Focused Feedback: The Engineers

On the 30th November Sandro Sammarco outlined some likely but not confirmed changes to the way Engineering works in the game. These can be seen below and in the Beyond: Core Gameplay Improvements thread. A feedback thread on the forum can be found here.


Issues with the current system

“There are a few of areas that we’d like to improve, in no small way influenced by player feedback”

  • Potential for Failure – Whilst it’s cool to have pros and cons in an upgrade system, the statistical level of variety in many blueprints ends up meaning that there is a chance of an upgrade being wrecked entirely, and basically not an upgrade, which undermines the whole process
  • Too much Random – There are multiple points where the player is left at the mercy of chance. Whilst some amount of randomness is not a bad thing, and can lead to interesting variety, in our crafting process we have three: the likelihood of correct materials being generated during associated activities, the range of known statistical options, and a final hidden set of statistical options. It can been argued that this is simply too much and makes everything unnecessarily complicated
  • Too Long – Elite Dangerous is a game where scale is very important, and very cool. It also means that processes tend to take longer. This includes the engineering loop. The grand scale acts to exacerbate when the results aren’t great and even when they are it can be argued that it’s just too slow a process.

Suggested Improvements to Engineering

“With this in mind, we’re looking to overhaul the system to try to achieve the following goals”

  • Guaranteed improvement – We want to make sure that when you upgrade a module the end result is always better. There’s a significant time/resource investment in engineering so we want to make sure that you feel it’s worthwhile before you even begin, so you can make informed decisions on whether to take part or not.
  • Increase efficiency – There’s always going to be a significant time cost to upgrading your ship, but we want to look at ways of sometimes mitigating where we think it’s appropriate.

These are the changes that we’re currently investigating.

  • We are removing all hidden statistical variables from the upgrade process. Whilst there is still a range of success when you craft, you will be able to see the range before you commit
  • One cool thing that secondary hidden statistical pros and cons did was ensure variety. Because we are losing this, we will try to ensure that each module has a set of experimental effects that allow you to tweak your module in a variety of ways, hopefully ensuring that there are a number of different options to aim for
  • We are removing any potential for the crafting process to result in an upgrade that is worse than what is currently fitted. Every time you pay for an upgrade, an upgrade is what you will get
  • All penalties in blueprints are fixed and only applied once per rank
  • You will need to maximize benefits from an upgrade before you can start applying higher rank versions to a module. We still want to reward the process of upgrading modules
  • You will be able to craft pinned blueprints at any starport that has outfitting. You won’t be able to gain reputation with the Engineer for doing this, and you won’t be able to fit experimental effects at this point, but it can considerably reduce the upgrading crafting loop
  • Experimental effects will no longer have a chance to occur during the upgrade process. Instead, each experimental effect will simply have a materials cost that you can pay to have it fitted to an appropriate module
  • This means that there is no way you can lose reputation ranks with an Engineer
  • We will institute a materials trader at specific starports. These contacts will allow you to trade (at loss) materials within the same class, allowing you to convert unwanted materials into useful ones
  • We will introduce a per material storage cap, probably around the 100 mark, to remove inventory shuffling
  • We’ll add quality and quantity indicators for materials in the game world, so you’ll be able to see at a glance how much you have of a material without having to check your inventory
  • We’ll add an ignore function for materials (and commodities, incidentally), allowing you to mine and collect more efficiently by preventing collector limpets from picking up ignored items and auto-venting refineries.

Grandfathering

“A lot of you will already have engineered items. We will “grandfather” these. This means you can still use them, their statistics and effects won’t be changed.

“However, if you want to apply further upgrades to them, they will have to be converted. Conversion will place them at the top of the previous rank (so a rank 4 upgrade would become maxed rank 3 upgrade) and would change all statistics and effects to represent the new blueprint.”

“In general, we will try to make sure that the new blueprints can max out slightly better than the old system (we want to encourage conversion).”

Beyond : Focused Feedback – The Engineers

As announced by Ed Lewis on the 30th November, the next subject for feedback will be The Engineers. Updated information will appear on the Beyond: Core Game Improvements thread.

