Best posts made by JazzFunkNoob

Grimplex and I have prepared something very interesting for the upcoming ladder sweatlords! We sat down and sandboxed buildorders for a selection of maps in the current ladder pool. We prepared cheat sheets for each of the BOs with replay Ids and images. The buildorders are aimed for people over 1300 ladder rating and tend to focus on one aggressive aspect. We suggest you watch the replay, sandbox the Bo once or twice and then use the cheat sheet to remember it. Currently we have prepared:

• Arcane
• Zeuthea valley
• TAG_Craftious_Maximus
• Bermuda Locket
• Emerald Crater
• Crossfire Canal
• Cobalt Valley
• Theta Passage
• Loki

Sheet: https://docs.google.com/document/d/1ct-M0BzpywXGy5ffmOsahknVcYl2weU8AczBkOliMOs/edit?usp=sharing

Feel free to experiment with the BOs, you will likely be able to find some improvements or develope a counter. Also feedback on the sheets is very welcome :).

Edit:
Selection for the current Month:
https://docs.google.com/document/d/1mpr8MnNwtvADwvYCWaskt3nNLmqu_jyLKDYB5a2XTHY/edit?usp=sharing

First off: Just relying on sniping is not good strategy and it is not desired that low rated players only know how to use that strategy. You will just put yourself into a dead end if you only focus on one strategy as a beginner. So from a game director perspective it is good to not allow single strategies to stand out, especially on a lower level.

A t2 air snipe usually is a massinvestment of 3k mass, which needs to be quite early in the game. Doing in a 1v1 usually means you will loose so much map control that you will autoloose the game if your snipe fails. So it is more of a last resort/ meme strat. In a 2v2 the frontline is more densely populated with ACUS which makes it less of a setback diverting resources into a snipe. And a snipe without Fullshare would still win you the game in 99% of the cases. So it is as effective as in a 1v1 with a lot less risk involved.

A snipe with fullshare is still very viable though. Even the best player is very voulnerable for the next minute after loosing his teammate. He has to take over the base and redivert units to fill the gap his teammates ACU left. This is when you must make your moves. Army movements force a reaction of your opponent, diverting his attention from managing the new base, arty drops have a very high chance to make it into your opponents base and causing huge damage. Snipes in 2v2 full share don't win you the game, but they win you the initiative wich usually is key to winning games.

There has been some discussion about trying to put Ais into the ladder queue. I would like to make this a proper discussion. Personally I would be interested to see how ais would actually perform in ladder, aka which rating they would converge to. So the question which ai is the best competitive player could finally be answered! This would probably not disturb the lower ladder regime too much as each Ai would only need to play 30 games or so to have it's rating converge.

A more interesting role of AI in ladder would be to have new players play their placement games against AI. I want to present a concept and go into some benefits and problems.

Ais would first play ladder games until their rating converges. New players would face these Ais in their first placement matches. Which AI they face could be determined by the AIs they beat in the previous placement matches to create a less frustrating experience. The rating of the Ai would not be affected by placement matches to prevent the biass from mostly facing new people shift the AIs rating. It would instead play some games against converged players in ladder to readjust it's rating to possible skill/ meta changes.

So why would we even want to do it?

It would prevent new players facing ~1300 rated players in their first matches and vice versa saving 1300 rated players from having to beating up noobs (This was quite frustrating when I was 1300). I don't have any data of how many noobs and players you face at that rating range but it is quite significant. I found out that there are about as many people who played one ladder game as who played more than one ladder game in the data from 2014 to 2020. So there were a lot of players who clicked the ladder button, got absolutely wiped and then never clicked it again.
Edit: Apparently fixed, new players get focre matched with people with lower rating. Getting a more accurate starter rating might still be desireable, as it causes less unbalanced matches.

The reasons I see against it:

Players might not want to play against AI in ladder but people would possibly prefer to face an AI over getting wiped by 1300s during placement. The games the AI would have to play in ladder to readjust its rating are probably neglible as I would expect it to be less than one per week. The initial games to get the AIs rating might cause some upset though.

I'm not even sure if it would give the result we are hoping for: For instance the AI might be too weak to be a meaningful opponent. If the AI constantly looses to new players the system won't adjust the players rating by a lot. µ and sigma only get adjusted marginally when a higher rated player wins against a lower rated one.
(http://www.moserware.com/assets/computing-your-skill/The Math Behind TrueSkill.pdf, page 29 onwards)
So extra information of the skill of a new player we get by having him face a weak AI is very small and he would have to face 1300s after his placement matches anyway.

