Welcome to Season 3!

Welcome to Season 3!
0

#1

Season 3 Rules
During Season 3, we are going to be trying out some changes to the standard rules of terminal to keep the game fresh. A number of changes have been made which we believe improves the game and / or allows users to try out new strategies.

We will be listening to feedback and making adjustments, especially in the first month or two of the season. The rules and docs will be updated to reflect these changes shortly.

Unit Speed: All unit speed and damage has been doubled. This is a quality of life improvement that reduces the granularity of damage, but halves the number of frames in a game. This granularity change is a nerf to units with high damage, especially the scrambler, so we will be monitoring that.

Scramblers: Scramblers have traditionally been difficult to use, except at the highest levels of play. Scrambler range has been increased from 3 —> 4 to make it easier to line them up with your enemies attack, and to compensate for the increase in ‘wasted’ damage that will be caused by the above change. Scramblers also move half as fast as before, making them more likely to have home field advantage and block enemy units.

EMP: These had the opposite problem of scramblers: They were very consistent across all levels of play. Their damage has been increased but their range has been decreased by one. This has two major effects:

  • The number of ‘safe’ tiles they get when walking directly into enemy defences has been reduced from 2 —> 1, making them less able to destroy them before reaching them (and earning more safe spaces by doing so). This will prevent a ball of 5-6 EMPs from destroying their opponents entire base
  • Reduces the number of tiles an EMP line can affect.

Destructors: Destructors damage and cost have been doubled. This change was made to increase the space efficiency of destructors, and to make it more difficult to ‘cover’ your entire area

Encryptors: Shielding has been reduced from 10 —> 3, but shield decay has been removed. The primary reason for making this change was to create variety in gameplay. We have a few ideas about how these changes might be utilized, but are excited to see what you all come up with.

General: Refund amount changed from 50% —> 75%. Refunds were used very rarely, and we are interested in the more dynamic strategies that may appear if refunding is more viable.

Resources: Starting Cores increased from 25 —> 40, Cores / turn increased from 4 —> 5, bit decay decreased from 33% to 25%. We tweaked resources in an attempt to change strategies.

Additional updates
In addition to a number of quality of life and minor bug fixes, the ability to download C1 algos off the playground will be enabled later today.

Congratulations to the S2 Champions!
The two groups for the S2 finale will be as follows. We want to showcase everyone’s best work, so message me which of your active algos you want to enter if you are in this list, otherwise, your highest Rating algo will be used.

Round Robin Group 1.
Ryan Draves
Paprikadobi
LuckyR
ParadoxRift
Scott-Hanpa

Round Robin Group 2
TEAM13_KEAN
IWannaWin
Quincyc3
Arnby
TambourineMan42

For more details on the S2 Grand Finals, view the below post.


List of all updates (Depricated, Filter posts by update catagory)
#2

Cannon builders, let me se your hands :open_hands:


#3

:raising_hand_man:

and good luck to everybody in the S2 finale!


#4

Very, very excited about these changes! Looking forward to some new and interesting game play designs and congrats to all the contenders :).


#5

Oh dear, I guess we will all have to spend some time adjusting our simulators… I like the changes though

Also congrats to all contenders in the finals, especially @Ryan_Draves, who held himself in the top 2 all season, I think a deserved first


#6

Thanks. It probably wasn’t in my best interest to stick around there all season (I think I won’t make it out of round robin regardless of what algo I pick), but I’m glad X2 was able to hold on to some elo.

I think these changes, especially the ones coming over from Terminal X, will introduce a lot of corner play. Can’t wait to see what utterly wrecks the Demux structure.

But on the plus side… guess who already has a well established strategy that uses refunds…


#7

lol guess who never used it and has no idea about how to integrate it :laughing:

But very excited about those changes!


#8

in some testing… scrambler escorts don’t work… 1/2 pace scrambler corner suicide cannons don’t work… in general defenses looks a lot stronger?

