options
    I always wondered why Greg consistently uses Brokhan's 2nd ability instead of his 3rd, and then I realized he can't because (usually) he has a full hand by the time he has full loyalty so there's nowhere for the creatures he pulls to go before he can cast them... :D