Youtuber Kornelius Briedis seems to have had it with Fdevs

Yeah, I don;t buy that. We already have NPC wingmen (in SLF) and AI cops that will help you. So, that's fine. We also have NPC's that will follow you in SC (and constantly interdict you) and follow you between systems.
This "it's too complex" doesn't hold me any water IMHO. It's just a lack of will as far as I'm concerned.

Those NPC's don't so much "follow you" as much as respawn where you are, they aren't in any way persistent & all you have to do to prove this is look at their weapon load out & (now) ship name. Both change, but the hunting NPC's name stays the same plus it always respawns at 100% health as it is a brand new enemy...
 
Those NPC's don't so much "follow you" as much as respawn where you are, they aren't in any way persistent & all you have to do to prove this is look at their weapon load out & (now) ship name. Both change, but the hunting NPC's name stays the same plus it always respawns at 100% health as it is a brand new enemy...

Ok, so just make the system that spawns the ships name an load-out the same system that keeps the name persistent, and job done.
 
all good points.... and all points which were allegedly being covered back in 2013 when the game was being made.

Stuff like how it fits in with the codebase is not really our problem. npc wingmates and crew were core parts of the game and for me, FAR more important when it came for me deciding to help fund the game, than the multiplayer which was always of 2ndry importance to me personally (which is not to say FD should have dropped it)


but imagine if we turned it on its head, and that after FD talked about ED being multiplayer, it was 2 years after npc wings had been launched and players could still not wing together.... this is kind of how npc wing mates, or lack of, affect me... esp when the ai have no such trouble getting in wings of 8!.

Stuff like how it fits with the current codebase may not be your problem but it certainly is for the person who has to develop any new feature. In fact none of the list is our problem as players but I wrote that down to give you an idea of the complexity of the task in answer to your question.

There are generally three stages of specification. Overall spec that gives that what would be liked and is a very hand-waving document. Then there's the next level down which attempts to list and answer the type of questions that I posted, finally there is a detailed specification that attempts to define the development in such small detail that the developer/coder can do their job.

I would expect that what was covered back in 2103 was the overall, broad specification, not one of the other two which you only start working on when the time comes to consider developing the feature. On very complex systems the requirements gathering can take months if not years and there are qualified engineers that do noting else but this.

Personally, I consider NPC crew to be an easier task than atmospheric planets and landings and both are just a tad difficult.

As for funding, I just wanted to play Elite: Dangerous as a Solo player, the rest was not relevant to me.
 
Okay, take one of those, the NPC AI, to be able to have an NPC Crew that is going to be a member of the crew you need to pretty much understand all the possible inputs and define an action for them. This could either be a set of hard rules, a number of broad rules that cover a range of behaviour, a neural net that learns as it plays, or some combination of all of these with some random fluctuation built in. There may be more possibilities here but these are the ones I can think of off the top of my head. So which of these is the better route?

You have to realise that this also needs to be developed in the context of Elite Dangerous as it is today, not some other game of 20+ years ago.

Other questions and in no particular order:

