Fractal Softworks Forum

Please login or register.

Login with username, password and session length
Advanced search  

News:

Starsector 0.97a is out! (02/02/24); New blog post: Simulator Enhancements (03/13/24)

Pages: 1 2 [3] 4

Author Topic: 10th Starsector AI Fleet Building Tournament  (Read 12521 times)

Tartiflette

  • Admiral
  • *****
  • Posts: 3529
  • MagicLab discord: https://discord.gg/EVQZaD3naU
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #30 on: June 29, 2020, 11:33:15 PM »

Well, this tournament is truly cursed but we have streams!

The first half of the 3 Round will be cast TODAY Tuesday June 30th on FallenShogun's channel, cohosted by Laharl, at 7pm UTC:

https://www.tickcounter.com/countdown/2059503/round-3-day-1
https://www.twitch.tv/fallenshogun

ORA+SRD: Purple Pirates Platoon
XE+VHT: Gamer Team 6
DA+II: Team Awesomeforce Prime
SRA+II: Probably Cancer
SRD+PV: The Galatia alliance
AL+MN: Light of Sura

NC+DME: Unlikely Alliance



The second half will be streamer Wednesday July 1st on Nemo-Naemo's channel, co-hosted by Aurica, at 7.30PM UTC:

https://www.tickcounter.com/countdown/2059509/round-3-day-2
https://www.twitch.tv/nemo_naemo

PRV+BrDy: Merchant Guild Fleet
DA+SN: Bad Dudes
SRA+LoA: The Light of Nait
VLT+SN: Yggdrasil's Electric Time Bandits
PP+MN: Spaceballs: The Tournament Team
THI+VMD: Dakkaholics Anonymous

DME+Red: The Beehive Fleet
ES+KoT: For Pizza
Logged
 

Titann

  • Commander
  • ***
  • Posts: 122
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #31 on: July 05, 2020, 12:52:34 PM »

when is next tournament? It was fun to watch.
Logged

SCC

  • Admiral
  • *****
  • Posts: 4112
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #32 on: July 05, 2020, 01:06:34 PM »

This tournament has not ended yet. It will end with the fourth round, which is going to happen in two weeks, on 15th of July or about that date. After that? Tart's going to take a break, a year or so should suffice.

Seita

  • Ensign
  • *
  • Posts: 15
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #33 on: July 14, 2020, 12:48:09 AM »

Tournament week?
Logged

Tartiflette

  • Admiral
  • *****
  • Posts: 3529
  • MagicLab discord: https://discord.gg/EVQZaD3naU
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #34 on: July 14, 2020, 01:53:56 AM »

I'll probably have an announcement today when/if I'm confident I can finish everything left to do.
Logged
 

Tartiflette

  • Admiral
  • *****
  • Posts: 3529
  • MagicLab discord: https://discord.gg/EVQZaD3naU
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #35 on: July 15, 2020, 01:24:42 AM »

Well, tentative stream announcement:

The first half of the Round 4 will be cast Wednesday July 15th on Nemo-Naemo's channel, co-hosted by Aurica, at 7.30PM UTC:

https://www.twitch.tv/nemo_naemo

The second half of the Round 4 will be cast probably on Gloracle's channel, otherwise possibly on Nemo's (I'm waiting for some responses)

https://www.twitch.tv/gloracle

The matches will be run in the order of the ranking, with hopefully 7 ranked matches on day one, the rest on day 2.
« Last Edit: July 15, 2020, 02:30:29 AM by Tartiflette »
Logged
 

Modo44

  • Commander
  • ***
  • Posts: 144
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #36 on: July 16, 2020, 12:27:50 AM »

So what was that 7th wave? We need to know, man.
Logged

SCC

  • Admiral
  • *****
  • Posts: 4112
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #37 on: July 16, 2020, 12:33:34 AM »

SINDRIII!
The mod will probably be released tomorrow, after the tournament has ended.