The fact that we are changing which rating new players face in their first matches might also change the mean of the whole rating distribution. But I'm not certain on that. Fact is that trueskill has a rating deflation over time. It is observable in the distribution of ladder µ over the last couple of years, but appears to be stabilizing. Avrg.rating should be mean. µ of all players that played at least one ladder game in the previous year and had a sigma<100:

In summary I think having players face AIs during their placement matches is a good Idea, given we have an AI that can actually beat the average ladder first timer. The other approach would be to manually match these players against weaker opponents. This is probably fine for new players but it made TMM quite a nightmare after the reset as non converged strong players were matched against converged noobs.

Pls no more FAF elections.

I have tried myself quite a bit on mathecatically describing balance. I am currently at the stage: Complex system even just for direct fire units you will only get to a precision of 10% of the balance mass cost at best. There are so many unit stats that change the strenght of a unit. Easiest to start is with effecitve DPS, HP and Masscost. But then you have a lot more stats like energy cost, movementspeed, range!!, firecycle (regarding alpha damage and overkill, which also quite depends on the army composotion you are facing.), turret turn speed, hitbox size, acceleration. And I am sure there are at least twice more stats than I have named here that effect balance. So a theoretical approach via those unit stats becomes very hard to do. It can be usefull when you have to first set a baseline for balance for a mod etc. I did help Dragun with his SCTA mod a bit with this and an efficiency calculation based on lanchesters laws. But that is only a base line after that you can only improve balance via playtesting.

I am going to talk about the medusa bomber stun issue as an example. I find it quite hard to compare the stun of a medusa and the bomber. Mostly because one is an air unit and one is a land unit. So the counters to each other are quite different. Also the bombers main advantage is that it is very fast and is not restricted by terrain. So it needs to be less efficient than a medusa because it can attack way behind the front line whereas the medusa is not capable of doing so. The stun is rather a gimmick for later game use of the bomber because it is cheap and can stun high tier units. Also medusa usually gets dodged, whereas t1 bombers attacking an army usually don't get dodged.

But the biggest issue in my opinion is that you are asuming all tiers of all domains of all factions should be equally strong. That is not the case and is not desired as far as I am concerned. Quite similar balance is required because of the map diversity present in FAF. Some domains have a huge advantage over other on some maps. This results in very similar factions if you want to have interesting games between all factions on all maps. But you still have a bit of wiggle room and in my opinion it should be used. So that is the reason there isn't even an intent to make all t1 bombers equally strong (even if it was possible). They need to be comparable but I am quite certain if you made a statistial analysis of all replays and just calculate the average vet of all t1 bombers in all games it would not converge towards the same value for all bombers. And quite often you don't even have an equivalent unit in the other faction.

Balance is messy and as far as I can tell it is currently done by a couple of the top players having a discussion about their intuition of what is op and what is not. Another problem on top of that is balance is distorted on what tactic (unit composition etc.) currently is in fashion. Also ease of use has a huge impact. Most memorable example that comes to my mind is the introduction of SCAU presets.

If you are interested this is the (old) combat efficiency of selected units calculated via a partial discretization of lanchesters laws wich only takes HP, Mass and avrg DPS into account:

As people age they have more money and less time. Can't wait till FAF is mainly played from retirement homes with 20 apm and players falling asleep during a game.

Just remove seraphmi, cybran and aeon from the game. All hail the uef mirror matchup. Perfect balance.

I did find a workaround to watch replays from a file while being in queue, which is a requested feature but is not implemented into the client. For that you will need the Python client in addition to java client, as it allows launching replays without the client actually running. Download here:
https://github.com/FAForever/client/releases

Then you will also have to add a section of code to the game.prefs file in order to launch multiple instances for FA at once:

debug = {
    enable_debug_facilities = true
}

source: https://forums.faforever.com/viewtopic.php?f=2&t=14434
Finally you will have to select the python client exe as the (default) program to open replays:

Now you should be able to launch replays while in queue. But the replay must be of the current game version. Otherwise the launch will fail I presume. (Afaik this is the reason it is not implemented in the client atm)

I'm all in for the labs and scouts. But I think any other unit should keep doing so for asthetic reasons alone.

Also FAF live should host a credits (points? mass? what ever it is you are gambeling with) leaderboard, since the voting is such a big deal in every stream.

If you are lucky this is a fixable issue. But I remember the original lead engineer saying in an interview that they had some issues with some hardware not being deterministic with certain operations (hardware accelerated sine/cosine calculation was the issue if I remember correctly). If the emulator is partially non deterministic then there is nothing FAF can do.

