Alpha test feedback June 28, 2014

Continuing the discussion from Alpha test feedback June 14, 2014:

Join us at for a 30 min playtest Saturday, June 28, 2014 at 8pm UTC, 3pm Central US, 4pm East US, 9pm UK

Release notes to follow…

Thank you for all your feedback from the last test! There were a considerable number of changes; some worked, some didn’t. We’ve been hard at work on a new and improved game engine; meanwhile we are looking to consolidate on last test’s changes based on your feedback:

These are the preliminary release notes for this week’s test:

  • Weapon is no longer an instant weapon, and targets need to be led - although it remains a combination weapon, with a faster fire rate.
  • Changed weapon sound - more suitable to its rate of fire.
  • Slower weapon projectile - more suitable to its visual effect, and important for leading.
  • Target leading - you will need to shoot where your target will be when your projectiles hit it rather than where it is now
  • Turbo no longer locks turn controls; we will revisit this when we have “turbo-type” choices.
  • Mouse dead zone removed - the area inside your reticule was a dead zone, making left/right adjustments hard
  • Non-linear mouse turn, turn adjustments as small near center of screen and accelerated as your mouse moves further out, though max turn speed remains the same, this will allow for greater precision.
  • Bots causing name craziness should be resolved
  • Player name display fixed in IE11

Some information on our upcoming new engine

  • Recent browser changes have allowed us to decouple networking and graphics into different threads, this means high network traffic won’t effect the framerate, and extreme visual effects won’t effect the network.
  • Efficient GPU pipelining, you’ve already seen a lot of ships and bullet hell - but we can do much much more, while still improving your framerate
  • New physics - expect collisions with other ships and space debris/asteroids for hide and seek!
  • Input controls - user assignable keys, improved touch and gampad support
  • Improved server messaging - where you are being shot from, who’s shooting you, targeting and tracking.
  • New weapons, new ships…
  • …and much much more

Obviously these improvements will all get their own spotlight in the release notes as they happen and will be explained in greater detail at that time.

If you want to know the bigger picture and where we are heading don’t forget to check out: About Age of Ascent

Thank you all for joining us on this journey!

Wow… Very impressive. New ships? Did I read that right? I find that I live for Saturdays now.

I remember there being mentioned that there was a major infrastructure overhaul last week, and that was the reason there was no playtest. I’m wondering if the pre-hit-detection-clean-up type of leading was not actually leading our targets, but compensating for latency?

The infrastructure is actually what I find most impressive about AoA. I remember in a stream or documentary, it was mentioned that stock market caliber infrastructure is used to produce the smooth and lag free environment. Would it be possible to shed some light on the benefits and challenges of using such a structure?

They will be in the upcoming engine; not this week but soon, we still need some time to finish the changes; but getting the controls and weapon/dogfighting right with a single ship is very important.

There’s a Intel/KillScreen: How Age of Ascent Aims to Become the Biggest MMO Ever article about it and I believe there is a more in depth one coming; so stand by…

Not compensating for latency as such; we were just avoiding the need for collision detection further down the player-interface line (ie event occurrences beyond the point-of-fire).

However, we’ve always known of the need to put in proper collision detection into the server-side interest-management/physics layer, and so we took the extra week out to get that process started (initially with high frequency items like bullets, and subsequently with higher latency items such as ship-to-ship collisions).

So there’s no doubt it will put a small amount of extra load onto the physics processing infrastructure, but we were handling that anyway in terms of the comms layers (ie even with an instant “beam-style” laser… when you press fire, there’s a round trip from you to the servers and into the interest management layer, decision on whether it’s a hit or not, and then back to the target to let them know they’ve been hit; we’re now delaying some of those decision-making moments to ensure that the bullets intersect in time and space with a target).

So it’s not a comms latency issue at all, and we can always further distribute CPU cycles across multiple boxes in a scale unit.

I think - of all the weapon types - it’s going to be seeking-missiles and drones (with their own AI decision trees) that will probably eat the most CPU cycles. We’re not concerned about that yet, though. To explain - when we spooled the system up to 50K concurrent players in the same battlefield a few months back (with 8 position updates and 4 player actions - such as firing - per second), we were handling some 267 million (useful) application messages per second (ie ship movements, bullets being fired, interest management etc). And that’s not a hard limit by any means: the architecture means that we can just put more boxes into the mix to handle the (distributed) load.

If that explains a little more?



Turbo no longer locks turn controls; we will revisit this when we have “turbo-type” choices.

Forwarded to wrong score date, scores are here:

Again totally awesome. Just a couple of niggles.
The team kill count stopped updating about ten minutes in. Blue team were stuck on 302 kills
The leader board didn’t load properly at the end. I got the headers and side bit, but the table didn’t load.

I like having to lead the target, even though it meant i went from around 6 kill streaks last time out to 1 this time.

It forwarded to the wrong date, you should be able to see it here? ->

A little above my head but I think I get the basic gist. Does this mean that slower projectiles will present more of an issue than faster projectiles as far as communication goes?

Anyway, had tons of fun with the playtest today. I love how slippery the controls get when you boost. Really makes you feel like you’re along for the ride. I noticed that it sometimes turned into spaceship soup in the middle, and I think ship diversity would help that. I found that if I boosted out and back in, I could get a really good bead on someone but I had an extremely difficult time securing the kill myself. Especially when GM Cerberus would line up right behind my target and block all my shots. Go team effort.