Tartiflette

  • Admiral
  • *****
  • Posts: 3529
  • MagicLab discord: https://discord.gg/EVQZaD3naU
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #38 on: July 16, 2020, 12:45:56 AM »

Sadly it's not in a state to be released yet. I'll try to have a barebone release Sunday, but no promise.
Logged
 

MrDaddyPants

  • Ensign
  • *
  • Posts: 16
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #39 on: July 16, 2020, 07:46:51 PM »

It's finally over 8)

Thx Tart and Nia for organizing.

My feedback on the tourney. (again it’s feedback.. not meant to be “ you made dis, I *** all over it post..” )
Wall of text incoming !

1.   I liked soft antispam rule. On one hand I hate antispam, because it means weaker fleets overall, on the other hands it softens blow from overpowered or wrongly priced ships, that some mods or vanilla have. Better balance is worth it.

2.   I don’t like vanilla weapons exclusion. Mods shouldn’t be balanced around its mods weapons exclusively, if your hulls are broken by vanilla weapons, its bad hulls for the tournament, or campaign for that matter. More weapon choices means more variety and more interesting builds in the tourney. For example I wanted to put flak canon in hujings turret. I believed it would be the best weapon for the slot in the OP budget I had. And flak canon is never considered OP or anything. Well I had Burst Flak Cannon in Tiandong, so I can’t use flak canon (tourney rules). Burst flak was unsuitable and was 10 OP. So I considered even a railgun or small needler. But sadly had to pay full 10 OP for HMG as it preformed best of all bad options. Sadly ship was probably 4-5% worse because of that. And this kind of sacrifices add up, some factions are arbitrary luckier that they have precisely the weapon you need for specific build and with specific flux/OP stats. Difference between your faction having +-800 range kinetic and HE medium ballistic weapon (which you are forced use) or not (which means having access to like 10 vanilla weapons instead) is an arbitrary advantage, by virtue of not more powerful weapons, but having a choice for more suitable weapon for specific purpose and flux/OP budget.

3.   Some factions don’t have enough ships (this is exasperated by weapon rule) to execute multiple (or sometime even single) viable fleet style/strategy. For instance I don’t think high tech could really build a standoff long range fleet and its only realistic option is rush down wide fleet. Vanilla fighter split is arbitrary and unbalanced. I think if factions were bundled together it (something like VHT+ THI+ PRV) it would mean better tourney because of better balance and larger variety of ships and strats and more space for creative fleet building overall.

4.   This ties in into 70% CR. Most fleets were not at 70% CR to really experience this issue, but it’s not a lot of time for “steady” or “aggressive” fleets that are standoff/long range to safely kill the wave. Which is one of the reasons we chose rushing “balls to the wall” wide strategy. For instance defensive kiting ora+sylph fleet might be very strong in this format, but for me as player it seemed too risky to go for, solely because of CR. Also phase ships, certainly phase frigates, even with hardened subs are not viable, because of 70 CR.

5.   I don’t think the pricing of cruisers and capitals is correct. I’d do (at least!) -15% discount on capital ships and -10% on cruisers to make them more competitive. I feel that most players with capital ships would have gotten more points overall if they just spammed DD’s with some cruisers.

6.   Refitting budget feels bad. I think it should scale up the less points you have so you can fix your fleet and come back. And overall ratio of testing and prep before rd1 and after rd1 combined was like 7:1 for me. There was basically so little to do after rd1, it felt boring. Paying for hullmods again is weird.

7.   I don’t think multiple front attacks or constant switching of wave directions is good idea, maybe for boss wave. But overall it just screws some fleets over and forces things like SO on carriers. While some factions have a reasonably fast carriers from the get go.. it’s not encouraging wide array of ships and strats, it forces players to field more “generalist” ships.. And finally combined with huge battlespace it basically invites randomness in. Randomness anchored in AI *** ups. If every wave was top down, on smaller battlespace, I doubt you’d see -+4 points variability in fleets, because there would be less room for AI rng.