Currently there are a lot of suggestions for new matchmaker queues. But I'm afraid the player base is way too small for having that many queues running at the same time. Also maintaining that many pools is quite a lot of work. Instead I would propose having only one additional queue which has a different theme every pool rotation. For instance only mapgen team games, only 5x5 1v1s, setons, phamton, SCTA, shared armies, survival and what ever else might be fun.
It would also give the matchmaker team the opportunity to test out certain maps/concepts that are too experimental for the competitive queues.
In other games these special events/ minigames are very popular amoung beginners/ casual players.

For putting it into practice: I expect that launching 3v3+ matchmaker games is only a matter of time. I suppose having mapgen maps in the matchmaker is also planned. How hard would it be to add mod support?

Maintaining the pool would be a lot more work than the usual 1v1/2v2 pool. As every month you would need a new concept that is somewhat coherent. I could imagine some modders/ mappers being motivated enough to propose a whole concept (which would highlight their work) which would only need some adjustment. At the end of the rotation it would be very important to evaluate how well the current concept was received. There will be a lot of bad ones for sure.

There should be an advantage to not store up 20000 mass until you have your unit.

Signing up Blast and me
1935
2077
Teamname: circus rejects

Probably everyone higher rated has noticed that TMM is pretty dead in the 1500+ range. Part of that is certainly that it isn't new anymore and the race for the top 10 on the leaderboard has been won.
But a big part of the inactivity I would put on the map pool. I personally stopped playing it cause I didn't like 70% of the maps I was getting. I still don't like the maps I'm getting and I am the guy that curates the maps for the TMM pool atm. Most maps I don't like are for the <1000 rating bracket, which makes up more than half of the pool:

On top of that, if I queue solo I have a good chance of being matched with a <1000 player against two 1400s so everyone in this constilation is locked to the sub 1000 part of the pool. So the effective chance to get a 1500+ map as a solo 1500+ player probably 20%.

I have suggestions to fix that: First one is to pick the maps not by the rating of the lowest rated player but by the avrg. rating of all players. I expect a low rated player to be able to handle a more complicated map with the guidance of a higher rated player, or the higher rated player just grabbing a bit more of the map.

Second one is to reduce the pick probability of a map if the avrg player rating is significantly higher than the bracket of the map. (exact Probability distribution to be determined)

Let me know what you think.

To add to @archsimcat calculations:

when you take into account what your mex is usually replacing the t3 mex suddenly becomes a lot more efficient. in the best case when you reclaim your t2 mex and have a lot of bp on building the t3 mex it has an even smaller repaytime than upgrading a t1 mex to t3. When upgrading it is still more efficient than ringing mexes.

So when someone properly enters the t3 eco stage it can be considered some kind of eco runaway. I would consider the t3 mex to have a 10% shorter repay time than capping t2 mexes in a realistic szenario where you are reclaiming your t2 mexes.

To put that into perspective: Usually a 1v1 is over when one player has a 10% total mass lead and isn't significantly behind in army size. All because of that exponential groth nature of eco scaling.

When switching to a 10% more efficient eco stage you have the 10% generated mass lead pretty quickly and the 10% total mass lead short after that.

My cunclusion is that the efficiency increase of switching to t3 eco makes the game quite unstable at the transition stage. If you aren't able to mirror your opponent in first t3 mex or deal big damage right after the transition you are screwed. The question now is if it is a bad thing to have this deciding point/phase in a game. On setons navy for instance it is meta to not invest into heavy offense before you have at least some t3 mex. I would consider that a sympton of not falling behind on the t3 eco switch.

I don't feel like I have understood the problem well enough to have an opinion on whether t3 mexes should get changed let alone what change would be good for gameplay. Maybe this theory approach helps to form a better decision.

equations behind the sheet, cause I am too lazy to explain:

Thanks for being so committed to tracking down the issues.

That would change gameplay quite signifficantly. You would be able to distinguish between a t1 tank and a lab in the beginning easily. Atm you could spot the different movementspeed but that is quite hard.
I don't think it would be good for gameplay.

I just saw pillars continue to focus an acu despite t1 tanks running into range of the pillars. I thought tanks were supposed to switch to tanks when they come into range of them. Otherwise sniping acus could still work by just rightclicking the acu once and then following it around.

At min 13 you can see it happening to grimplex: #14540450

@ThomasHiatt
I suppose it is just a matter of taste. I didn't enjoy learning bos or even standard tactics at all up to the point I got like 1300 global. Was still fun.
Now I don't enjoy learing bos but I accepted i have to do it to at least some degree. Or at least get a feel for the map/ knowing about the tactics that can work on the map.

Playing bad is still fun for a lot of people. But unfortunately not for the majority, as they are too afraid to leave the astro/gap safe space.

I don't think that the people in lower ranked ladder mind playing on unknown maps. They already are comfortable with a lot of maps.