The real problem with RNG (and a solution!)

[Spyro Jance]

The real issue with RNG is it hides problems with the game.
After seeing Obsidian Ant's latest video and noting that Frontier have confirmed there is an issue with Unknown Probes spawning, I am left wondering how many other things are broken in the game but have been hidden by RNG?

How many other times has an issue cropped up regarding the spawning of particular mission rewards, USS, materials etc and been excused as RNG?

Having worked in customer support I know many organisations use troubleshooting flowcharts (or programs). I can easily imagine "RNG" would be the end of one of those branches, and provide a handy way to end that investigation (we used to call them "get out of jail free" cards).

What Frontier need to do is introduce a modified RNG which reduces the odds every cycle until after a set number cycles the odds become 1:1.
In this way Frontier can always run the cycle and confirm the results are working as intended.

It also means that players have a guarantee that for any RNG event (engineering, materials, mission rewards) they will (eventually) get the result they want.

This doesn't get rid of RNG, but it manages it in a way that is much more productive and satisfying. Lucky players can still gloat about how quickly they got a particular mod/material/reward, while unlucky player will have the guarantee that they will eventually get the result they want (instead of potentially never).

 

[NW3]

Well said. I'd like to see that implemented. Rep++

 

[SpeakerToAliens]

The players would just complain that they have to try n times before they get the perfect role and why not just let the engineer work their magic perfectly every time?

My solution? No RNG. If someone was "chipping" your car you'd expect to get what you paid for or the CarHacker's rep would fall rapidly and everyone would go elsewhere.

 

[Thunderchief]

My solution? No RNG. If someone was "chipping" your car you'd expect to get what you paid for or the CarHacker's rep would fall rapidly and everyone would go elsewhere.

This, eleventy millions times this!

one of the reasons I have not bothered with anything to do with the engineers is the pointless RNG system. I don't know why its set like this "magic" system, it should be a set of bonuses that are appied positivly and negitivly.

you want a longer jump range, ok fine but it will take longer to spool up.

you want faster thrusters fine but you will pay for it in ware and tear. or use more heat.

and so on.

The magic RNG is just silly.

Simon.

 

[Jack Schitt]

The thing I'll never quite understand is not so much the RNG element, but the decision to gate some "loot drops" (I hate that term but it's universally understood) behind other, largely hidden, variables as well. So there might be a 50% chance of commodity x appearing in a USS, but only if you're already in a system with a population greater than y and a system state of z.

These have mostly been worked out by the community now and are on sites like Inara, but the whole mechanic is very clunky and this information really should be presented in-game. Yes, there is a certain logic to some of the elements once you realise what's going on (ships in Outbreak systems are more likely to drop pharmaceutical isolators) but many of them still seem arbitrary. If it wasn't for this forum and the community at large I don't think I would ever have figured out what the rules are.

 

[Lunchmoney]

[off topic]

If someone was "chipping" your car you'd expect to get what you paid for or the CarHacker's rep would fall rapidly and everyone would go elsewhere.

As a car tuner I dislike your analogy. Two cars that roll of the production line together and get the same mods may end up with different torque and bhp after a remap. Yes, you see an improvement but not the same as the other car.
[on topic]

OP, seems like a good idea.

 

[Shiro Akai]

What OP suggests we can observe in WoW - one of "core" things of current expansion is a "legendary" item quality drop. Everyone aims for those as they provide significant boost to overall performance. But, as always, drops are hidden behind RNG. And while legendaries are highly desirable, RNG may interfere heavily in obtaining them. So Blizzard introduced "bad luck protection" system. Everytime you face a situation in which legendary could drop but it didn't, your chances are bumped a bit. It's still few places after the comma but the number there is bigger. So, the longer you have bad luck, the higher chances that you actually get what you aim for. Very simplified explanation on an already existing example.

But.

Despite overall mechanics existing I don't think "fixing" engineering materials drops, mission rewards (to some extent) is the right way. Current design (a chance for finding material that is used for a chance for a good roll) feels like a RNG within RNG. It's highly punishing and unrewarding. First you spend lots of time looking for materials that (by RNG) just refuse to drop/show. And when you finally get enough of it you go to the engineer for another round of roulette. With a high chances of loosing which makes all your previous farming useless and more frustrating.

To be clear - I'm not against RNG per see. I just don't like when it's tedious, pointless and plain boring. So here's my proposal:

As OP mentioned, let Frontier incorporate some kind of "bad luck protection" for engineers rolls ONLY. And remove RNG from engineering materials. They can up quantities required a bit to compensate.