The 25% bit decay, and 75% core refund, this seems to point to massive 3-4 turn temporary layout attacks. I just can’t see 3 shield encryptors being worth it, but with no decay perhaps could lead to some row zero encyptor batteries, still not sure 3 shields is enough.

ryand temporary defense changes were very clever, was limited to simulating one filter at the front line by my apx ~100 path simulation speed Next season want to start using reinforcement learnign and building a database useful layout options in the almost infinite state space of the game.

this game is actually similar to my day job of high frequency trading, for example coming up with scoring formulas and constant improvement and adapting to what others are doing. Def makes sense the 2 sigmas (halite) and citadels of the world are interested in games like this.


#9

Interesting changes. The EMP range change is absolutely game changing. Im kinda worried about the encryptor change and defense balance though. I feel like defense is so strong that people will just throw a lock on the front line, maybe some scramblers, and then stack encryptors in the back for days until they break through.

The encryptors can build all the way in the back so no risk of losing them. They might as well have 1 life. We can no longer use an escort strategy so attack options are limited. And because we get so many cores from turn one we can put up a strong defense from turn one…

The scramblers move slower, so they can no longer be used offensively, taking double damage. Before they where better than pings for scoring. Their use might be less than before. Or at least less versatile. With scramblers so strong defensively I feel like we dont need 40 cores to have a guaranteed defense turn 1-2.

I guess we will have to see how things play out though. Changes are always interesting.

[Edit] I think increasing the range of the scrambler, but keeping the speed the same of EMP makes it a stronger unit then when it is slower, and will make the attack options more interesting. It would allow for escort strategies that have to cope with weakened EMPs so the mix will be more intricate, and Scramblers can still be used as scoring/blow up units. What do you guys think? From all the changes made I disgree with this one the most. As for the others they will sprout new strategies. And with the recycling efficiency more dynamic algos.


#10

I agree with you, the slower scrambler makes 2 of its most common uses today not possible:
EMP escort, safer replacement for ping rush scoring.
But make it more usable as a defence tool (especially with the range buff) and it becomes “moving tower”,
which I guess was its original design, and is hardly used so far.
I think the game will be more interesting this way.
And to compensate for the lack of escort, or stronger rush then PINGs, we have Encyrptors with no decay …

And no, Encryptors are not so OP, a good maze , to strip your cores, or a good cannon, with well timed unstoppable attacks … will not let you “safely” build Encryptors in the backline


#11

I agree partially with ping cannon. Even though the stronger destructors will allow you to cramp more firepower in the corner. The maze however… The decrease in range for EMPs is a gigantic nerf for that strategy.

Actually interestingly with these changes, Demux no longer works so everybody can go to the drawing board again. I’ll see whether I can take advantage of that end of august :smiley:


#12

I mean no disrespect, but it seems to me people don’t really understand the mechanics of good cannons.
There is limited space for the enemy to build a proper defence,
Most of the “projectiles” path, is in friendly territory, thus controlled by the cannon builder
Over investing in defence, makes a good target for EMPs, and suicide bombs, aiming for Core damage.
Over investing in corner defence, makes you vulnerable … in the other corner and mid field … and you need way more resources to properly defend a corner, then the mid field.

All this does not mean Cannons are OP, but it is hard mind game then is a bit simpler for the cannon builder


#13

I know, ping cannons have been the very first strategy that was developed. I have been playing this game since September last year. Though interestingly this strategy was absent in the season one finals, and then popped back up in the season two finals. This has something to do with the meta that is going on. If ping cannons are the prevalent meta, then the ping cannon counter becomes extremely effective (Ping counters usually are quite committal, but ping cannons themselves are also quite committal so it checks out). The point of using EMP’s to chip away at the corner to counter ping cannons like that is a piece of logic that exists in my original Demux code. Having a structure that can both use EMP and ping cannon at the same time is a very specific one that also has counters. I personally think that the increased firepower of destructors relieves some of the pressure from the corner, making it easier to survive for more turns. After turn 40 or so, ping cannons can no longer be stopped statically.

actually the original ping cannon required scrambler bombs to smash open closed off corners. But right now those things move twice as slow, with destructors with twice the fire power. You are going to require quite a few resources to smash through the closed off space.

The meta has been shifted between: “Who can ping cannon better than the other ping cannon” to “who can build the best maze” to “who can make the best Demux” to ping cannons again (?), and now… we’ll see :slight_smile:. Though I am pretty sure with the new balance patch ping cannons as at least a substrategy will be very important. Especially late game.


