Thread

1. hnedeotes · 1790969271 · focus · HN ↗

   I think that what makes these games beatable repeatedly is that they&#x27;re static. Not saying an algorithm properly trained won&#x27;t play better than the average player a game like MtG, or my own <a href="https:&#x2F;&#x2F;aethersummon.com" rel="nofollow">https:&#x2F;&#x2F;aethersummon.com (specially now while it has under 90 possible scrolls only) but if you have a regular release cadence (say weekly or bi-weekly) of relevant new &quot;cards&quot;, then I think the playing field is much more even for humans.

   Those new additions can invalidate the whole training data by a single new &quot;card&quot; that changes completely the dynamics and would be easy for a player to understand and incorporate but not for an algorithm (perhaps with enough compute to re-train it regularly it could) - that along with the decision trees being orders of magnitude deeper, wider and with more conditionalities than go, chess or stratego - even through the same turn with the same cards available and same table state - would probably pose much harder problems for a compute bound algo.

   1. xpct · 1790974932 · focus · HN ↗

      You can definitely try to regularize against ruleset changes by generating a bunch of cards and making the agent play in randomized subsets of those cards.

      I didn&#x27;t look for prior work on this, but my estimate is that it&#x27;s probably within 2-3 orders of magnitude of additional training compared to a static game. (Still a lot!)

      1. hnedeotes · 1790976145 · focus · HN ↗

         But wouldn&#x27;t (couldn&#x27;t) the model then hallucinate play patterns and get itself into problems when playing against a real opponent?

         1. xpct · 1790977940 · focus · HN ↗

            Well, if your training includes regularization against ruleset changes, the model should simply handle it. (that would be the expensive option, and require vastly more training)

            When the Dota 2 bot was made, they retrained the bot only partially when new patches came in, so it was definitely cheaper to adapt.

Unofficial Hacker News client; not affiliated with Y Combinator.