“Hi everyone,

“Lead Designer Sandro Sammarco and I will be streaming tonight. We’ll be saying a quick thank you for the feedback from the last Focused Feedback topic and we’ll be introducing some of the ideas for the new topic: Engineers.

“This livestream is to do with the Beyond Series of updates and how we’re looking for your feedback and help with the direction of some elements of the game.

“We’ll be going live at 7PM GMT.

Link to Live Stream

 

Live stream: Discovery Scanner 2

Frontier will continue their peak behind the making of Elite: Dangerous with ‘Bringing a Thargon Swarm to Life with Mark Allen’

Hey everyone,

Join us here tonight

‘For episode 2 of Discovery Scanner, we’re joined by Lead Programmer Mark Allen, who will talk us through how he brought the Thargoid swarms to life and created their fascinating behaviour patterns.

Discovery Scanner is a new Elite Dangerous livestream series where we take deep dives in to the creation of some of Elite Dangerous’s most exciting features and aspects.’

Enjoy! See you at 7PM GMT

 

 

Live stream – Beyond: Focused Feedback with Sandro Sammarco

Frontier have announced that there will be a live stream tonight, Thursday 16th November 2017. You can also read some of Sandro’s answers to questions raised by the player base, here.

“Hi everyone,

“Tonight I’ll be joined by Sandro Sammarco at 7PM GMT to talk about Crime and Punishment in the Beyond series of updates (2018).

“We’ll be discussing the focused feedback forum – why we’re doing it, how we’re doing it and talking about the feedback we’ve received so far, and how that may influence game development.

“See you at 7PM.”

Live stream – Discovery Scanner 1: Update

Following the live stream, Dr Ant Ross answered some questions. You can see them below, on the official forum here, or on Reddit, here.


1) Is “The Bulge” what is referred to in game as The Bubble?

The galactic bulge, or galactic core, is the blob shaped region of stars in the centre of the Milky Way. It is a separate thing from The Bubble, which is the region of human inhabited space around Sol.

2) Why did Elite Dangerous decide to use light-years and not parsecs in-game?

I don’t believe there was a specific reason why. The original Elite and its sequels also used light years for interstellar distances, and it is nice to be consistent.

3) Does the game allow for any catastrophes? Novae, rogue-planet ejection or capture, system collision?

During the generation stage, rogue-planet events and collisions are taken into account. However, these are historical and not visual effects. When you arrive in the system, you only see the results of those effects on the resulting stable system.

4) When do we get to hear about what is coming next for Stellar Forge? Like comets, flyable gas giants, etc.

Oh, yeah, that’s scheduled for MMPPPPPPH ( Ant gets dragged away by Zac ). Ahem, hm, apologies, but I can’t really talk about that at the moment.

5) Why does Elite Dangerous only use circular orbits? Are elliptical orbits too complicated?

Elite Dangerous does use elliptical orbits. The inclination, periapsis, longitude of ascending node, semi major axis and eccentricity information is all tracked.

6) There are some planets in the game that look like they would be torn apart by tidal forces. Did you say you actually simulate this? What means those close planets are actually possible?

There are some bodies which may be near the mathematical limit of that collection of matter being clumped together. However, we don’t currently have the rendering feature to display what the surfaces of those look like or how you would interact with them. In reality, you may see a distorted molten mess.

7) The in-system barycentre-orbital distribution in the game is quite vast, given the limited observed bodies (just Pluto/Charon). Is that likely true to life?

The distant orbits which are created should follow Kepler’s laws of planetary motion and classical (non-relativistic) orbital calculations. The resulting hierarchies of orbits should be feasible, but it will be very interesting to see what comes from the future of orbital mechanics analyses and exo-planet searches.

8) Is there anything orbiting within Lagrange Points at all in Elite?

There is a section of code running capture interactions in the virtual protoplanetary disk which can place bodies into the L4 and L5 Lagrange points. I believe this ends up being mostly asteroid clusters, but I have not had a look at them in a while.