8.   And finally maybe waves 1-4 should be known before the round, or even at the beginning of the tourney. And also battlespace. So participants can actually build for the challenge they’ll face. The jazz around building waves so there is a good spread of scores, seems by definition that it would discriminate against some fleets, while helping other fleets.. We don’t have 10, or 20  or 100 rounds. For most “generally capable” fleet to win, instead it’s the one who got lucky that it performs well enough on unknown battle space against unknown waves. Having waves known beforehand would also transfer work load to make it all balanced from the organizer unto the players, and would hopefully prevent some uncompetitive 1 point flops…

9.   Officers. (I know they were implemented 5 minutes before the tourney..) I don’t think giving everyone same officers that automatically level up is good design. Officers and their levels should be part of the hull budget or some officer budget. Get 5 level 1’s. Or 1 lvl 5 or get another ship. As they were implemented you basically put an officer on your strongest ships, there was no min maxing or any real decision making involved. You could basically tell by looking at DP points of the fleet where the officers would be, and where the most experienced officer would be.

10.   Hull budget, while I had enough good frigates to make it work for me so I didn’t have an issue spending most of my budget every round. But I had this issue when I tried to build a high tech fleet. Sometimes you just can’t field useful hulls for your desired strategy that would fill the whole budget and you have a hole of 30k unspent or being 20k over... Being able to spend increments of (1000 credits) on officer exp would help with this. Or being able to transfer funds to the next round… Again balance wise it’s completely arbitrary if hulls in your faction are capable to fill your round budget well or not, it’s another 4-5% chance of your fleet being better if it does.

11.   And lastly I’m not a fan of vanilla tourney patches that change ships stats for worse. Like what happened to omen. If it’s was just badly priced. Just change the price, don’t break viability of a ship itself.

Logged

SCC

  • Admiral
  • *****
  • Posts: 4112
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #40 on: July 17, 2020, 12:35:09 AM »

3.   Some factions don’t have enough ships (this is exasperated by weapon rule) to execute multiple (or sometime even single) viable fleet style/strategy. For instance I don’t think high tech could really build a standoff long range fleet and its only realistic option is rush down wide fleet. Vanilla fighter split is arbitrary and unbalanced. I think if factions were bundled together it (something like VHT+ THI+ PRV) it would mean better tourney because of better balance and larger variety of ships and strats and more space for creative fleet building overall.
For what it's worth, high-tech isn't meant to be able to do that. High-tech ships have good flux stats, speed, short range weapons and are expensive, they're built to be aggressive. If they could reliably destroy enemies from afar, they could just kite enemies to death, with enemies being unable to do anything about it, unless they spam enough fighters or are faster than high-tech already. You could say that the same isn't true of low-tech, because of safety overrides, but, well, that's getting banned for the next tournament...

6.   Refitting budget feels bad. I think it should scale up the less points you have so you can fix your fleet and come back. And overall ratio of testing and prep before rd1 and after rd1 combined was like 7:1 for me. There was basically so little to do after rd1, it felt boring. Paying for hullmods again is weird.

8.   And finally maybe waves 1-4 should be known before the round, or even at the beginning of the tourney. And also battlespace. So participants can actually build for the challenge they’ll face. The jazz around building waves so there is a good spread of scores, seems by definition that it would discriminate against some fleets, while helping other fleets.. We don’t have 10, or 20  or 100 rounds. For most “generally capable” fleet to win, instead it’s the one who got lucky that it performs well enough on unknown battle space against unknown waves. Having waves known beforehand would also transfer work load to make it all balanced from the organizer unto the players, and would hopefully prevent some uncompetitive 1 point flops…
Refitting budget was meant to be "prepare for the next round", but it ended up being "fix your mistakes, if you know how to". Without knowing anything but the vaguest (and not even entirely true) description of low-tech, midline, high-tech, everything-lol rounds, there's no specialisation to be made, just do the best you can against "everything".