What are all the possible behaviours that the AI is to exhibit?
What are the limitations of these behaviours? You don't want the AI to have instantaneous reflexes for example, they need to be approximately human.
What can go wrong and what is allowed to go wrong?
What about damage to the NPC, how is that going to be modelled? Is it going to be modelled at all?
What are the differing roles required for the AI NPC and how are these different?
Across the differing roles what are the common behaviours?
For each role what are the unique behaviours?
What are the required skill levels? We have Harmless through to Elite are the NPCs gong to have the same and if so what does that mean in terms of the behaviours? Faster reflexes? Better coordination? and so on.
How are the skill levels gong to affect the common and unique behaviours for all roles.
Are there likely to be new roles required in future?
How can all of the above be parameterised so that changes to roles and behaviours as well as new roles and behaviours require minimal new code, preferably none.
How does all of this fit with the current game engine?
How does all of this fit with the existing code?
What are the likely choke point in the code?
What possible optimisations can be made to speed up the code? Probably going to need to wait until the code is written before attempting to answer this one.
How is this going to be developed?
How is this to be tested? That is alpha testing and not beta testing.
What is the likely timescale for this?
Is just one role developed to start with and others added or should a generic crew member be created first and then specific roles added?
How many developers will be required to develop this?
Is there anything already out there that will help with all of this?
Are specialised developers needed to help with this and should they be contractors or permanent staff?
How are the new crew NPCs going to be recruited?
Are NPC crew going to be rewarded for their part in the game and if so how.
Are the NPC crews to be persistent or transitory?
What roles and behaviours have the player base been asking for?
Are the requested roles and behaviours reasonable?
How does all this fit with the vision of the current game?
Should this development be extended to NPC with space legs in the future? If so, what are the likely roles? Bartender, mechanic, accountant, mission giver, tourist etc.
What are the testing scenarios?
How are QA going to be involved and at what point?
How many QA staff will be required to work not this once it hits alpha testing?
How are the NPCs to be rendered?
Can the Holo-me suite be used to render the NPCs? If so what parameters are required for each NPC role?


The list goes on.

These are just a few of the questions that need to be answered before development can even begin that I can think of off the top of my head. If I were to sit down and approach this as a proper development task, bearing in mind that I write complex business software for a living and have been in the software business for over 45 years, so I like to think that I know a little about the subject, I would expect this basic specification to run to at least 50 pages of A4. The task would require one project Manager and 3-5 developers and probably around 6-12 months to get a playable model in the lab, with public beta testing in 12-18 months.

I could be way off with those last, as I said my experience is in business application not game development but I'm probably not that far off.

I wouldn't even begin to be able to look at the costs or whether or not this was even a justifiable project, I leave that to other experts to answer those questions.

And we haven't even begun writing code yet!

So, does that give you a better idea of the complexity of the task?

An excellent post Mark, rep to you.

Game developing is hard work. I once attempted to write AI scripts for employees in a “game” of mine where the employees work at their machines but move job carts around, while also putting some random behaviors in there like going to the bathroom or getting a drink. It’s amazing how what seem to be simple things in video games are huge undertakings when coding and are far more complicated than most gamers ever realize. And yeah, the planning phase of game dev can get very in depth. I have GDD’s (Game Design Documents) that are dozens of pages and still incomplete.

But then, I’m just an amateur hobbyist game dev. Frontier are the professionals. :cool:
 
Last edited:
Elite has always been a niche title... there is nothing what so ever wrong with a game not appealing to everyone

I agree with that sentiment 100%. The problem is that while ED might be a niche title, it's also a title that's online only and the players ability to play is dependant on FD maintaining the server infrastructure for the game. That costs money. That means the game needs to keep bringing money in just to break even. That means it's entirely possible for the game to go "into the red" and for FD to decide that it's time to turn the servers off, at which point nobody can play.

With that in mind, being a niche title doesn't seem like such a positive attribute.
 
Okay, take one of those, the NPC AI, to be able to have an NPC Crew that is going to be a member of the crew you need to pretty much understand all the possible inputs and define an action for them.

I don't actually think this needs to be as difficult as it's being made to seem. Our NPC fighter pilots already do the things NPC crew would need to do.
Let's look at the individual roles:

Fighter Pilot - already covered - done.

Gunner - Fighter pilot AI already knows how to operate a ship and aim and shoot, so this role would only need to be "lightened", as no power management would be required. They would need only simple orders:
1. Aggressive - fire on any valid target in range (Fire at Will)
2. Focused - fire only on the same target as the pilot (Fire on my target)
3. Defensive - Fire only on targets firing on mother ship.
4. Stand By - Don't fire at all

Engineer - purely a power-management role, which again, the NPC fighter pilots already do. A simplified order set would really be all that is needed.
1. Aggressive - prioritize power to weapons (maintain full power to weapons as long as other capacitors are not drained, shift power when other capacitors are less than 10%)
2. Defensive - prioritize power to Systems. (maintain full power to Systems as long as other capacitors are not drained, shift power when other capacitors are less than 10%)
3. Evasive - Prioritize power to Engines. (maintain full power to Engines as long as other capacitors are not drained, shift power when other capacitors are less than 10%)
4. Balanced - Prioritize balanced power. (maintain balanced power as long as other capacitors are not drained, shift power when other capacitors are less than 10%)

They would need only to shift power around when necessary, and then would do so based on the priority orders given.

