Monday, 23 June 2025

22/6/2025: Cell games

I was checking my old phone earlier today for any remaining files before hard resetting it as family backup phone.

I checked all the pictures and downloads, whatsapp backups, as well as all the abandoned apps. I then  came across to a particular app: the tentacle wars. That is the ported version of the old kongregate game that I spent so much time on. I played the custom levels on toilet, before bed or simply all the time back in the days. That's the screenshot I took today:


That means 3349 total (auto generated) random levels solved. Taking levels deemed too easy to spend time on into consideration the actual number is easily above 5000, and that's only on this particular phone. On previous phones the counts are all 5 digits. 

The total score is the sum of scores across all non-random levels, and the score is essentially a measure of remaining time -- or equivalently the time you used. The higher the score, the lower overall time you used to clear all stages in the game.

That reminds me of the various speedrun categories where players care about time spent on independent stages separately. It might sound less exciting than non-stop categories but they also provide all sorts of fun plus extra technicalities since you can study every detail to the finest bit. Examples include punch-out, Mario Kart 64 and Smash Melee break the targets -- when you know these are all games mentioned by Summoning Salt you know these categories must be spicy!

Fortunate for me, a framerule (cue the bus analogy here) system applies here where the clock/score ticks only every second, so optimization is much easier. I once thought 110k+ was impossible but then proceeded to reach that, and I am quite confident that this is very very close to perfect now.

I found myself so addicted to such simple game. Or rather, I found myself so addicted to a whole genre of games. Referred to "cell games" on Kongregate, it's a branch of RTS that is almost extinct since the discontinuation of flash.

Simply speaking it's a type of RTS around cells which are preset immovable architecture, and the goal is to conquer all cells on the map. Assault is usually done by connecting two cells where units are sent from one to conquer the other or boosting self's architecture. Health of the cell is usually the only measurement -- it represents both health (it flips when HP reaches zero) and strength (how many units you can send/how powerful your attacks are) at the same time.

Most cell games separate from each other by permutation of the above rules, but that leads to completely different strategy. Just to give a few examples:

Tentacle wars
Architecture: Power = HP, upgrade/downgrade automatically with HP
Attacking means: by extending straight line tentacles (that costs HP) that attacks, power depends on attacker level, no collision
Strategy: fully connected network, overloading (tentacles output much more if it has grown to size limit and consecutively receives from other tentacles)

Cloud wars
Architecture: Power = HP, no upgrade, homogeneous structures (with variations from 2nd game)
Attacking means: send half of its own HP towards others on straight line, with collision
Strategy: Aggressive, concentrated attack, make use of collision to defend

Little stars for Little Wars
Architecture: Power = HP, predetermined production power for each star
Attacking means: consume ALL of its own HP to attack on designated straight line (graph structure), with collision (only when path cross over each other which is rare)
Strategy: create chain of power transportation, create line of defense/min-cut analysis

Solarmax:
Architecture: Power = HP, predetermined production power for each star
Attacking means: consume part (adjustable) of its HP to attack, do not collide
Strategy: swift, concentrated attack, high precision high frequency control, game theory

Auralux
Architecture: Power = HP, spend fixed amount of HP to upgrade (for higher HP and production) for some stars, reverts to lowest level 1 if conquered
Attacking means: consume part (adjustable) of its HP to attack, do collide
Strategy: swift, concentrated attack, rely on chaotic randomness on multiplayer field\

I really how such simple rules lead to different optimization approaches. They are complex, but not complicated enough so that near full optimization is possible. It also matches my interest on investigating the "ultimate meta" of a game. 

Sadly the tentacle wars app is no longer compatible on my new phones and with the upcoming hard reset it would be gone forever on my phone. But for whoever interested in the game you can still find that in Kong (using ruffle). Alternatively, Little Stars for Little Wars and Auralux (2) are both on app store that you can give it a try.

Tuesday, 10 June 2025

Thoughts on CTWC2025

Hello and welcome to my CTWC thoughts 2025 edition.

I finally watched it live again. That's probably due to my huge interest on TGM4 listening to its OST and watching every pushing attempt, but that's topic for another day. Youtube on the other side, sensed my taste and started pushing everything about CTWC to me, from player interviews to day 1-2 qualis and of course, the finals on Sunday.

There were about 2k livestream watchers early in the finals. That number is surely much lower than the numbers we saw during 20-22 twitch live streaming, but that isn't a bad number either. It is hard to say whether it is better or worse comparing to 2024, this is something we can say much much later.