I don’t have much else to say. I feel that the current balance is really beginning to scratch an itch I’ve had with space combat. I can’t wait to get my hands on a faster ship with a more impressive strafe ability. It was really humorous drawing a bead on someone and watching the bullets fly past my head as I dodged them using my ship’s strafing abilities. It also added a much more interesting element to staying on someone’s tail. I remember reading strafing not allowing the ship to exceed its set velocity, so by strafing in the right direction you can sort of slow yourself down without reducing your turn speed. By creating extra distance away from your target using strafing, you can maintain slightly higher speed while covering less distance towards your target. I believe a similar tactic is used by nascar racers when they are being lead around the track because of debris or something.

Speaking of turn speed, I feel its relationship to engine output is spot on awesome. I found myself slowing down a lot to get a more accurate bead, and I definitely felt the effects of the turn speed reduction. I’m absolutely positive that someone somewhere is going to find this really inconvenient and kind of distasteful, but I’m sure there could be some work around for their play style through a different ship loadout or engine type or something. But for the default ship, I think it’s great.

I’m really liking how maneuverability is absolutely vital for survival. Hitting the boost and doing some crazy loops makes you virtually impossible to draw a bead on. This is often when I’d start looking at the map and see who’s shooting me, and engage them in a crazy dogfight. I remember one instance where Yihurt ducked into a large ship thinking he was good under cover. I bet he doesn’t use the gravidar map like I do. But I felt I could play very defensive when I knew I was getting pressured. No more helpless feeling.

Really excited to see what work on the new engine pumps out. New weapons? New ships? When collision is a thing, that whole soup thin in the middle will turn into a deathfest.

Video will be up tomorrow likely. Taking a while to render.

Yeah, we agree that it was a definite step in the right direction for making close quarter dogfighting more exciting, more involved, and also reducing death. You have to consider that - when this is a full on MMO game in the future - death will have consequences (in terms the economic loss of a ship), so this is an interim step towards bringing that in. Player “time-to-death” is a key MMO metric, and I think we’re heading in the right direction!

Yeah, though I wonder if a bit much. At times the rendering engine (admittedly on my test device - a 3 yr-old single processor $400 laptop with a default intel gfx card) seemed to show ships manuvering under boost as closer to the fx I’d be looking for in a top-end guided missile :smiley:

Some adjustments needed there, I think; but boost definitely is now a valid strategy for getting out of trouble.

We had some issues regarding the “hitbox” when relative angular momentum is very low. ie, when a ship is coming directly at you (or you’re right behind it) the “current” hitbox is being partially calculated as a leading target area - which needs to scale better with closer distance than it did.

Yeah, I agree that for a default “fast scout”-type ship the handling is definitely in the right ballpark. It felt fun, fast and loose, and with some valid defensive strategies. I also think the introduction of missile and AoE damage weapons will very much help with the fish-ball cluster-fest!

I experienced some issues with ship identification that we need to work on (sometimes no name appeared above a ship onscreen, and I was once told that I had been killed by myself!). We’ll look into these and get them fixed.

Yeah, we may need to readjust the kill streak medals downwards a touch!

I think with the concepts of target lock (and cycling through pre-selected targets) will allow leading fire indicators to indicate where you should be firing to hit a particular target (as well as arrows off-screen, showing you where to fly to get to your locked targets, and a highlight on the gravidar). This should help with leading accurately, as well as helping us to debug any hitbox issues.

We’re taking another 2-week gap before the next playtest, partly because we could really use the time to really nail this ship-one-on-one down, but also because it’s 4th July weekend next week. Please do keep any thoughts and/or comments coming though.

I’m going to write a devblog in the coming week about our thoughts for alternate ship hulls and module fittings (alternate weapon types, defenses, engine upgrades etc), so please keep an eye out for that, as all your feedback is really valued.

But, all-in-all, a real step in the right direction for dogfighting balance. gj Illy team!



That’s funny you mention that. I definitely noticed a thing where I would strafe to get an angle on the top side or belly of my enemy. I thought this was because it was better because it was a larger silhouette of my target and I would have an easier time hitting them. Really, it makes sense. Smaller silhouette, harder to hit.

There were instances where you could be firing right into the center of someone’s ship (firing directly from behind) and seeing the bullets actually pass through the model, not making hits… and that’s not right.

However, I do (completely) agree with target size presentation (silhouette) as in underside vs topside etc. Better piloting still needs to be rewarded, full-stop. And I was very pleased to see generally the familiar faces on the top of today’s leaderboards. Whilst it’s really early days with the newer systems ofc… it’s a fairly good indication that we haven’t actually nerfed piloting skill!

I think we’re talking about a fairly narrow “up the tailpipe”/“point-blank-in-the-face” extension of the default hitbox / angular momentum calculations, rather than making it generally a larger hitbox.

I think it’ll become much, much clearer when we start putting much larger ship models in. Imagine having the same issue with one of those carriers?! When the models scale, hitboxes become more of an overlapping mosaic of hit-“zones” and “regions of impact”. And for the larger sub-capital and capital ships… when we start introducing modular shield arrangements (“All power to front shields, Scotty!”) then the player angle of attack and damage done goes well beyond a single “hit or not” choice.



Video here!

cf your video from yesterday (tyvm)

at around 7m25 through 7m35 ish. No health on target (and no name above target either), but definitely should have gone down.

Sorry for not embedding it right.

And I see what you mean now. I assumed there was some lag going on, so I switched targets. I think the same thing might have happened to me at around the 10 minute mark.

Watching some of the video again now, I every now and then came across the issue where I would lose control when my mouse left the play area. One way to circumvent the issue is to use fullscreen (f11), but then you have that popup near the top of the screen.