9) Have you ever forgot the location of a special feature, that could be found by a player?

If I did, I have forgotten what it was. Sorry I did once make the galaxy produce Mr. Braben’s face in stars as a test for being able to inject very bright O class stars, but that was removed pre-pre-pre Alpha.

10) How close is this identifier system to any existing official naming or identifier protocols currently used by scientists? Like, could we input future data we collect in real life into Elite?

Each generated body needs a unique identifier, but also any “authored” body, say from the stellar catalogues. Authored bodies, however, have their information pre-calculated, so only do partial proc-gen at most. The system is sensitive to adding new authored bodies to the system, as it could displace the identifier of other proc-gen bodies, changing the galaxy. It is possible to make changes, but we must be careful about it.

11) Why aren’t there dark regions on the other side of the galaxy?

Do you mean lower density regions between the spiral arms? If you could please forward on an X,Y,Z sector location (as seen from the Galaxy Map grid lines) I could have a closer look and confirm. Depending on where exactly you are on that side, there are still under-dense regions, but there may also be some bleed through from the galactic bulge data placing more matter in that region that immediately appears from the image in the Galaxy Map.

12) Will Nebulas one day resemble the Hubble painted shot?

One of the common discussions on the dev floor is about the level of visual exaggeration we portray to the players. For example, most nebula images you can see are false colour ones of non-visible EM wavelengths. Do we make radiowaves visible in-game to reflect these beautiful photographs? Do we go the realistic route of showing these hydrogen clouds as greyish, slightly greenish diffuse blobs? No promises, no guarantees, nothing to announce at this time, but I am always interested in potential improvements to the rendering of stellar bodies and phenomena.

13) Does the planet surface rendering use adaptive subdivision when making the hills and craters of the planets?

Currently, the subdivisions are in smaller patches with the same amount of points in the same square arrangement. This is used, rather than non-uniform dynamic tessellation, because in our system it allows a much faster and economical generation of both render and physics meshes. We blend between detail levels of patches though, to smoothly transition from low to high vertex density.

14) Have you thought about creating permanent, full textures for popular planets that get loaded frequently?

Such large texture collections would take up a lot of graphics memory. I understand and sympathize with some of the popping that users have been experiencing, but using up that available memory would mean that other textures in the game would pop quite a bit worse.

15) How come some stars are older than the Universe in Elite Dangerous?

The oldest stars in the real Milky Way can tend towards 13,800,000,000 years. The sectors themselves have a maximum average age of 12,810,000,000 years. Unfortunately, though, there is a small, rare, bug in resolving the range of ages across a few sectors, and there have been a few cases of the stated age if a star being around 18 billion years. Yup, that’s an oopsie. Bluntly fixing this will upset some generation and alter the galaxy, so a careful approach is needed for fixing this one up. It’s my list of work to address, apologies for the delay.

16) Does ED generate hidden geometry?

The word “patch” in this answer refers to a square of terrain. For planets? The base patches on the faces of those cubes that become the planets are always there, as the data might be needed at any moment if one is traveling fast enough. The finer detailed patches are only generated when it’s reasonable for the game to expect you to need that information, depending on your altitude and speed. There might be patch information which would be beyond the horizon and so not needed for rendering in this frame. In this case, the patch is culled from the list of things to render, reducing the work on the GPU.

17) Why are there regions of stars near the core where there are defined lines of difference… i.e. one side dense star fieldother side hardly any stars at all?

A good question; I do see the reddit posts about this one fairly frequently. There are some sectors which have a large density gradient between them. One sector is over a threshold for generating stars of the size that its layer looks after. The neighboring sector may be just underneath this threshold. As this neighbor doesn’t create any systems, it donates its mass to all its child sectors and so creates many bodies. The first sector has made some heavier stars, so donates a lesser amount of leftover mass. The result is that there is an overpopulated sector and an underpopulated sector. Not all cases of this happening were smoothed out before the initial release of Elite: Dangerous, and making any alterations now would upset the proc-gen of the entire Galaxy.