That way acquiring engineering materials would always be fruitful and it would be just a matter of time. Players would still have to spend time getting those but it'd be less frustrating. And the only lottery would be actual roll.

While I completely agree with SpeakerToAliens and his car tuning example (heh, I once used that example myself) I'm a realist that RNG won't disappear.


--- addendum---

As a car tuner I dislike your analogy. Two cars

And this is the part where you're wrong. While engineering Sidewinder and Anaconda may lead to different tuning results I would expect same results while tuning two Sidewinders. This is what I call quality. If my neighbor who has Ford Focus went to a car tuner and installed one chip and got certain performance boost I'd expect the same performance boost with my Ford Focus. Same car, same tuning, different outcomes is what we would see if Frontier applied their logic to your field.

 

[Agony_Aunt]

I like the idea. Would still mean people have to hunt but there would be an end to frustration when the RNG decides to hate you.

Doesn't help with all RNG hunting, but would help with things like UPs.

I had a thought about things like materials and data as well. Partially its about being a bit more specific in the in-game text of those things and partially just tweaking the spawn rates a bit as well.

For example, take a look at Chemical Manupulators. Description say they can be got from trade ships. Well, actually, they drop fairly regularly from T9s and Belugas, sometimes from T7s and Orcas, and rarely from T6s and Dolphins. So go a little further and add to the description "... most commonly found on T9s and Belugas" to help people focus, and perhaps double check the spawn rates to make sure they are acceptable.

Also, with regards to spawns of things that occur in locations where other things can also spawn. Take for example Pharmaseutical Isolators. High grade emissions in Outbreak systems. Seems simple enough right? But, if that system is a Fed system, then it will be competing on the spawn tables with Core Dynamic Composites. I'm guessing here that the game looks at all that can appear and then randomly selects what should appear from that list, rather than rolling for how many of each. The downside of this is that when you are looking for PIs in Fed (or Imperial) systems, you're reducing your chances of getting them compared with doing the same in an Independent system. But there is no information in game telling you this, and this can lead to frustration. So, either review descriptions to say search in Independent systems (as well as anything else that might block you), but that's messy, or just be a bit more generous with spawns and spawn everything that can spawn in those systems in a certain amount. Yeah, it makes things a bit easier, and people could kill two or three birds with one stone, but at the end of the day, people already have to sit through one level of RNG waiting for the right SS to appear, and how many appear is also a RNG factor, so doesn't hurt allowing people to avoid another level of RNG.

Which of course does point to the main problem with all this. RNG by itself isn't a bad thing. Its the multiple levels of RNG piled on top of each other that causes most of the frustration.

 

[Agony_Aunt]

And this is the part where you're wrong. While engineering Sidewinder and Anaconda may lead to different tuning results I would expect same results while tuning two Sidewinders. This is what I call quality. If my neighbor who has Ford Focus went to a car tuner and installed one chip and got certain performance boost I'd expect the same performance boost with my Ford Focus. Same car, same tuning, different outcomes is what we would see if Frontier applied their logic to your field.

I think i agree with Lunchmoney on this. No two cars are identical to start with. There will always be slight differences. You start chipping, and it probably exaggerates those differences as well. It might be minor, nowhere near the level of Engieners for example, but it probably would be detectable.

Same with integrated circuits like CPUs. They are massively complicated things, and no two CPUs are identical. This is how the manufacturers can sell two CPUs from the same line as different processors, because they do tests after production to determine which mean the required performance for a certain brand and to check for errors. Oh, a problem with part of the cache? Disable it, sell it as a lower spec CPU.

Or two people build identical PCs, one person's PC runs Elite Dangerous with no problems, the other person's crashes whenever a certain event happens. Its rare, but it can happen. People complain on forums about rubbish devs, and community doesn't believe them, because they have no problem with that event. And the problem is, its just the components, sometimes just a particular combination of components, that make it happen, or certain CPU instructions happening in a certain order, that causes the whole issue. Such things are rare, but they happen.

 

[Jmanis]

After seeing Obsidian Ant's latest video and noting that Frontier have confirmed there is an issue with Unknown Probes spawning, I am left wondering how many other things are broken in the game but have been hidden by RNG?

They confirmed it? Linkage?

Anyways, I don't like sliding-chance-windows which converge on guaranteed success.

There's nothing inherently wrong with RNG per se. The problem the Unknown Probe spawns have is that there's nothing to interactively do to verify if you're going to find one anytime, except sit down and wait.

