FAQ Authors

don't use the new proton_experimental. while it works for FA from steam FA run from FAF fails; this is what you'll get if you try to run it with proton experimental : https://askubuntu.com/questions/1309419/trying-to-run-proton-wine-app-getting-nice-issue

On another note I'm getting another issue. On my new install of ubuntu 20.10 and with proton 5.13 (in which case the nice error goes away), but my FA stops immediately with no log output as you can see :

2021-01-20 03:53:57.860  INFO 396709 --- [      Thread-24] c.f.client.fa.ForgedAllianceService      : Starting Forged Alliance with command: [/home/t/faf/run, /home/t/.faforever/bin/ForgedAlliance.exe, /init, init.lua, /nobugreport, /log, /home/t/.faforever/logs/game_13674228.log, /gpgnet, 127.0.0.1:13904, /mean, 1261.25, /deviation, 93.3293, /savereplay, gpgnet://127.0.0.1:45827/13674228/tatsu.SCFAreplay, /country, FR, /clan, BPT, /numgames, 2373] in directory: /home/t/.faforever/bin
wineserver: using server-side synchronization.
2021-01-20 03:53:59.067  INFO 396709 --- [pool-3-thread-5] com.faforever.client.game.GameService    : Forged Alliance terminated with exit code 53
2021-01-20 03:53:59.391  INFO 396709 --- [      Thread-24] c.f.client.fa.relay.ice.IceAdapterImpl   : ICE adapter terminated normally

Ladder (or rather matchmaker now) has adjusted the new player algorithm to be based on a linear extrapolation that eventually converges to player mu after a few games rather than the traditional matchmaking process of trueskill. New players are now usually facing 500 rated players in their first game. If they win, they face players closer to their mu, and so on.

Do these alternatives not already exist? Half of all FAF games are AI games whether coop, survival, comp stomp, whatever. I should probably also say that I think using AI to "train" for a pvp game is not an efficient usage of your time if you plan on improving and has never been considered a viable strategy in any rts, or really any game, I have ever followed. It just seems like another form of feature bloat to make people feel comfy without really making them much better relative to just playing a game and then asking someone to help them out by watching the replay.

I don't have a problem with making not terrible training material, MazorNoob.

But this is what FAF has as a market force for new players:

game is good

That's it.

The people that "get kicked, get no games, get wrecked" get a shit game experience and now leave the game. The only dudes staying around to say "no, I don't care about this terrible experience" are those so absurdly stubborn or so deadset on the game being good that they are willing to take the step to actually improve themselves. Contrary to popular belief, people don't like to actively improve. Most people do it by accident because they find something fun and then get better at it. Usually, you actually need to pay people to improve at things they didn't find fun in the first place. This is why I don't see the point to this AI stuff. If I didn't have a fun game experience I'm uninstalling this garbage client and going to go play something fun instead. Why would I go and play AI lol.

If they found the game fun, we are VERY likely to NOT lose these guys. This is why I do not see this as a retention solution. This is a solution to making it easier to improve yourself at the game, which is a subset of the bigger problem.

There is basically ZERO community built around the game to attach yourself in. There is BRNK, but the problem is that he doesn't really have any desire to associate with FAF itself so it's impossible to use him as some sort of leverage for retention improvement.

Just to reiterate, if people want to make better training material feel free. There is zero pushback against it. But it isn't going to do much at all related to retention.

Can someone explain why playing AI training games retains players? It feels like you guys are conflating “trying to make players not lose” with “retaining players” because the latter is infinitely more difficult and basically impossible to artificially do.

You essentially need some sort of cultural force that can push some sort of zeitgeist at lower levels and create a community. It’s what Yuri did with creating his PRO clan and his weekly Russian tournies as well as the giant focus on astro games. Likewise, you see BRNK’s Discord do a similar thing where you basically have this segment of the community that entirely interacts with itself via community/stream events.

Any sort of half-assed “yeah just dump all Germans into a German text/voice channel” is the same as your mom telling you to go play with the neighbor’s kid. You’ll do it because mom said so, but it won’t lead to much. Likewise, this training stuff is just avoiding the real problem and attempting to answer a totally different one.

Step 1 is creating some sort of cultural platform that can actually reach new players. Whether through the client, youtube, streams, or whatever else.

Step 2 is finding people with the time/energy/existing cultural capital to actually have an impact. Obviously you would typically gravitate to FAF personalities that already exist, but if none fit the criteria, then you need to make someone that will.

Step 3 is everything else.

If you really care about retention, become a regular streamer that does community replay reviews and tries to get new players to do small “grudge best ofs” against one another. Try to make a discord for this stuff. Help do things like a monthly podcast explaining FAF news or some sort of fireside chat with FAF people just talking about life or the game. Literally do anything that shows new players the community has a pulse and that they can interact with it or at least form some sort of parasocial connection with already formed “groups” on FAF.

