This wiki is a read-only version of the Stardew Valley Wiki. The official editable wiki maintained by ConcernedApe can be found at stardewvalleywiki.com
Talk:CalicoJack
| This talk page is for discussing CalicoJack. | |
|---|---|
|
|
CalicoJack probability simulator
The decompilation on GitHub by the user veywrn at the repository StardewValley shows that draws are not weighted uniformly for either the player or the dealer as of 1.4. Player draws are more likely to result in a 20 or 21 when the player hits with a 16 or higher. Dealer draws are more likely to bust when the player has an 18 or higher, and this can result in the dealer receiving cards that would otherwise be impossible. Based on these probabilities, I wrote a small program to calculate the odds of winning if you hit vs stand for any given player total and dealer. You can find the source code and its output at (apparently I can't include a link here, it's a gist on github. User sgrif, gist id 0f319e5ccedbf2fb2231bf1400cc323f). This program demonstrates that the increased chance of the dealer busting, and the decreased chance of the player busting means that you should always hit on a 17. The program considers a set of odds "better" if the chance of loss is lower. This means that it would prefer a higher draw chance to a higher win chance, as long as this decreases the overall chance of loss. The only time this is relevant is if you have an 18 and the dealer is showing a 9. If you hit, your chance of losing increases from 46.6% to 47.7%, but your chance to win goes from 34% to 47.3%. I'd argue that a 13% increased chance to win instead of draw is worth the 1% increased chance of loss. This means the optimal strategy is to hit until you have 18 or higher, unless the dealer shows a 9 in which case you hit until you have 19 or higher.
OLD PROBABILITY SIMULATION BELOW:
I've spent a few hours figuring out whether this game is worth playing or not, from the point of view of trying to win (of course the fact that it's fun counts for something too). The conclusion is it isn't. I've drawn this conclusion based on writing a script that plays the game according to a certain strategy, and simulates 100000 of these games and reports the outcomes. If someone who is good at calculating probability wants to verify these outcomes 'the math way', that would be awesome.
The short explanation: The only decision you can make is clicking hit or stay.
- If your current score is 12 or lower, you always choose hit, because no card in the game can kill you at this point, and higher is better.
- If your current score is 21, you always stay, because every card in the game now kills you.
- What's left is to make decisions on scores between 13 and 20. I've simulated games with different 'stopping points', meaning the number of points at which you always stay. E.g. if your staying point is 16, you will click hit if your score is under 16, and stay if your score is 16 or higher. Then either you die or the dealer plays, and we see the outcome.
Assumptions made (if these turn out to be incorrect, please let me know):
- The game doesn't use a pack of cards (in which cards that have been dealt can not be dealt again in the same hand), but a random number generator, making all dealt cards equally likely, regardless of what has already been dealt.
- I am not taking into account decision making based on the one dealer card that you see.
This is the result of 100000 games each played with 'stopping points' between 13 and 21:
| Tested stopping at 13. Wins: 36287, Losses: 58312, Draws: 5401 |
| Tested stopping at 14. Wins: 37360, Losses: 55706, Draws: 6934 |
| Tested stopping at 15. Wins: 37879, Losses: 53479, Draws: 8642 |
| Tested stopping at 16. Wins: 37999, Losses: 51855, Draws: 10146 |
| Tested stopping at 17. Wins: 36939, Losses: 51125, Draws: 11936 |
| Tested stopping at 18. Wins: 35491, Losses: 50844, Draws: 13665 |
| Tested stopping at 19. Wins: 33311, Losses: 55543, Draws: 11146 |
| Tested stopping at 20. Wins: 26960, Losses: 65117, Draws: 7923 |
| Tested stopping at 21. Wins: 16290, Losses: 79505, Draws: 4205 |
As you can see, the chance of winning is always smaller than the chance of losing.
-- Alvilda (talk) 11:41, 3 December 2016
- As of v1.4, odds of winning have changed and made these stats obsolete. Butterbur (talk) 02:41, 2 December 2019 (UTC)
Easy Qi Coins!!!
It seems that the outcomes for each round are locked in at the start of each day, with some variation based on when you opt to stay. I was able to get over 1,000,000 Qi coins starting from the initial 100 purchased coins using the method outlined below.
- Start on a good luck day, adding food buffs if desired (this seems to help increase win outcomes)
- Go to the casino and buy some Qi coins, use these to play a set number of low stake hands (I did about 35-40)
- Make note of your card counts and whether each hand wins or loses
- For the first run, you should stay on the initial draws (I found that the dealer often busts themselves even when they have a high enough card count)
- Reload your save and repeat the process, this time going for 21 or bust each hand
- Reload and repeat once more, playing the winning hands from the previous runs and highest before bust on the losing ones
- Combine the wins from all 3 runs and make special note of where you have more than 1 win in a row
- For the final run, you want to chain together as many wins as possible while betting double or nothing
- Chain wins on the low stakes table until you have enough coins for the high stakes table
- Wherever you have a losing hand, switch back to the low stakes table to play it
- Keep playing high stakes on winning hands and doubling on win chains until you have collected enough coins
I got really lucky and had several good win streaks with only 2-3 losses total, but the results will obviously vary. Kaori kins (talk) 23:49, 9 August 2020 (UTC)
