Counting cascades

Volrak

The rant thread has gotten all statisticy, and that's not ranty at all.  So here's a separate thread about collecting data for counting cascades.

If you're interested in contributing data to help work out whether the AI or the player gets more cascades, I've created a new shared spreadsheet, inspired by Quantius, where anyone can enter their cascade data as they play a game (specifically, number of gem matches per turn): https://docs.google.com/spreadsheets/d/1K4ITAgZXf_EzXY8NVAJjYuYvhEAGPmR4sdyrzqE9_e8

The only "pure" data is from turns in a game in which no gem changing effects have yet been triggered, so gem-changers need to be marked on turns they influence.  The sheet is also collecting player name (because I expect some variance depending on who is doing the matches) and game type (since whether you're casually messing around in TG, or focusing for a coalition event, your approach could also make a big difference).

Quantius

Figured I'd post my spreadsheet here too:

https://docs.google.com/spreadsheets/d/1Wx5umd9eZCXBYy1brd40UvuuPnYmv7sjsq1YZY3PU5s/edit#gid=0

I'll also note that I'm not running things like Hour of Promise in my tests (I am using other support style changers which I think reflects in certain games more than others). So far I've surpassed the AI once and tied once and the AI has surpassed me 10 times for a total of 12 games as of this post.

The biggest AI cascade has been 13 matches in a game I was otherwise ahead in terms of matches. I've lost zero of the test matches (but the heroic gauntlet is pretty wimpy).

I'll see about adding entries into your sheet from now on.

Now I'm wondering if my gem matching style is resulting in me (and maybe others who feel they suffer from AI cascades) allowing the AI to find more matches. I generally look for the highest mana match I can with swaps taking priority. If I see an off color cascade option, I'll only match it if I will net more mana than an on color match. Cascading for a total of 2-4 mana hardly seems worth it to me, but maybe it is in terms of hate-matching against the AI. I dunno. This might be exacerbated during events in which the AI have their mana gains tuned higher so I'm giving away bigger gains because it's not a valuable match to me.

Volrak

Quantius said:

Now I'm wondering if my gem matching style is resulting in me (and maybe others who feel they suffer from AI cascades) allowing the AI to find more matches. I generally look for the highest mana match I can with swaps taking priority. If I see an off color cascade option, I'll only match it if I will net more mana than an on color match. Cascading for a total of 2-4 mana hardly seems worth it to me, but maybe it is in terms of hate-matching against the AI. I dunno. This might be exacerbated during events in which the AI have their mana gains tuned higher so I'm giving away bigger gains because it's not a valuable match to me.

Yeah, all of those things sound like possible factors.  I think another factor is the PW you're using.  If it's a walker with balanced mana gains across all colours, then it's often better to pick for cascades over on-colour matches.  E.g. for Kiora or Dovin, 2x off-colour matches for 6 mana + 2 loyalty (plus higher chance of triggering off-screen cascade with 6 gems destroyed) is better than 1 on-colour match for 6 mana + 1 loyalty.  Whereas with someone like Koth, you'd pretty much always go on-colour if you can; a side effect of that would be to leave more cascades for the opponent.

Quantius

Did my TG this morning, tried it two ways. With Bolas I played as I normally do, looking for best mana gains. It's usually short games with my ID deck, but game 2 the AI came hard out the gate and happened to be a Lili PW so I kept losing my combo pieces until I managed to have enough cards in hand to protect it.

Then I played my next 2 games with Angrath, a slower control deck. I aimed to prioritize cascades of any kind over on-color matches. First game with Angrath I came out hard with 4 cascades giving me a 15/15 right off the bat. So the game took about half as long as normal. I would have stayed ahead of the AI in terms of matches until it took out the board and cast double deploy. Luckily I just needed 1 more turn to win because I would have been dead the following turn.

Second game with Angrath was more it's normal speed, and I eeked out ahead with matches.

So far what I'm noticing is that on-color matches dry up quickly without gem changers and you're left matching fresh drops in the top row where it's really unlikely to get cascades since there's nothing to cascade into. This has been my normal matching style so I'm usually just getting 1 on-color match. Multi-matches tend to be deep on the board and while it's RNG, it can simply land in your favor leading to more cascades as the many empty spaces are filled with new gems.

Second big takeaway is the getting the fresh matches means often leaving new cascade/swap triggers for the AI to snatch, so I may be causing the AI to find more cascades due to how I was doing things. I'm the culprit and the victim in one!

Unknown

This content has been removed.