Science Officer - Not really a current role, but would function much like a Gunner, only operating ADS and DSS

Commando - No idea what this role might entail, as we're stuck in our ships and paralyzed from the waist down.

But I do think these roles could be readily adapted from the current NPC fighter pilot AI routines.

As for how to recruit them...

Crew Lounge would be a good start. Rescued from Occupied Escape Pods would be a great addition.
Recruited as part of a mission or mission chain would also be a great addition.

While you do raise a number of valid questions, most of these are already answered by the simple presence of NPC fighter pilots - as for how this fits with the vision of the game - mechanical crew and roles were something initially pitched, with a fair amount of detail given about their operation, as well as biological crew. It's rather old-hat, and just has not yet been implemented.
 
Which has more to do with ED being for a niche crowd and those games being for the masses.

Like ED could be a micro brew beer that is good for people who like that style of beer. Those games are like Bud Light.

That's no excuse. There's nothing niche about it any more than WoW or any other. They delivered. That's the difference.

Niche, the idea we fly "complex" spaceships with pips here & there? Look at the UI during a wow raid.

The only thing making it niche is its blandness and lack of features combined with the idea of "potential" driving into compulsive purchase.

All these games have an identity, where-as Elite is still completely undefined, unpredictable and already has a terrible track record.

The only thing defining us so far in Elite is our Spaceship containing a stick-figure glued to the chair which we just about learned to project into some fictive body. Because otherwise it wouldn't make sense right?

Whether you take your character on an adventure into the woods fighting some imaginative monster or picking up your space ship to fight a pirate the core is the same.
When one offers you an interactive story around that monster and some loot leading you into more interesting events and the other still just is about a no-name pirate. Don't blame "niche", blame quality.
 
I don't actually think this needs to be as difficult as it's being made to seem. Our NPC fighter pilots already do the things NPC crew would need to do.
Let's look at the individual roles:

Fighter Pilot - already covered - done.

Gunner - Fighter pilot AI already knows how to operate a ship and aim and shoot, so this role would only need to be "lightened", as no power management would be required. They would need only simple orders:
1. Aggressive - fire on any valid target in range (Fire at Will)
2. Focused - fire only on the same target as the pilot (Fire on my target)
3. Defensive - Fire only on targets firing on mother ship.
4. Stand By - Don't fire at all

Engineer - purely a power-management role, which again, the NPC fighter pilots already do. A simplified order set would really be all that is needed.
1. Aggressive - prioritize power to weapons (maintain full power to weapons as long as other capacitors are not drained, shift power when other capacitors are less than 10%)
2. Defensive - prioritize power to Systems. (maintain full power to Systems as long as other capacitors are not drained, shift power when other capacitors are less than 10%)
3. Evasive - Prioritize power to Engines. (maintain full power to Engines as long as other capacitors are not drained, shift power when other capacitors are less than 10%)
4. Balanced - Prioritize balanced power. (maintain balanced power as long as other capacitors are not drained, shift power when other capacitors are less than 10%)

They would need only to shift power around when necessary, and then would do so based on the priority orders given.

Science Officer - Not really a current role, but would function much like a Gunner, only operating ADS and DSS

Commando - No idea what this role might entail, as we're stuck in our ships and paralyzed from the waist down.

But I do think these roles could be readily adapted from the current NPC fighter pilot AI routines.

As for how to recruit them...

Crew Lounge would be a good start. Rescued from Occupied Escape Pods would be a great addition.
Recruited as part of a mission or mission chain would also be a great addition.

