Currently set to 29th at midnight, arbitrarily
Okay, I noticed that matchmaking was really fast around 1 hour ago but now it’s gotten noticeably slower. I suspect that it’s due to more people uploading algos now that the deadline is extended. Is it possible for the matchmaking to not be as slow again?
The original deadline was midnight on the 27th at midnight ET, and was extended to tonight at midnight ET after concerns were raised in this thread. I don’t plan to extend it again, as this is a season-long competition and algos ideally should be iterated on across the entire season
2 Likes
Yeah I agree that algos should ideally be iterated on across the entire season but many players have recently started uploading their algos quite late into the season (about 1 week before the deadline if I’m not mistaken). Personally, I think that this defeats the purpose of Terminal C1, not to mention that some of these people copied my team’s layout and even attack. I started noticing these ‘copies’ of my team’s algos only a few days ago so up until then everything was fine in terms of everyone iterating on their algos. Even though I started playing Terminal in Season 5, I personally think that Season 7 had the most people copying algos so far…
I agree with this (as you may have noticed I was also copying). It has always been a valid strategy though, there was even an entire thread of discussion about it:
Maybe we should rethink this in future seasons.
I tried to make my own strategy but due to fundamental flaws in my base-outline I was losing no matter what I did. Then (because I wanna win) I started to copy your superior base (which multiple people were already using)
Respect to @acshikh (and to the person who created the strategy I copied) for making such a unique strategy, I’ll try to do the same next season.
3 Likes
By copy I pressume you’re talking about the “V” (IDK if there is a better word to describe it, they are vaugly “V” shapped so “V” layout lol) shaped type layout most of your algo’s, that mine’s played against, and other people’s algos I started to see them use. My algos ended up that way starting with the first IDHC, after I tried many different types of defensive firewall layouts, I noticed the main differences between what I ended up vs what a lot of other people went with I deploy starting factories center, then it decides to go left or right each side mirroring the other, based on enemy unit health, or if enemy unit health is equal on both sides it just picks one randomly. As of now the one that’s performing the best waits 3 turns, other variants of it wait 2 turns, or start deploying firewalls on turn 1, the one that started turn 0 just didn’t work out, i pressume becuase some other algos will deploy defenses differently based on how the enemy does it as well.
I noticed some of newer versions of yours started doing that as well, where it deploys center then seems to decide whether to fully commit on the right or left side, and I’ve been working that angle since the start, (like litterally since season 2 when i first started playing lol i always try and do that now) mines been using a similair firewall layout to yours for over a month now though, since I started uploading and running games since day one, and I can’t rember when I uploaded the first version of IDHC, it was atleast a month ago.
As for attacks, mine randomly selects from a small number or possible deploy locations until it’s decided to either commit to the left or right, at that point it has code to choose type of attack depending on enemy player health, number of scouts affordable, and enemy unit health. I don’t think submissions are going to be open again this season, i suppose if they were it would be unwise to give as much information as i did. IDHC, the the different varations of enodev, and notenodev, have slightly different attack code mostly due to being close to deadline and not being able to adaquitly test each one when i made the changes.
This isn’t new though I don’t remember what season, i think i first saw something maybe it was in forums from the previous season when i started playing in season 2, so season 1. Like every top algo was an EMP maze algo. I think I saw soemone from C1 state somewhere, is due to the nature of the game, they can’t really private replays that players algos are in, since part of the point of any game is being able to see where you messed up and fix it. Granted a completly blind season, might be fun, if a little frustrating, where the only games you’d be able to watch would be from the bosses in the playground, since like i said you wouldn’t be able to see specifically why your algo is winning or losing, a specific match, so you’d have to guess i suppose, that could lead to an interesting finally live stream though.
1 Like
Let’s be honest, referring to past seasons and other people copying isn’t a good way to justify copying other people’s algos (and if you want to win for that matter). The people you mentioned who were also copying just so happens to include 2x Terminal Season Champion @Felix, much to my surprise and disappointment. He was actually the first to completely copy my team’s layout which eventually encouraged more people like you to copy my team’s algo. The only reason why my team kept quiet about this was because we didn’t want more people to copy our algo (but that obviously didn’t work out).