At a very crude example, a barely adequate fix would be to go in range of an ammonia world (where they're meant to spawn), honk a "local discovery" scanner, which reveals a half dozen USS. None of them may be a UP location. Fine, great.. given the procedural nature, drop out of supercruise and back in, do it again.

Maybe you could put the USS that spawn on a 1-hour rotation, so that if you don't find one at a planet at, say, 1315, you won't find any until 1400, when the RNG cycle "ticks" and it randomizes the specific ammonia worlds they spawn at again. That way, if you don't find them at one ammonia world, you're encouraged to go check others out in the hopes of finding a site. Ultimately, the goal is you're not sitting there for literally hours or days on end wondering "Well, is it going to spawn in the next 30 minutes? I have no idea...."

A better system would make USS be a difficult activity where you actively do some sort of minigame to identify and track down particular signals, then finding UPs becomes a matter of skill, and you could still throw an RNG layer on it.

That's the biggest problem with UPs, there's literally no way to work out if you should move on or just sit tight, and nothing to do besides that. Dropping out at Degraded Signal Sources just wastes more time that could be used cycling for a UP spawn. That's bad.

 

[kofeyh]

I think the core issue, really, is the layered approach Frontier took.

There is RNG associated with materials/ data collection, RNG with stats during roll, RNG with secondaries during roll and RNG (with a cost to override) for applicable specials. This would also otherwise include RNG for commodities, if they weren't disabled. This isn't the sort of 'fudge' that can shift legendaries around a little bit in other games; it's up to five layers of individually random elements; currently four. Four. I don't think I've ever seen a game go to just that extreme before. Now, I get why Frontier did this, a) to cause module variation (and thus counters) and b) time sink, but in practice? It's just b) timesink.

Most everyone will naturally gravitate to 'optimal' during ship config. The game essentially rewards that. Hence why multipurpose ships are so popular. They're they most optimal solution (more than actual dedicated ships) but I think there's a specific thing going on here, and frankly one of the two actually disturbs me and yet is the mostly likely reason it's still highly random and arbitrary.

The first, is that this was really just a massive miscalculation, and they just did not realise how really unworkable it is with so many layers of RNG combined with an almost complete lack of in-game referencing. The other? The disturbing bit? Is that it was always intended to be a massive time-sink and it's a very intentional and calculated model designed purely to keep people playing.

Personally? I think it's the latter. They've fought tooth and nail to keep as much of the RNG, with arbitrary locales for drops, as possible. That's not the response of a developer keen to make things a little more workable and consistent if they've just misjudged a game element (at least from a drop rate and locale) - it's the response of a developer who has pegged their very future on layers of RNG that massively distort outcomes, virtually enforcing compliance (and thus time spent).

And this? This is what gives me pause for Thargoids. Because the overwhelming temptation will be adding yet more, multi-layered RNG, on top of what we have. That? I think may alienate a lot more people. And may well end up the proverbial straw. I just.. I don't know. I don't entirely know what to think. Sure, some won't play ball, and refuse to grind. But it's evident to me, that that isn't the majority. Not by a long shot.

But I am concerned. Because I don't think anything will be learned, and 'more of the same' will feature strongly post 2.4, as the thargoid story continues to unfold. More RNG. More of the same.

 

[StiTch]

Nice try...I don't quite think it works like that, though.

For a start we complain about RNG in ED a lot but the stuff that really matters is a tiny section, namely in the RNGineers. Almost every modern game uses the RNG almost constantly. You can't fight it, just make it more sensible.

So you try to implement a rule that asks RNG odds to tend towards 1 as it rolls through. Given that's technically feasible, how many RNG timers do you think you'd be putting out of whack in the time it takes for you to make a single journey?

Perhaps instead of trying to make the random events of the universe not-exactly-very-random you could look at something a little more...elegant.

 

[kofeyh]

Almost every modern game uses the RNG almost constantly. You can't fight it, just make it more sensible.

There is random chance, random stat or random encounter; Frontier have elected to combine multiple RNG elements for engineering, with wildly variable outcomes. It's designed to be a timesink. RNG has always been designed as a timesink. Frontier have, however, taken it to a whole new level. It's clear they've pegged a lot of content on RNG. Arguably, this is to be expected.

Consistency is sorely lacking in ED. I don't think removing all RNG is necessary, but imho the balance is way out. And to me? It's always been "how can we artificially keep people busy" expressed as a mechanic, not "does this enhance the experience" to keep people sucked in and enaged.