While you do raise a number of valid questions, most of these are already answered by the simple presence of NPC fighter pilots - as for how this fits with the vision of the game - mechanical crew and roles were something initially pitched, with a fair amount of detail given about their operation, as well as biological crew. It's rather old-hat, and just has not yet been implemented.

Possibly but that just adds a whole load more questions on top of the questions I've already listed. Such as:

Is what we thought up several years ago still valid?
Should the document form the basis of the new development or should it be archived and started again? Things will have moved on in the intervening years and it may not longer be valid.
Can existing code be re-used?
If not can it be enhanced to be reusable? What effect will this have on the correct use of that code?
What changes to the testing need to be made?
What changes to the QA scripts needs to be changed

and so on.

Just because you have possible stuff that can be re-used does not stop you from having to answer the questions in the list I posted. You still need to design the feature before you figure out how you are going to implement it and re-using existing functionality and code is an implementation detail. You would also probably revisit the design document for revisions in light of things found during the implementation design phase.

You do not consider existing features or code during the design phase, or try not to, so that you do not limit your design by existing solutions. Just because you implemented the functionality one way previously does not automatically mean that it is the best way to implant it now.

So yes, there may be features and code that can be re-used but you do not, or should not, take these into consideration during the design phase.

There is also the possibility of the reverse being true. You design and implement the NPC crew and then decide that the existing NPC AI can be replaced by the new functionality if it is modified a little. So more questions, design and development.

It is not a trivial piece of development at all.
 
Last edited:
I agree with that sentiment 100%. The problem is that while ED might be a niche title, it's also a title that's online only and the players ability to play is dependant on FD maintaining the server infrastructure for the game. That costs money. That means the game needs to keep bringing money in just to break even. That means it's entirely possible for the game to go "into the red" and for FD to decide that it's time to turn the servers off, at which point nobody can play.

With that in mind, being a niche title doesn't seem like such a positive attribute.

it all depends on the size of the niche tho and how loyal those players are. I have no idea how many players makes ED a success or a failure in FDs eyes.


if we say on average the average player paid £20 for the base game....
and 50% of the playerbase bought horizons at an average of £15

that would be a cig packet income of £27.50 per game sold.

Ed has sold...................... what, in the region of 2 million copies? of which its mosly all digital sales so that is 55 million quid.

some of that will be a steam oculus cut, so maybe take 15 mil from that, which leaves 40 million

no idea how much they pull in on cosmetics, but given its on their engine, and they self publish, with no licencing fees or celeb voice overs taking a "Gary oldman" type cut they have less of a middle man taking a cut.

the game uses P2P which cuts down costs, and also didn't FD develop a game exclusively for amazon to get a good deal on their cloud services?

obviopusly these are all woeful guesses, but, my point is, ED may not need 15 million sales to make enough money to keep on trucking.;

I'm not using "ED is a niche game" as an excuse quite the opposite in fact. I think many of the changes FD have made have been with a view of attracting the main stream gamers, and it is to the detriment of the original vision.

examples for me of this are
docking has been dumbed down
ships maintenance and running costs are now a joke
money earning has been insanely boosted to a point which logically makes no sense at all
CQC -
and then the big one, engineers with its magic lasers of healing, and all the other nonsence which is right out of the generic handbook of augmented weapons with runes of secondary effect.

whilst I did not expect DCS level of flightsim detail of micromanagement... I DID exepect a plausible game which at least at a passing glance had an internal logical consistency.
 
Last edited:
In your opinion.

Play the game, or don't. The zone between those two extremes is reserved for the masochistic, the whiny and arm-chair developers.

Hmm, why do you pretend that "the whiny" play the game and complain at the same time? Not all people who complain play the game, I'm one of those.
 
whilst I did not expect DCS level of flightsim detail of micromanagement... I DID exepect a plausible game which at least at a passing glance had an internal logical consistency.