My team was the first to adopt this layout and it worked for the majority of the season as we held the #1 spot. Fast forward to now and you have half of the leaderboard being copies of my team’s algo (or some other very close variation of it). This not only discourages innovation from players in general but is also a big frustration point for people like myself who spent most of their time in coming up with such a layout and attack. (You can imagine how it feels to lose to a ‘copy’ of your own algo). I think players like @acshikh, @Toastyone, @HOSO and more were the only ones who came up with original algos including my team and I.
Moving forward, I would like to hear @RegularRyan’s input on how this issue will be addressed (for both this season and future seasons) because if it isn’t, it will only repeat in future seasons (as shown in the reply above).
I mean there is always another way to look at it, you have the orginal algo with the orginal layout, make an algo that beats it, you’d need to make sure it could beat other algos as well obviously, so I would say maybe test it by uploading it for a little while, or test some varient of it live, then right before the deadline. Upload it, especially if you notice a lot of very similair algos that are running the same or similair strat. or heck have several different strats that can beat the one that is being copied and upload them all right before the deadline, so everyone who took we’ll say inspiration from your algo won’t have time to really make a counter to it. You could also bake it all into one algo. If you have a base starting defense for it, then wait a couple turns, if it looks like fromt he first few moves going against the algo’s main strat, as a clone, or varient, you can have it pick a counter to it. I know people on the forums here in past seasons have talked about this idea, where waiting a couple of turns with some form scrambler stalls then making a decision based on how enemy defenses have been set up to use a completly different strat.
1 Like
I agree with @Toastyone here, copying will always be part of this game. If you want to avoid this, you could try to make algos that cannot be copied, just as @acshikh did or @arnby or @Aeldrexan in previous seasons.
I don’t think there is a way of getting people to not copy…
1 Like
I would say aside from a blind season at some point some part of any good algo will be cloned to some extent, people would probably want a few more bosses in the playground for testing though. I think a blind season would be interesting, if a little frustrating as i stated previously. Honestly though if there were to be a blind season i wouldn’t mind contributing some of the algos I have towards more bosses. Or take like the top ten algos from the previous season, or from a few different previous seasons, and making them bosses (with the makers permission of course), but then having actual matches be blind so you couldn’t watch replays.
I think at some point we should draw the line between a straight copy and just being apart of the “meta”. With factories being introduced in S7 and a nerf early on in the season, I think the majority of players have factories figured out. We see that in how players prioritize factories over other defenses etc. etc.
Is copying an algo’s layout OK? Is copying an algo’s attack strategy OK? That’s up to the players to decide. Though in my opinion it’s quite difficult to prevent others from copying you.
S7 quickly evolved towards a “meta” of the “V shape” algos at the end of the competition, rather than slowly building towards that meta throughout the season. This, coupled with the server issues and slow match making definitely frustrated some players as we did not expect the match making to be this slow.
I agree with @Toastyone that we should have more bosses in future seasons. Allowing voluntary submissions for boss algos would be nice, especially if match making becomes slow again.
Also shoutout to Max K’s mint algo for withstanding the test of time in S7 and staying in the top 10 for months without touching it.
3 Likes
Like I said, this season was the worst in terms of copying and it will only get worse if nothing is going to be done about it. I only replied to this thread because I wanted to encourage the moderators to take action.
I’ve noticed that the number of players in the newer seasons have dropped dramatically compared to the number of players in the first few seasons. I wonder why… 
If it gets to the point where everyone starts copying the best-performing algo(s) during a single season and justifying their rationale with simply adapting to the ‘meta’, the game becomes unplayable in my opinion.
Newcomers can just scout the leaderboard, see which current algos are the best and straight up copy them right before the season ends. That’s not adapting to the ‘meta’, that’s just being lazy and disrupting the ‘meta’.
I notice the dwindling participation in the global competitions as well. People copying is one thing and should be discouraged, but the prize pool is another, as mentioned on another thread. Most people are drawn into Terminal for the live events since that’s where most of the competition and prizes are. Not to mention the Citadel recruiting opportunities only seem to apply to Terminal Live events as well.
I suggest we find ways to discourage and stop copying, maybe with a blind season as @Toastyone suggested. And ways to boost participation in the global competitions.
1 Like
Personally I’m against a blind season.
It could be interesting of course but it would become near impossible to actually improve your algorithm if you don’t see why you lose games (once you beat all bosses).
Like I said before, yes copying isn’t great but there is almost nothing to be done about it. I think it’s a fundamental part of the game and it’s not like any new player can just look at the best players and copy, since for example my algos rely on an action phase simulator, or other rules which can’t just be copied.