7.   I don’t think multiple front attacks or constant switching of wave directions is good idea, maybe for boss wave. But overall it just screws some fleets over and forces things like SO on carriers. While some factions have a reasonably fast carriers from the get go.. it’s not encouraging wide array of ships and strats, it forces players to field more “generalist” ships.. And finally combined with huge battlespace it basically invites randomness in. Randomness anchored in AI *** ups. If every wave was top down, on smaller battlespace, I doubt you’d see -+4 points variability in fleets, because there would be less room for AI rng.
I don't know if that would fix distribution. I'd have to ask Tart to open 4th & 6th tournament score sheets, because I don't remember how it looked exactly, but at least in 6th, with waves only from the top, there almost was a tie for the first place, second place was a tie and there was a bunch of people with third best score.
I do agree that current setup is very hostile to dedicated support ships. I think only your fleet could have gotten away with it, as the sheer width meant that there's always a warship to hide behind, but everyone else had to use their fast carriers, if they had any, or battlecarriers that could survive long enough for support to come. I would have to wait until Tart releases final round's variants, but I don't recall anyone besides SRAARS going for battlecarriers and using fast (sometimes SO'd) carriers instead. Its effect can be also seen in how low-tech fleets are made: that is, as high-tech as possible, with fast, manoeuvrable ships, because this playstyle that doesn't handle flanking well was in a tournament format that was about flanking you all the time. And also because SO is bonkers.

11.   And lastly I’m not a fan of vanilla tourney patches that change ships stats for worse. Like what happened to omen. If it’s was just badly priced. Just change the price, don’t break viability of a ship itself.
Cruiser-tier shields on a frigate is a bit too much, don't you think?

Tartiflette

  • Admiral
  • *****
  • Posts: 3529
  • MagicLab discord: https://discord.gg/EVQZaD3naU
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #41 on: July 17, 2020, 12:39:41 AM »

1. Antispam was too weak, making DD spam too affordable.

2. Most factions had access to vanilla weapons though, only a handful couldn't: SRA, DA and Brdy, and they all performed well.

3. The main issue here is that fleets are too big, at least 2, maybe 3 times larger than what I'd like to see. Ships variety wouldn't be an issue if an end fleet was a Capital ship and 2 cruisers. Additionally having distinct fleets to cheer for and playing a faction as best as you can are some of the main driving forces of these tournaments.

4. CR stuff is imperfect, but it is also a game design imperative since some mechanics rely on it, and there is a technical limitation component too.

5. I'm not sure I would agree given that officers help larger ships more. Sure DD spam showed its strength, but if there were more rounds I believe tall fleets could have taken the lead. Plus I don't want to get into one-size-fits-all solutions because it would have to be applied to everything, therefore everything becomes a point of argument. I'm working on a better system for the next tournament anyway.

6. It was a bit too short, I wanted to bump it to 150K but people have been screaming at me with every change I made to the rules or the mission mid-tournament. As for scaling, the problem become to balance it properly, therefore it would be a potential point of arguments. Hullmods are too cheap if you can buy them once and for all, and it would be an insane job to track who bought what when and make sure they is no cheate... nobody made a mistake.

7. Oh yeah, I'm sure you would have loved to be able to establish a solid impenetrable line of ships from side to side in a tiny map that no wave can outmaneuver. It would have been so interesting to watch!

Besides, you can test it: everything from the top favor the waves. It makes them more concentrated and drives up attrition against most fleets. Varied spawn zones made the fights more interesting to watch too. And after the Admiral AI fix, I don't think I saw carriers get caught alone: escort ships always had time to rejoin them. It also forced fleets to have such escort around instead of solely relying on "ships of the line". You say it invites randomness but most fleets were very consistent in their results, the exception being the fleets built atop a single ship that can be taken out, but for those, the battlesize doesn't matter.

8. No, deal with it.

9. I disagree completely, the point was to force participants to choose whether to invest in a large ship to make the most out of their allotted officers. Besides, officers are impossible to price (and would be yet another point of constant arguing), and doing so would compound with the issue of large ships being too expensive. Think of the current system as if capitals had their officer price "baked in".

10. Tracking leftovers is something I'd like to have, but is a nightmare to implement in a way that isn't another large overhead on the organizer workload. It wasn't a big issue in previous tournaments when there was a shared budget for hulls and weapons since it was easier to max out each round.

11. Most of the changes are vanilla changes for the next game update. other changes in the past have been integrated to the base game (such as sabot pods firing two missiles instead of 4 for example). Therefore, deal with it.
« Last Edit: July 17, 2020, 08:07:06 AM by Tartiflette »
Logged
 

Hiruma Kai

  • Admiral
  • *****
  • Posts: 878
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #42 on: July 17, 2020, 12:16:13 PM »

Cruiser-tier shields on a frigate is a bit too much, don't you think?