This year however, I would not cover anything further about change in popularity. I felt bad about missing technical contents in 2024. Sure I missed the tournament in live and surely wouldn't spend more than a few hours to re-watch the whole thing, but it is bad in the sense that the missing technical content created a hole, something hard to recover when I come back trying to figure out what was classic Tetris in 2024. Think about it -- I mentioned "lv29" ZERO times in my 2024 thoughts. How was that even possible?!?

I believe my argument on the tournament and my vision over the "endgame" is pretty well laid in the past years, so there is no need to emphasize that every year. Is the lv39 super kill screen killing all the fun? I might say yes in 2024 but this year I started to appreciate that -- we will talk more about that later.

As a compensation I want to simply focus on the players' performance and the shift in the meta. Let's start, shall we?

Gameplay before lv29

The community is getting so good that lv19 speed does not pose too much additional threat comparing to lv18 speed, so we may as well talk about both of them in one go.

Although I didn't say anything about gameplay last year, the progression from 2023 or since rolling is clear: optimize up to lv28 as much as possible. Given their ability to roll, it is possible to take near maximum risk and survive consistently. 

This was a goal, but this is precisely what top players are doing in 2025.

The best example is Tristop with how he maintains crazy (80%+) tetris rate in the majority of games. He even entered second transition once at 1.3M+ score and another time at near 1.3M, both would be tournament record breaking. It is unfortunate that we don't see him progressing further, because his ability of taking huge lead would be such a big threat aginst those who survives up to high 30s to hope opponents to top out.

Second transition and so on

I think this is what truly separates very good players (say, the top 32) and champion contenders. Assume that everyone survives up to lv28 what would you do?

They would have taken the risk if they can afford to, but the speed is simply too much for most players. What you see in the tournament is basically what they think is the best balance between risk and reward for players' self. If their strategy is not aggressive enough they can only hope that opponent would unfortunately top out -- if your opponent out-maneuvered you from lv29, chances are there he holds the lead before lv29 as well.

And for those title contenders, what's fascinating is that each of them take enough risk so that they probably top out before lv39 aka the second kill screen. If you think carefully, reaching lv39 is actually a waste because that means there is space for extra aggressiveness for potential extra tetrises that are worth more than potential score you can get up to lv39. Very few games actually ends with double second kill screen. I shall appreciate that as medal of players seriously taking this game and format to a very competitive level.

In short, a player's potent under such format is decided by their effectiveness in the last 10 levels. If you read the games in that way, you won't be surprised that Alex T won the tournament unbeaten at all. He is the only guy who can not only comfortably survive, but also consistently turning mess into tetris ready shapes. Again I have to mention Tristop here, he is the only one who can outperform Alex before lv29 and then try to survive Alex T's counterattack. (What about Meme? I think he is very well-rounded, unfortunately weaker than Alex on every edge, putting him in a very bad position in case of 1v1 matchup...)

Qualification and seeding

Remember how seeding works? It is all about getting maximum number of maxouts within a time limit. For some time I do not treat an accurate way to assign seeding, possibly due to impression from distant past where all you need is to get a few maxouts and chill.

Not anymore now.

Getting a few maxouts merely gets you into the gold bracket...before getting kicked out by top rankers even if you survived the top 48 stage. You need to press really hard to score 12 maxouts which all top 16 did. The aggressiveness is precisely what is needed in the knockout stage. 

I wasn't sure about the correlation between aggressiveness before and after lv29. Like...is it possible for a player to be very skillful that mastered lv19 speed but simply can't react after lv29? The tournament proves that these two are properly correlated i.e. it also properly seeds the player in the knockout stage.

At a point I wasn't sure about whether kicker makes sense. Is it better just to instakill a game to get a super small kicker (so that it isn't zero) than to lose intentionally at high pre-maxout score? Well I was certainly overthinking there. Most top players ended with a high kicker, not intentionally but the aggressiveness level naturally leave them a kicker like that. What about players with zero/low kicker? Well, that means they get an extra maxout instead of every single possible high kicker, which is, at the end of the day, more important.

The same qualification format has been used for a very long time, but I am glad to verify that it is actually a good one.

***

Before closing down my thoughts I just want to raise one last observation: lv29 gameplay is extremely stamina and concentration consuming. 

This is quite easily overlooked when you thought players could play casually up to a few million points (in the unlimited format), even up to the glitched levels or even close to reborn. Not only that such high level gameplay only applies to a few players, it is also about tension and pressure from the tournament. Scuti is one example where he slipped and misplaced bad piece on column 2 forcing him into rescue mode. If that happens on a player who reaches reborn, the same could have happened to anyone else.

When it comes to extreme e-sport, I always feel stamina being a very important factor. RTS games, MUGs, and of course competitive tetris. Perhaps a secret of Alex T's domination is his training from other sports...?

That's all! Another year, another fun weekend to enjoy. I will see you again in 2026.