I wonder if the lack of participation this season is what leads to so many copies in the top ten? Copying has certainly happened a bunch in past seasons but it hasn’t been quite so blatant or widespread.
In my opinion, a blind season would be bad for Terminal since it would make it very difficult for new players to start playing the game. Playing against bosses is not nearly enough feedback to make a good algorithm.
We need to accept that copying of ideas is going to happen to some extent. For anyone who is frustrated with copying, I think the best thing to do is to focus one’s time on developing the decision-making logic of their program, as that is much harder to copy and can lead to enduring performance gains.
1 Like
I know you can just go to, well these api pages
https://terminal.c1games.com/api/game/leaderboard?page=(page number)
https://terminal.c1games.com/api/game/algo/(algo id)/matches,
then the watch page.
https://terminal.c1games.com/watch/(match ID)
I don’t think a lot of new players would know about this, well unless they search the forums. I guess the dilimma comes up, blocking access to the api all together, would cause issues.
Limiting access to replays by what your algo has played in would allow you see why your algo lost and work on it, I imagine that could possibly cause other potential issues though.
A way i imagine that could work, would be assigning each player a unique api/replay access key, and that key can only see replays from algos associated with your account.
I imagine that could curb it slightly, granted lower ELO players would still be able to see some higher level algos, since all algos start at an ELO of 1500, so a top performing algo would still show up in lower ELO algos api replay access(if that algo played against it shortly after upload), at that point though it would still come down the skills of the programmer in order to implement it successfully.
Granted I have never setup an API before, I’ve interacted with APIs, so I don’t know if that would be too difficult, or complicated but new players couldn’t go right to the API and view matches of the top 10 algos, unless they were able to get high enough ELO to match against those.Or matched against them when they first started at the starting ELO of 1500, (Assuming matches are based on a range withing each Algo ELO rating)
@IWannaWin That makes sense, i think that’s a reason @RegularRyan and other c1 staff have brought up when the issue of copying has come up in the past. Your algo is more than just the defensive layout, it’s the attack logic, the logic behind how you deploy your defenses etc. I remeber one season ping cannons were extremly popular, and on the surface they all appeared mostly identical, under the hood they were different though which allowed some ping cannons out perform others.
2 Likes
I agree with Max that the large amount of copying is a result of less overall participation in this season of Terminal. I think the most interesting parts of terminal to me have been my work on my decision engine and simulator. Perhaps my lack of copying certain standard structures has also hampered the ability of my algo to perform as well as I had wanted this season… I think that factories as we currently have them are a strong force in making certain setups perform better than others. I mean, you have to focus on building and protecting them and try to destroy your opponents factories too! If you lose the ability to do either of those things, you will be hamstrung, and this singular focus prevents as much effort on the defensive structures themselves since they have to be so lean so we can build factories!
1 Like
@eric574 I am curious and i noticed this during the season. I think it’s possibly one of the reasons why your team algo had such a big lead on the second place algos in terms of ELO, and i suppose this is just a side effect from being in teams. Since each team member can upload 6 algos and then the team can have 6 algos, and possibly more, based on how a bug i found, seems to behave, but I guess i never put too much in an attempt to reproduce consistently. I noticed my algos playing against algos with users who didn’t appear on the leaderboards, I presume this is becuase they’re in a team. So if a team uploaded their 6 team algos, and each team member also uploaded their 6 algos that were all possibly varients of the teams main algo, ELO was still affected normally, I’m guessing your teams algos were able to maintain such a big ELO lead becuase of this?
The bug i found I made a thread in bugreport, and @RegularRyan has been looking into it, but I noticed when editing algos using the web editor, sometimes algos would get dupped, exiting the play/edit page I’d notice a new algo with the same name. The reason I believe this could be used to upload more than 6 algos per user, is becuase at one point i deleted an old algo from my algos page when it was full and then it ended up being full again, with a clone of one of the other algos, the reason i think it could be used to upload more than 6 algos is becuase the new cloned algo that showed up had an ELO rating like it had been playing in matches. I haven’t put too much effort into being able to reliably recreate this issue, and it mostly happened by accident. If i notice the same thing when next season starts I’ll maybe be able to put more time into reproducing it, so it can actually be fixed, I feel bad becuase I know with the info I gave Ryan, he was basically stumbling around in the dark.