The expected resources is about 5.0 resources, same as Emergency Cache, in standard token setup. Many "5" give slightly better result, but many above 5 (-6, -8) give poor result. In this reason, this card is not suitable for expert difficulty. If you consider "fun", 21 or bust is good choice unlike Henry Wan, since 21 or bust at least achieves E-Cache.

Additionally, if your deck has Prophetic, 21 or bust may be better than E-cahce. Since 21 or bust has fortune trait, you can play 21 or bust for free with Prophetic. Sadly, Prophetic doesn't have , so that only off-class investigators can do.

Note: additional -5 or symbol token slightly increases the expected resources, and -7 or -8 token significantly decreases the expected resources.

The below is my stratagy. I believe your intuition is better, so please refer this just for fun.

Here is my stratagy to achieve E-Cache. According to the combined value of revealed tokens,

• 11 or less: reveal 1 additional token
• 12, 13: reveal 1 additional token only if and/or was revealed. (exceptional case: stop if you revealed but you bust if you treat as 11).
• 14, 15, 16: reveal 1 additional token
• 17, 18, 19: reveal 1 additional token only if was revealed and treated as 1.
• 20: reveal 1 addtional token with above condition and if 2 or more "1" tokens exist in bag.
• 21 or more: stop

Nevertheless, I believe many of you don't want to stop when the total value is 12 or 13 (including me) for fun. I also give the expectation value when you follow above stratagy but don't stop at 12/13.

NotZ Standard bag] ideal: 4.98, proposed: 4.96, proposed(non-stop): 4.86, 17-stop: 4.76
NotZ Standard bag + ] ideal: 5.01, proposed: 4.99, proposed(non-stop): 4.95, 17-stop: 4.85
NotZ Hard bag] ideal: 5.02, proposed: 5.00, proposed(non-stop): 4.97, 17-stop: 4.88
NotZ Hard bag + ] ideal: 5.07, proposed: 5.03, proposed(non-stop): 5.03, 17-stop: 4.96
Here, "ideal" means ideal stratagy. I reveal one additional token only if the expectation will be increased considering all cases. It's the upper bound if you don't know the future information.

With above stratagy, you can achieve below table in standard NotZ.

• 17stop: we stop whenever the sum is 17 or above
• 12/13/17stop: In addition to 17stop, we stop if the sume is 12 or 13 (not 14~16).

(modified Oct 8, 2021)

Sometimes in Arkham Horror we tend to get obsessed with Min-Maxing. 21 or Bust is a card whose greatest strength is that it is just fun. It's enough of a gamble that the team feels tense when you are at 17 and want to pull another token, it's not game-ruining when you fail, and it's an incredible experience for the group when you actually manage to hit 21.

Outside of gambling, this card does have some utility through the reveal token text. If you're running Favor of the Moon for curse mitigation, this is a really good card to release a sealed curse on. The curse token itself has no downside, and you gain a resource as part of the effect.

Side Note: If anyone tries to give you crap for running this instead of Emergency Cache, remind them that there are dedicated fighters out there running .35 Winchester.

Can Jim Culver treat as 0 while resolving this card? And can I choose the value of after i draw next tokens? Or I have to choose it right when I draw it? 200 characters, so have a good day if you read this!

This card is interesting thematically because, while Rogues have a lot of betting in their thematics, they tend to NOT gamble! A rogue, before this card, said "Gambling is for chumps" like the ace hustlers they were and what their cards were more about them being certain about the outcome and deciding to make calculated wagers, not gambles. They leave that for Mystics and Survivors!

It isn't going to affect the mechanics one way or the other, but it does make my rundown of Rogues for new players a bit less accurate!

This is mostly worse than our good old friend Emergency Cache. If you play it honestly and pay normally it's only better if you hit a 20 or 21, slightly worse on an 18 or lower, and much worse if you bust. I feel that it's pretty clearly only worthwhile if you have some kind of trait synergy going with fortune or gambit, or you're already doing token manipulation.

As Elkeinkrad points out, you can pay for this with Prophetic if you have it in play, which notably improves the value proposition if you weren't using that card otherwise that turn. With mystic access, one could also try fishing for desired numbers with the new Olive McBride, although generally anyone who can run Olive can run Voice of Ra to probably-better results.

Preston Fairmont's Family Inheritance has an interesting interaction here that may make genuinely gambling worthwhile, because his resources are use-it-or-lose-it each turn unless he spend an action to bank them. With two action remaining, Preston could play 21 or Bust with his inheritance money, and then spend an action to bank his winnings if he wins big and move on with a shrug if he busts.

Finally, this card's fortune and gambit traits have some interesting potential with the premier green tutor, Friends in Low Places. Selecting gambit lets one search up a number of decent resource economy options (Bank Job, "Watch this!", Kicking the Hornet's Nest), some bonus actions (Honed Instinct, Swift Reflexes), and a handful of other utility options (Copycat, Stir the Pot, Money Talks and deck thinners (Daring Maneuver). There are also some powerful cross-class options like Act of Desperation, Transmogrify, and Practice Makes Perfect.

Fortune probably doesnt have enough hits in mono-green to be worth selecting although it has the very powerful All In and Hot Streak, but for green/red players it opens up a huge library of failure-prevention and failure-exploiting cards in Survivor.