All of this is just dancing around the fact FAF has a terrible social presence across, well, everything. Find or become the guy that is willing to cut off his ear to show his passion for the project and then FAF itself can do its best to actually promote you and create a positive feedback system for retention results.

When I read arma's suggestion, I get the picture that it is appealing to the guy that is going "damn, this game is tough but I really enjoy playing it, I'm sure glad I have this easy to access material to improve myself" which is hardly the sort of person that FAF needs to work towards retaining. Would it be better to have easier material to allow these players to improve? Yes, absolutely. But that's improving people already captured and enthralled with the game. Not enthralling them.

To put some of my own personal experience with regards to games and FAF in:

I don't think I've ever stopped playing a game because it was so hard that it demotivated me. I have, however, stopped playing a game because everyone that I know relative to that game stopped playing. I've even stuck to games I personally hated playing just because friends wanted to keep playing it.

With regards to FAF, I've taken several year+ breaks from the game. I first played FAF around 2013-2014 and basically farmed up to 1k rating on gap. I pretty much hated the map, hated the people, and kept blaming noob teammates for why I was losing. So I quit and left FAF for like a year. I came back, played a couple more games, then quit again for a year. Then I came back around 2016, played with Farm, shitposted with him, made a clan, and suddenly I'm now a council member.

You could make the game so easy to understand that a monkey could play it. I wouldn't care. I would still leave sporadically because I have no chains attaching me to the community. Your tutorials/training/babying could result in people getting 200-300 rating higher than they would before, but in the end they'll still leave randomly since they feel no sense of obligation towards staying.

It's popular with Russians almost entirely because of Yuri and his casts.

I’m curious what FAF’s retention rate is relative to other games, particularly AoE and SC2. I know in app development, a 10% retention rate after a month from download is considered amazing.

I do wonder if relative to other games FAF is performing at expectation or even above expectation with regards to retention. Not saying it can’t be improved, but some of the ideas here are quite drastic and generally are more likely to turn people off than encourage them to stay.

Like forcing people to play an ai game before they can play online? That sounds like the ultimate turn off for a lot of competitive players, I know I would have hated it when I started out. If I wanted to play AI I could have done that on Steam.

If you really want to make the most efficient bang for buck effort wise, you improve retention by creating a community. Stuff like making ladder less scary doesn’t do much, we already have teamgames (which are less scary and more popular) as well as teamgames with deeply simplified gameplay (that lowers the amount of things you need to worry about).

How do you do that? No idea. New players don’t seem to like talking to other new players as evidenced by the utter inactivity of #newbie channel and even new players on the discord and in aeolus. Some sort of incentive to get interaction going among new players would probably be the first step.

Cybran being worse at T2 stage isn't a problem when they have the capacity to address the weakness through a quicker tech transition or through other forms of attack. When you adjust mex values, their range of viable options/margin for error decrease. This means what was once NOT a problem is NOW a problem. It isn't some RANDOM problem that always existed.

It's almost like the game was balanced on the assumption of certain things like mex output staying consistent. Crazy right?

"FTX, you continually engage with an arrogant attitude and consistently refused to actually make an argument, rather offering attacks and unsupported denials of others valid arguments, until I pointed out you weren't actually making any relevant arguments. You can do away with the hyperbole."

Pot meet kettle, shut up dude lmao.

Just utter delusional garbage based on some experience you have in dual gap or whatever? "Illusion of choice" like there is no such thing as a t2 factory in a teamgame. No one does a t2 push in any of your teamgames? Maybe play a decent teamgame some day. We made a matchmaker just so you could experience one. Would you like me to link you some teamgames that ended at t2 stage so you could see that games don't converge to t3 mex just because player count is larger than 2?

I fail to see how it's disingenuous. Later tech units are more destructive and they hold more mass. This means that as games go longer and longer, the margin for error decreases as a push failing leaves more and more mass behind. At some point, you will simply lose the game due to one bad push due to the mass invested into it, and this is what allows games to conclude. If the game gets to the point that there is no room for massive pushes because both players are averse to the possibility of donating mass, then the game goes to game enders. You making later tech units appear later means we spend more time with higher margin for error, therefore increasing the length of the game.

This transition is reflective of your adjustment in mex output. You make teching less effective, you make units more effective. It's zero-sum.

As an example, if I make a t2 mex cost 90k mass to upgrade, do you think that this has no consideration in unit balance? How effective do you think Aeon gun com is going to be on maps where you will never be able to get beyond 30 mass income?

Do you think on "mass glut" maps that certain factions did not have an advantage? Why did UEF and Cybran get picked all the time on Ditch, especially before t3 nerf? Was it perhaps a combination of units that make them incredibly strong on high mass maps aka the amphibious siege bot and drones? Do you think these considerations matter as much when I make a t2 mex 90k mass on Ditch?