Sorry, this statement is a personal pet peeve of mine, and always rubs me the wrong way.  And I often see the sentiment expressed on the twitch streams.  I don't think its the shields on the Omen that make it stand out.  Omen has high tech frigate tier shields.  At best I can see it stated it has low or mid-tier destroyer shields.   Its shields are 33% better than a tempest or 40% better than a Wolf, if they were built in the same way.

I mean, if I outfit ships in the same way most people outfit the Omen, the statement clearly becomes false.  Cruisers can be outfitted to have 2-3 times the effective shield capacity of a maxed out shield Omen.  It is the combination of incredibly high speed beyond that of even most frigates (155 base is only beaten by the Tempest, Afflictor, Hound, and Hyperion), with a powerful and free ship system that is basically a medium+ weapon (seems better than a Flak cannon for example) with a low flux cost, and the expectation that ship doesn't need to do real damage, which then combines with high tech frigate tier shields.  The fact is the Omen can be built ignoring offense since it has built in offense already.  The whole package is really powerful, but its not because of the shields.  Drop the speed to 120, or swap the ship system with the Shrike's, and I don't think people would complain about the shield.

I'll note you can build a Shrike such that it has 360 degree "Capital-tier" shields in the same way that an Omen has "Cruiser-tier" shields, but I don't think I've ever heard anyone say that Capital-tier shields on a 8 DP Destroyer is a bit much.  It can even be built to have more shield per DP than an Omen, but I don't think I've ever seen that done.
Logged

intrinsic_parity

  • Admiral
  • *****
  • Posts: 3071
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #43 on: July 17, 2020, 12:32:21 PM »

I don't think it makes any sense to say 'cruiser tier shields' at all. Shield efficiency does not scale with ship class (shield efficiency can be any value for any size ship, it mostly depends on the design), and shield upkeep scales inversely with ship class. The only metric that scales with ship class is capacity, and the omen does not have exceptional capacity for a frigate.

Hiruma Kai hit the nail on the head with regards to the ship system. The omens system is super strong since it effectively counters fighters locally (frigates biggest weakness IMO) and gives a bunch of free shield pressure which is generally lacking on high tech ships. The speed is nice but I don't think it matters that much. The system is the reason to use it (and the great campaign hullmods), the survivability just lets it use the system.
Logged

SCC

  • Admiral
  • *****
  • Posts: 4112
    • View Profile
Re: 10th Starsector AI Fleet Building Tournament
« Reply #44 on: July 17, 2020, 01:15:53 PM »

Omen has 0.6 shields, unlike Wolf or Tempest, meaning it starts with 33% better shields, it starts with similar flux stats, its weapons are few in number and its EMP emitter is a very efficient and versatile, if random at times, weapon. It can tank quite a lot, simply because there's not much going on, unlike other ships, so it can disregard vents entirely. I wouldn't say ~12k effective shield HP is low-mid destroyer tier, it's closer to what Eagle or Aurora start with, about 90% as much. Unlike cruisers, it enjoys being a small, fast target, meaning that it's also less likely to get hit in the first place.

You might be better at building Shrikes than me, but my 360 degree 23k effective shield HP Shrike ends up being heavily reliant on sabots and in a way starved for dissipation, because even if the only thing it has is heavy blaster, 390 dissipation is still quite short of that. Regular Shrike builds are closer to 12k shield HP, which is the same as Omen, except on a more expensive, hotter-running platform. It's also a bigger target.

I guess that Omen doesn't truly have cruiser-tier shields, only almost cruiser tier shields that are hard to hit that its effective durability is around cruiser tier shields.
As for its EMP emitter, it's a part of what makes it good (and as a weapon, it's incredible), but for shield durability, the most important part is the flux draw, which I already mentioned. People like it because it's not just a durable distraction, but a mean anti-fighter frigate as well.
Pages: 1 2 [3] 4