I don't think anybody on this forum doesn't want that, do they? I know I certainly do and I've seen others crying out for logical consistency throughout the game. ED is analogous to the Star Wars prequels. It's got bits of everything shoehorned into it in an effort to appeal to everybody the cost of that being that they've strayed from what made Elite/Frontier what it was, which leaves some fans of the originals (like myself) not quite knowing what to make of it and feeling hugely disappointed with what could have been. I'm trying to speak for anybody other than myself here, this is just an observation.
 
FDev recently had Obsidian Ant do some PR for them (which was seen for what it was) whereby it was claimed 100 staff or more were working on the game, but obviously no numbers where given of how much of that was the imminent PS4 release (probably most) and how much was the current or upcoming PC content.

It's also noted that even though he was under a NDA, he wasn't exactly shouting from the rooftops "HEY GUYS - YOU HAVE NO IDEA WHAT'S COMING BUT YOU'RE GONNA LOVE IT!!!!! " he sure didn't sound that enthused to me.

Come on, it's Obsidian Ant. He could be telling you he'd just won the lottery and Scarlett Johansson was popping round with the cheque and wicked intentions and he'd sound like he was dozing in his armchair on Sunday afternoon. :D
 
Okay, take one of those, the NPC AI, to be able to have an NPC Crew that is going to be a member of the crew you need to pretty much understand all the possible inputs and define an action for them. This could either be a set of hard rules, a number of broad rules that cover a range of behaviour, a neural net that learns as it plays, or some combination of all of these with some random fluctuation built in. There may be more possibilities here but these are the ones I can think of off the top of my head. So which of these is the better route?

You have to realise that this also needs to be developed in the context of Elite Dangerous as it is today, not some other game of 20+ years ago.

Other questions and in no particular order:

What are all the possible behaviours that the AI is to exhibit?
What are the limitations of these behaviours? You don't want the AI to have instantaneous reflexes for example, they need to be approximately human.
What can go wrong and what is allowed to go wrong?
What about damage to the NPC, how is that going to be modelled? Is it going to be modelled at all?
What are the differing roles required for the AI NPC and how are these different?
Across the differing roles what are the common behaviours?
For each role what are the unique behaviours?
What are the required skill levels? We have Harmless through to Elite are the NPCs gong to have the same and if so what does that mean in terms of the behaviours? Faster reflexes? Better coordination? and so on.
How are the skill levels gong to affect the common and unique behaviours for all roles.
Are there likely to be new roles required in future?
How can all of the above be parameterised so that changes to roles and behaviours as well as new roles and behaviours require minimal new code, preferably none.
How does all of this fit with the current game engine?
How does all of this fit with the existing code?
What are the likely choke point in the code?
What possible optimisations can be made to speed up the code? Probably going to need to wait until the code is written before attempting to answer this one.
How is this going to be developed?
How is this to be tested? That is alpha testing and not beta testing.
What is the likely timescale for this?
Is just one role developed to start with and others added or should a generic crew member be created first and then specific roles added?
How many developers will be required to develop this?
Is there anything already out there that will help with all of this?
Are specialised developers needed to help with this and should they be contractors or permanent staff?
How are the new crew NPCs going to be recruited?
Are NPC crew going to be rewarded for their part in the game and if so how.
Are the NPC crews to be persistent or transitory?
What roles and behaviours have the player base been asking for?
Are the requested roles and behaviours reasonable?
How does all this fit with the vision of the current game?
Should this development be extended to NPC with space legs in the future? If so, what are the likely roles? Bartender, mechanic, accountant, mission giver, tourist etc.
What are the testing scenarios?
How are QA going to be involved and at what point?
How many QA staff will be required to work not this once it hits alpha testing?
How are the NPCs to be rendered?
Can the Holo-me suite be used to render the NPCs? If so what parameters are required for each NPC role?


The list goes on.

These are just a few of the questions that need to be answered before development can even begin that I can think of off the top of my head. If I were to sit down and approach this as a proper development task, bearing in mind that I write complex business software for a living and have been in the software business for over 45 years, so I like to think that I know a little about the subject, I would expect this basic specification to run to at least 50 pages of A4. The task would require one project Manager and 3-5 developers and probably around 6-12 months to get a playable model in the lab, with public beta testing in 12-18 months.

I could be way off with those last, as I said my experience is in business application not game development but I'm probably not that far off.

I wouldn't even begin to be able to look at the costs or whether or not this was even a justifiable project, I leave that to other experts to answer those questions.

And we haven't even begun writing code yet!

So, does that give you a better idea of the complexity of the task?

What coffee shall the Devs drink? What time should they take their lunch? What's the meaning of life? What Party should the devs vote for? Why is it so hot?

Sorry, mate, but a lot of the questions you ask are either already dealt with (by adding misc NPCs / NPCrew into the game) or are just generic problems of building any addition to the game.
 
it all depends on the size of the niche tho and how loyal those players are. I have no idea how many players makes ED a success or a failure in FDs eyes.


if we say on average the average player paid £20 for the base game....
and 50% of the playerbase bought horizons at an average of £15

that would be a cig packet income of £27.50 per game sold.

Ed has sold...................... what, in the region of 2 million copies? of which its mosly all digital sales so that is 55 million quid.

some of that will be a steam oculus cut, so maybe take 15 mil from that, which leaves 40 million

no idea how much they pull in on cosmetics, but given its on their engine, and they self publish, with no licencing fees or celeb voice overs taking a "Gary oldman" type cut they have less of a middle man taking a cut.

the game uses P2P which cuts down costs, and also didn't FD develop a game exclusively for amazon to get a good deal on their cloud services?

obviopusly these are all woeful guesses, but, my point is, ED may not need 15 million sales to make enough money to keep on trucking.;

I'm not using "ED is a niche game" as an excuse quite the opposite in fact. I think many of the changes FD have made have been with a view of attracting the main stream gamers, and it is to the detriment of the original vision.

examples for me of this are
docking has been dumbed down
ships maintenance and running costs are now a joke
money earning has been insanely boosted to a point which logically makes no sense at all
CQC -
and then the big one, engineers with its magic lasers of healing, and all the other nonsence which is right out of the generic handbook of augmented weapons with runes of secondary effect.

whilst I did not expect DCS level of flightsim detail of micromanagement... I DID exepect a plausible game which at least at a passing glance had an internal logical consistency.

Of cause sales and players are two different things?

I suspect one ideal model would be lots of sales but as few actual players as possible :)

A sale = +£, but a player is probably = -£ due to them using servers and reporting bugs etc.

There's got to be a business model out there to do the best marketing ever with the weakest content ever to back it up? :)

ps: Let's ignore paint jobs and the like for my silly example.
 
Last edited:
Play the game, or don't. The zone between those two extremes is reserved for the masochistic, the whiny and arm-chair developers.

hand on heart do you honestly take your own medicine on this stance on all things.... or is it just elite dangerous?

Despite liking the game, warts and all, I feel given the amount i have paid to help get the game past its KSer goals, I have a right to comment on where i think the game is coming up short based on the sales pitches made by FD.

Of course if you do not like to read negative posts, no one is forcing you to read them... Even people i feel are way off base on their complaints still have the right to voice them, it is after all a discussion forum.
 
There's got to be a business model out there to do the best marketing ever with the weakest content ever to back it up? :)

you mean one which sells packs of space ships for $15,000???

I can think of 1... and I bought into that one too :D

(there was a time i would have put No mans sky into that bracket but to be fair it has has a lot of content added since launch........... I dare say if NMS had launched as it is now it would not have been panned as much as it was)
 
Last edited:
you mean one which sells packs of space ships for $15,000???

I can think of 1... and I bought into that one too :D

(there was a time i would have put No mans sky into that bracket but to be fair it has has a lot of content added since launch........... I dare say if NMS had launched as it is now it would not have been panned as much as it was)

LOL! Perfect :)

Yeh, buying a jpg image for $15,000! :)
 
Last edited:
Back
Top Bottom