We are aware of it and are still mulling over which direction to take this in.

18) Do you have anything to say on the real life discovery of Trappist-1, with it being so close to the system which the Forge created?

“Well done StellarForge!”? Also, “Well done astrophysics and astronomy researchers! You rock!”

19) I’d love to know how you integrate the catalogued star systems (stuff we see from earth)!

We created a compiler that analyzed the contents of the Hipparcos and Gleise stellar catalogs of the time and created stellar forge system entries from the details. Some entries had the full details of the star, others only the brightness, colour and direction of the star, and others somewhere in between. This compiler would take the available data and fill in the gaps using the same calculations that would be used for proc-gen systems. It would also assign BodyAddress identifiers based on the sector they would be found in. When creating a sector in StellarForge, these “authored” systems are loaded in first, their pre-compiled data taking place of generated results. The authored system also take away the mass designated to the sector, like a generated system would.

The proc-gen systems are created later with the leftover mass. The game then treats the authored and the proc-gen systems identically. As for the contents of the star systems, we have other game data resources which Mike Brookes organised which link to a system for full or partially pre-made system layouts.

20) Will we be getting those high mountains and deep ravines back at some point?

The crazy ravines and mountains (mostly on “Rocky Ice” planets) were quite striking, yes. Unfortunately, they were also causing issues in the creation and handling of the physics mesh that you would drive on (or fly into if distracted). The errors that would be encountered were severe enough to cause a crash (this was seen in public as us pre-emptively kicking you to the main menu on approach to the planet). Removing the crazy examples of ravines/ridges was necessary to:

  • allow the creation of physics patches
  • not be entirely scientifically inaccurate
  • avoid terrain which was sharp enough to look really unpleasantly jagged on VR setting.

There also used to be a few cases of a bug in the noise graph allowing the occasional mountain taller than a planet sustain, taking into account material properties and gravitation. These were just reduced to the maximally supported mountain height for the planet. We want to keep working on the planets and deliver improvements. Official people will say official things when the official time is upon us. For now, the short term work involves revisiting the planet colouration, as per the keynote from FX2017

21) How close was the old Elite 2/3 system generation compared to Stellar Forge? Are there any notable changes compared to David Braben’s original take on it or was he pretty close?

Mr. Braben worked very closely with the team to help ensure the approach was consistent but building upon his systems and that StellarForge covered any additions that were wished for at the time of Frontier and First Encounters. The maths behind the coarse level details of the models of stellar and planetary formation have not changed drastically since the 1990s, and so many calculations would be similar. What has changed is the understanding of things like how common binary pairs are and other statistical particulars.

22) Where does the initial mass component to a galactic sector come from? Real life observation, procedural generation or guess work?

Real life observations give us the approximate mass of the entire galaxy. There are various papers out there, some looking at the orbital velocities of globular clusters in the galactic halo, some trying to count types of stars in a region and extrapolate that across whole arms, some measuring the speed of stars at different orbital radii away from the core. The value we used was taken as an average from available research papers at the time. Observations of our and other spiral galaxies gives us a good idea of the distributions of that mass in 3d space.

23) How does star system naming work?

The unique identifier is transformed into a name. The sector address is encoded into a proc-gen “stellar catalog” name. The sector layer can be encoded in a single letter. Some of the numbers refer to the system within the sector at that layer.

 

Live stream – Discovery Scanner 1

As mentioned on the 3rd November, Frontier will be beginning a new series of live streams. Tonight’s stream can be seen live from 7pm and will essentially be the same talk Dr. Ant Ross gave at the Frontier Expo, but with some more in-game examples and description. You can read more about Dr Ross, here.

“Discovery Scanner is a new Elite Dangerous livestream series where we take deep dives in to the creation of some of Elite Dangerous’s most exciting features and aspects.

“For episode 1 of Discovery Scanner, we’re joined by Programmer Dr Anthony Ross, who will talk us through the creation of the Elite Dangerous galaxy – including how the stellar forge was created. Let’s go behind the scenes like never before.”