 

[StiTch]

There is random chance, random stat or random encounter; Frontier have elected to combine multiple RNG elements for engineering, with wildly variable outcomes. It's designed to be a timesink. RNG has always been designed as a timesink.

You misunderstand me.

RNG checks are made constantly within a game like this. Traveling through SC? Multiple checks made for PoIs. Module damaged? Continual checks for malfunction. Combat? Being used en masse, with several used for a single shot.

Introduce a blanket rule to say that RNG odds always approach 1 after five minutes...not to say that PoI spawning would start to look a tad bizarre, can you imagine all modules malfunctioning after a max of 5 minutes once they're damaged enough? I use that as a rough example but the message here being RNG is used in the background way, way too much to just slam a rule like this over it all; it would be literally game breaking.

You'd have to pick out singular RNG checks and modify them independently. It'd look awful, because the odds of a spawn at certain times are higher and so spawns would be grouped/inconsistent, and it's a lot of work when one could sit down and create a better solution.

I'm all for the tempering of RNG where it matters - like as I said, the RNGineers, which are overtly liberal in their application of randomness to the point of borkdom. But that's a tiny facet of RNG within the game; don't mess with it where it doesn't need messing with because you got indignant you couldn't find something for a while. Leave PoIs as a random and think of a better solution to making harder to obtain materials more obtainable - though I've always said one of the motivations to improve RNGineers would ideally leave the mats as hard if not harder to obtain than now, and rewrite the RNG so we arguably don't need more than one or a few rolls per ship for something we're pretty happy about.

 

[Valethar]

For a start we complain about RNG in ED a lot but the stuff that really matters is a tiny section, namely in the RNGineers. Almost every modern game uses the RNG almost constantly. You can't fight it, just make it more sensible.

The problem isn't so much that ED uses RNG, it's that the idiots that came up with Engineers thought that RNG was a good way to build a 'crafting' system. I can't think of any mainstream AAA games that use such a horribly botched approach to crafting as this.

A far better approach would be to have slots for upgrades. Every upgrade has benefits and drawbacks. Each ship module would have a specific number of upgrade slots that could be used, based on class and rating. You mix and match what goes into the upgrades you want on that module to get the performance you think will suit you best.

The player would still need to hunt down the materials required to manufacture the upgrade, and would still need to unlock various engineers in order to be able to craft them. Better upgrades would require higher quality materials, higher reputation with the engineer, a specific level of reputation with the local faction(s), specific ranking in the three main Pilots Federation careers, etc...

Each engineer would still be responsible for specific things, like they are now, but the ridiculously tedious idiocy of feeding the slot machine and hoping for a jackpot would be replaced by something that actually makes sense.

What we have now isn't crafting. It's Gambling Simulator.

I agree with you. They definitely need to come up with something that's more sensible and/or elegant. RNGineers is neither.

 

[Aashenfox]

This, eleventy millions times this!

one of the reasons I have not bothered with anything to do with the engineers is the pointless RNG system. I don't know why its set like this "magic" system, it should be a set of bonuses that are appied positivly and negitivly.

you want a longer jump range, ok fine but it will take longer to spool up.

you want faster thrusters fine but you will pay for it in ware and tear. or use more heat.

and so on.

The magic RNG is just silly.

Simon.

The trouble with this model is whatever is optimal would soon be learned and everyone's then flying around in the same ships as each other, back to square 1 when engineering didn't exist. RNG is a necessary evil. It is an EFFECTIVE mechanism of gating the best things to the most dedicated (or very rarely plain lucky) players. I don't think it's ideal, but I don't want engineers to be pointless overnight, which they would be if the RNG was taken away. I don't have a better suggestion at teh current time.

 

[Aashenfox]

And this is the part where you're wrong. While engineering Sidewinder and Anaconda may lead to different tuning results I would expect same results while tuning two Sidewinders. This is what I call quality. If my neighbor who has Ford Focus went to a car tuner and installed one chip and got certain performance boost I'd expect the same performance boost with my Ford Focus. Same car, same tuning, different outcomes is what we would see if Frontier applied their logic to your field.

On the contrary, any tuner who told me he could guarantee getting my focus the same result as my neighbour would be treated as dodgy, cos you can be sure, he tunes to a number, rather than tuning the engine he is tuning. A good tuner will tell you when you've reached the optimal limit for YOUR engine, they might give you the option to run riskier settings to bring it up a little more to match your neighbour, but that would be folly. The skill in tuning isn't making cars faster, it's making them faster without making huge compromises in reliability.

