The Tao of Gaming

Boardgames and lesser pursuits

Results from my Slay-the-Spire/Bridge Training

As you recall (or can read), I was using techniques given by Kim Frazer in her Gaining the Mental Edge at Bridge in Slay the Spire. My goal was to win 25% of the games I played, and the results are in: I played the same # of games as my sample size (where I won 12%) and got

17W – 33L (34%)11W – 39L (22%)6W – 44L (12%)34W – 116L (22.67%)
Not quite…

So, I missed, but not by much. I won more games with each character from the fifty prior games, nearly doubling my win rate. (I’m still much better at Ironclad than the other two characters, but that’s because its easier to play).

So, did it work? Well, it’s complicated.

(There’s an XKCD for everything)

This shouldn’t be taken as a complaint about the book. The real stumbling block is that it is incredibly difficult to categorize strategic mistakes in Slay the Spire. Bridge is easy by comparison. While it has grey areas, many of the typical mistakes are easy to diagnose by simply replaying the hand. Bidding has borderline cases, but the play and defense can be fairly rigorously analyzed after the hand.

In this respect, bridge is like shooting. You get instant feedback if a shot was good or not. In this comparison, Slay the Spire is … not quite Calvinball, but at least Cricket as understood by Americans. So, to take some notes from a random loss. “I died hitting the worst possible elite at the time, then drawing poorly. Despite that I might have won if I’d not used my potion a turn too early in this case, if I’d saved it for a turn later — with the draw I got then — I would have won”. So, clearly some bad luck, but also a micro-mistake (any mistake inside a single fight I call a “micro” or “tactical” error). But if I hadn’t used my potion and didn’t draw that particular card (about a 50/50) I’d lose in all cases. So, should I have used the potion? Maybe? I could math it out, but that’s just inside one fight. Many of the StS issues are “Should I rest or smith” and you don’t get feedback (dying) until five floors later, but smithng did save some health, and you had a few random events. Feedback is incredibly noisy.

There were some losses that clearly had horrible luck. But how much? Difficult to say. After about 1/3rd of the trial I realized that my guesses as to why I lost were pretty random. Even right after the game I sometimes couldn’t tell. I suspect that (in the future) it might be best to track more specific information.

Another reason for caution is that this last month was fruitful one for my outside learning. In particular, three StS streamers talking shop about their respective recent win streaks for 3 hours was an invaluable resource, and I probably got a few extra victories after watching that (and reading Jorbs debrief after his Slay the Spire Marathon).

But certain aspects did help:

  1. Mindfulness. My checklist wasn’t perfect, but it did catch some common errors I made. I might revise it.
  2. The act of reviewing the notes. I haven’t done a detailed review of the most recent set of 150 games, but I suspect there is data to be mined. (Some StS streamers appear to have all their runs in a DB where they can run queries to answer it. I haven’t gone nearly that far).
  3. Instead of trying to quantify why I lost, I switched at some point to just writing down a one sentence summary. That may help in clarifying thoughts.

Anyway, with all that said, I suspect that the techniques from this book will work quite well to help the intermediate (or better) bridge player, or really any game where you can quantify the mistakes easily. (Perhaps some StS players who are better than me can, so this would help them more than help me).

So — What’s next? Clearly more Slay the Spire … (its pretty much my pandemic relaxation). I’ll try for a 30% winrate for my next 150 games and we’ll see if maybe my last set was just regression to the mean (as I think 12% was low).

Written by taogaming

February 27, 2021 at 2:46 pm

10 Responses

Subscribe to comments with RSS.

  1. What level are you playing at?

    Also, you have found the Spirelogs website, right? I found that improved my game a lot.

    Frederic Bush

    February 27, 2021 at 9:53 pm

    • A15. I’ve gotten to A20 (and beaten it with all but watcher), but I enjoy having wins be roughly double digits, not 1%. I haven’t used spirelogs.


      February 27, 2021 at 11:11 pm

      • When I was playing a lot, I used the spirelogs stats to guide my choices. I didn’t actually upload my own games.

        Frederic Bush

        March 3, 2021 at 4:04 pm

  2. I enjoy StS, but not enough to research how to improve my play. But since you’ve done that research, any broad takeaways to pass on? I’m around A7 or 8 with each class.

    Peter Sarrett

    March 1, 2021 at 2:35 am

    • 1) Be aware of the need to balance fast damage (“How much damage could my deck do if I played each card once the first time through the deck”), scaling damage (“how can I get up to 400 damage as fast as possible), block (fast and scaling), and sustain (healing). Sometimes “How to get to my best cards faster?” (if you have a key card). Once you have an area “good enough” focus on the next weakness. (Fast Damage first, then fast block, then scaling. Sustain depends on how you are doing). The key struggle is to define “enough” and “when to switch”. (It’s like the Talmud. “Be Good” and then 10 books of commentary).
      2) For some fights (elites/bosses) holding one key potion is often enough. One relic that synergizes likely beats 3-4 random relics, so saving cash for a store is good.
      3) Great things — Dead Branch, Intangible, Deck thinning, Blade Dance, any card that “doubles” is a great scaling card (Catylst, Echo Form, LImit Break, Burst). Snecko Eye and Runic Pyramid (not together).
      4) Deck archetypes.
      Ironclad — a)Corruption/dead branch, b) Strength Scaling, take the hit, reaper to recover (feed to let you take the first hits). c) exhaust synergies (often with A, but also to make a small “infinite” combo.
      Silent — a) Blade Dance Cheese. b) Poison cheese. Add intangible cheese to either, but sometimes if you get enough intangible you can win by just stabbing things.
      Defect — If you can get setup, you can beat anything. How are you going to get setup? How will you heal the damage you took on T1-4? a) Typically Orb Slots + Frost + defrag b) lots of lightning. c) COre Surge + Biased Cognition.


      March 2, 2021 at 3:04 pm

      • Above list is probably not nearly complete, just random quick thoughts. I may write about this at some point.


        March 2, 2021 at 3:04 pm

      • You should also look at the boss and have use it to guide your deck choices. Slime boss needs fast damage, Time Eater rewards scaling and ruins decks that play a lot of cards.

        Frederic Bush

        March 2, 2021 at 10:31 pm

      • So are you trashing cards in the shop at every opportunity, or saving that case for relics? Wondering if I’m undervaluing trashing base cards.

        Peter Sarrett

        March 3, 2021 at 5:14 pm

      • Trashing base cards is a strong option (see also, Dominion, Chapel). In general when hit a shop if something is good (and trashing often is) I don’t often save. (The big exception is that if I have a Maw Bank not buying somethign will let me buy something big in the next shop, if the current buy is borderline).

        Looking at my last few runs that made it to Act IV, the number of strike/defends I had removed are: 3, 7, 7, all (Pandora), all (pandora), 4, 6, 6 …. not all of those were removed by shops, but it seems like I’m removing about half of my starting cards in general. I think I will go ahead and write a larger article….


        March 5, 2021 at 11:06 am

      • Base card trashing depends a bit on the character. I do it slightly less often with the Silent, relying more on card cycling since it starts with 2 extra base cards. I’ll certainly do it when given the option, but I’d rather spend my money on other things at the shop.

        Mark Delano

        March 5, 2021 at 11:48 am

Comments are closed.

%d bloggers like this: