yesterday’s playtest was awesome[1].
not sure yet if projectiles or hitscan is more fun. lots of balancing left to do, but the movement already feels amazing.
who wants to 1v1?
cash bounties coming soon.
sell cpp17 code in the item shop soon after.
541 karma · joined June 15, 2012
yesterday’s playtest was awesome[1].
not sure yet if projectiles or hitscan is more fun. lots of balancing left to do, but the movement already feels amazing.
who wants to 1v1?
cash bounties coming soon.
sell cpp17 code in the item shop soon after.
step one is acknowledging that a problem exists. this can be hard, especially when one is part of the problem.
step two is overcoming learned apathy. this is often harder.
step three is making things less shit, one day at a time, in any capacity that you can. this is awesome. everyday can be the best day of your life.
good for you. can’t wait to visit melbourne and interact with even a single, slightly less shit, piece of engineering.
viva le renaissance.
only thing worse than dysfunctional companies is dysfunctional technology.
someday we’ll get the incentives aligned properly, and thrive.
my point was that it’s not really needed.
if you’re migrating off, you already have an sdlc that you like and want to preserve from provider degradation over time.
simplify it a bit, and encode it directly into aws or ovh. then it’s permanent, and grows insanely robust over time.
sdlc doesn’t not need to be constantly evolving. evolve it for the next greenfield product.
a go binary in a zip builds and uploads to lambda in 1 second. handle routing and everything else in binary, don’t use aws features. you don’t need em.
lambda, s3, and dynamo all scale to zero with usage based billing.
toss in rds if you really miss sql or dislike scale to zero.
once this is too expensive, move to ovh metal, which sells significantly faster compute, ie epyc 4244p ddr5 5200.
more importantly, they sell bandwidth for $0.50/TB instead of $0.10/GB, which is aws price before paas wrapper markup.
the ovh price is after the 1Gbps unmetered you get for free with any server.
most companies will never even need metal, and between lambda and metal is ec2 spot, which might be of use anyway for elasticity.
ovh metal bills monthly, ec2 spot bills by the second. recently i learned ec2 spot in localzones is x2-3 cheaper than standard spot. i only use r5.xlarge in los angeles these days.
ovh metal has an old fashioned but perfectly workable api. aws has a great api.
spend a few days figuring out a decent sdlc, and then freeze it permanently. it will grow insanely robust over time.
i haven’t published my ovh metal workflows. my aws workflows are here[1].
lower overhead means more interesting products coming to market. lower friction definitely doesn’t hurt. viva le renaissance!
then you will be banned from ranked, unless you have a clever way to synthesize perfect handcam video.
in which case, you get the entire bounty pool and 10% profit for 6 months.
then you will be banned from ranked.
the bugs won’t discover themselves. i salute you.
yet cheaters will still cheat.
non-human inputs IS cheating. period. there is no information cheating possible in this game.
input provenance is the only anti cheat that has a chance.
full fidelity 120hz server replays with handcam seems a good place to start.
we’ll see how it goes!
there are however legions of fortnite players who would mortgage their lives for a better game. a handcam without audio during gameplay seems a more reasonable and low price to pay. time will tell what gamers think! very exciting.
it’s not about casual vs competitive. it’s about how the game feels to play, and how the adrenaline hits your blood stream.
handcam-anticheat is ranked only. pubs has good old fashioned try-our-best anti-cheat. aka thoughts-and-prayers anti-cheat.
what does fix this is handcam anti-cheat, which makes it trivial to identify the operator by their hand movement patterns, arm size, etc.
you can review not only the match you are in, but all previous matches. full fidelity replays + handcam = a brave new world.
account sharing is not allowed in ranked. in pubs, go crazy.
very well, ranked will then require specific hardware. i’m gonna have to build more test pcs. 9800x3d 7700xt b650m looking pretty clean.
handcam anti-cheat and replays will handle provenance of inputs. non-human inputs are cheating.
you could try to boot in vm, but performance will suffer. the server moved off vms for the same reason.
the only thing i really care about is non-human inputs. if you get really creative with booting the game, and aren’t nefarious, i’ll set phasers to stun.
handcam anti-cheat and the replay system is how we deal with non-human inputs. ranked only.
in pubs, you may encounter a cheater. hopefully owning the whole os makes that easier to detect, but i’m not sure it will.
the goal is to be minimally invasive. you don’t have to install anything!
i don’t fight your misconfigured pc, you don’t worry about me scanning your c drive.
nothing to install, configure, save, or lose.
i can't tell users to go into bios and turn off smt. i can tell them to boot an iso, and have it preset to do that via kernel cmdline. owning the os means i can tune the network stack, the os, the kernel, anything that helps performance. the game itself has minimal config, only keybinds.
i'm really targeting one player base: fortnite ranked/competitive players.
fortnite competitive does exist on console, but barely. it's an after thought, and the game can barely run, even on ps5. most importantly, it doesn't let you set graphics to the settings every pro uses. on console you have default high graphics. all serious players migrate to pc.
the game will run on beater laptops, but it's not a great experience. the players i'm targeting all have desktop pcs and 1080p240 monitors already. if i can get 1% of 1% of fortnite ranked players to try, that would be inglorious. if it's just me haning out with 99 aws servers, that's fine too.
i want there to be an on ramp for new players, but optimizing for low end spec is not a goal. especially with 100 players all standing next to each other, low end hardware falls over. will probably have to limit to 50 or even fewer players depending on how much low end hardware shows up to play.
input cheats are common on console for fortnite, i'm not sure about other games. reads video from hdmi, mitm controller outputs. multiple generations of that, some of it works on pc too, and then pc has a whole new host of similar tech.
input provenance is the only real solution that i can see. the rest is shadows on a cave wall.
[x670e][~/repos/WickedEngine/Protoverse/reboot-to-play][better-game][us-west-2][-][master]
>> bash flash.sh /dev/sda 1224736768 bytes (1.2 GB, 1.1 GiB) copied, 52 s, 23.3 MB/s
when you lose a fight, you drop into replays without leaving the game. scrub back like in a video editor, and watch the engagement from the other players perspective. the handcam footage is available here. cheating should be obvious. report or no, then return to the game, where you can spectate the rest of the match like a ghost.
fortnite replays are commonly used by pros, but less so by more casual players. their main issue is that you have to leave the game to get to replays. they also only kind of work, and don't show full server fidelity. our server ticks at 120hz, and replays are full fidelity.
lots to figure out still, but that's the idea. ranked won't launch for a while, but i wanted to design around anti-cheat from day 1.
failure is a distinct possibility with this game, that's ok. success would be interesting too.
i'm not quitting the day job to build this game. if it never makes a dime that's ok.
i started this to study fortnite and understand it's tradeoffs. turns out, it's mostly just tech debt and accidents of history. i wanted to see if i could do better, and i could.