#14

I’d consider myself a bit of an expert on ping cannons and I see a few nerfs to ping cannons in the new update.

  1. the slow down of scramblers coupled with the increased damage from destructos makes a scrambler bomb infeasible unless the corner is only defended by filters
  2. since pings/scramblers travel such a short distance within the cannon, ecryptors are less effective. This combined with strong destructors mean less pings will score on each run. This will be offset by the decrease in bit decay but I think it will still be a net decrease in scores/turn.
  3. the cannon itself is fairly core intensive. With the increased cost of destructors it will be harder to defend your other edge. Again this will be offset by the increased cores and I haven’t crunched the numbers but my gut tells me the increased cost out weighs the added cores.

I think the changes will lead to more diversity in strategies among the top players and I’m really looking forward to seeing what people come up with.


#15

After a little bit of playing around, I definitely agree with the points @quincyc3 brought up. I think the “traditional” ping cannon isn’t much of an option alone, but I don’t think the strategy itself is suddenly doomed. Rather, I think the changes most affect strategies that depend heavily on the ping cannon alone. The increased refund cost makes it more viable to switch to a ping cannon and then have some other base. As we saw in the finals of season 2, this was already quite effective given certain layouts, but I think this will be accentuated given the increased damage from pings.

Edit: When I say damage I mean damage done to static units (firewalls), not the enemies health.


#16

Whoah, wait did they change that? I missed something. Where do they talk about new damage? That might change things

[edit] whoah the EMP does more damage, and the scrambler does twice the damage too. damn. It does more damage than the new destructor which is already twice more than what it was previously.

[edit2]: pings do the same amount of damage right? Previously it was speed 2 and 1 damage, and now it is speed 1 and 2 damage. But that is the same amount. It is the EMP who got a buff. These new changes are even more interesting than I thought… The meta will change a heck of a lot


#17

I’d look at this thread, and the stats have been updated on the rules page.


#18

Overall yes, the damage should technically be the same. However, I don’t think it will be in actuality for certain cases like a ping cannon. Specifically, I’m thinking of when a unit gets in range and the “extra” damage that is done to a unit that doesn’t carry over to other units.

For example, if some pings are in range of 1 destructor and there is a second destructor in range next frame, I believe the damage would be different because the pings would get in range of the second destructor sooner, although maybe it cancels out since destructors got doubled as well, haven’t thought any of this through carefully.


#19

I think there will be a large change in meta for sure, and possibly some initial variety, but I wouldn’t be surprised if it solidified into a general static algo that makes pretty good use of the changes. I think the range reductions and speed changes take a lot out of existing strategies, and if I’m not mistaken 2 filters in row 13 and 1 destructor in row 12 should cover the corner sufficiently, although that is 16 cores to cover the corners.

My best guess is a modified defensive Demux that covers its corners, sends out scramblers while it builds out encryptors in the back, then cleans house with encrypted EMPs.

Edit: @kkroep’s below: I’m guessing if you’re using defensive scramblers, you can afford to make channels near the front that sweep out any early attacks. The only risk is letting EMPs strafe that line while the scramblers don’t hit the opposing line. EMPs can’t hit scramblers without getting hit back now, which makes matching them on oldschool “emp lines” much more efficient. And now the scramblers are slow enough to catch just about anything with only 1 line going across.


#20

I think @Ryan_Draves algorithm is going to suffer heavilly from this slower scrambler thing. He had this algorithm that could catch large EMP attacks perfectly. But that is no longer possible if the scramblers are too slow. An attack at the sides can be blocked but an attack through the middle will always be faster than the scrambler. Only scoring can be prevented. This changes the balance quite a bit. @Ryan_Draves do you have thoughts on this?

[edit] ryan above, the problem is scramblers work twice as slow. So if you start out of range you can actually do the middle part as a front line EMPline with no risk for scramblers. Because they theoretically cannot get there in time.

Demonstrated in the picture is when the scrambler will catch the EMP if done perfectly. For pings this is even further back, so it essentially turns scramblers into goal keepers. That is actually still useful, but also quite specific. So it is much easier to use them as goal keepers, but all other uses for scramblers are no longer working. I think that is kind off a shame, because I really liked this unit as a swiss army knife with many interesting and specific uses.