Additionally regarding RNG, you will never get the same result twice from the same car, tuner, dyno, or mods on the same day. NEVER. You'll get wild variation if you go in summer compared to winter, morning or afternoonl. Real car tuning is a very organic thing, and you can go and have your car tuned one day, then run it for a few days, and then take it for another tune, and squeeze a touch more out of it, as various things adjust. I drive a modified 350Z, I'm lucky, I have a strong one and access to 100 octane fuel, so my tune with the same mods is quite a bit stronger than a UK car with the same mods (just one example how the same 'car' can differ). I also run slightly heavier oil cos I live in a hot climate (Brit living in Greece). Secondly, no modified car owner wants to be identically tuned as any other. One migth want more torque low down for traffic driving, another might want all their power up top for track and drift. Some tuners will specialise in making a comfortable road tune that makes decent power (Evos for example, you can tune them to be undriveable but making insane top end power, or driveable with a bit less top end), others will specialise in squeezing every last horsepower out for track racing. This is really why 'chips' that you buy off ebay or whatever rarely make good power, every tune must be tailored to the exact car, not a whole model line. Two people with identical cars could apply the same tune, and one could run well and the other rough.

Since I typed these two posts, I came up with the answer... Let the overall VALUE of the roll stay the same, but allow people to tweak the ALLOCATION after. For example, the game simply sees a roll between 1 and 50, nothing more, then randomly splits those 1 to 50 points to the various sliders. Then we can pay a reputation fee (like special mods) to tweak the values as we wish, spending only that amount of points. perhaps only within a certain range. Anyway, that's one idea that would take away some RNG without removing all variation, while making it much more likely someone will get a roll they are satisfied with.

 

[Komemiute]

The real issue with RNG is it hides problems with the game.
After seeing Obsidian Ant's latest video and noting that Frontier have confirmed there is an issue with Unknown Probes spawning, I am left wondering how many other things are broken in the game but have been hidden by RNG?

How many other times has an issue cropped up regarding the spawning of particular mission rewards, USS, materials etc and been excused as RNG?

Having worked in customer support I know many organisations use troubleshooting flowcharts (or programs). I can easily imagine "RNG" would be the end of one of those branches, and provide a handy way to end that investigation (we used to call them "get out of jail free" cards).

What Frontier need to do is introduce a modified RNG which reduces the odds every cycle until after a set number cycles the odds become 1:1.
In this way Frontier can always run the cycle and confirm the results are working as intended.

It also means that players have a guarantee that for any RNG event (engineering, materials, mission rewards) they will (eventually) get the result they want.

This doesn't get rid of RNG, but it manages it in a way that is much more productive and satisfying. Lucky players can still gloat about how quickly they got a particular mod/material/reward, while unlucky player will have the guarantee that they will eventually get the result they want (instead of potentially never).

I like this idea but would like to provide an alternative.
it all stays the way it is now... EXCEPT... you can choose to "throw" more materials in the mix.
I'll explain better: if a recipe requires 2 of element#1 and 1 of element#2 you can chose to put more element#1 and more element#2, for each extra material the negatives reduce progressively* until 80% of max best value.

You still require some degree of luck to cover the last 20% (only of the good side- blue- of the RNG bar) of the odds but at least you can be sure your hard-earned materials are put to a good use.

That doesn't mean it'll take 100's of element#1 or 2, just a bit more.


*The progressive reduction of the RNG bar can be linear to logarithmic going from Tier 1 to Tier 5.

 

[Factabulous]

The problem isn't so much that ED uses RNG, it's that the idiots that came up with Engineers thought that RNG was a good way to build a 'crafting' system. I can't think of any mainstream AAA games that use such a horribly botched approach to crafting as this.

Diablo III comes to mind, I'm sure there are others as painful

It's designed to be a timesink. RNG has always been designed as a timesink.

Yup - this is why they will not get rid of the RNG. I think the most you could get is them to actually look at the spawn rates so that they know they have stacked them incorrectly - shouldn't rely on the community pointing out that CIF / UP etc etc are not available, they should be able to see that (ideally during the Beta phase) - looks to me like they are using RNG without monotoring to see it's actually working.

OP: The idea of just making a perfect roll inevitable will never happen - there are enough grinders on here who will (un)happily continue just in order to get to that perfect roll - OCD has a lot to answer for.

 

[Brrokk]

I like randomness in finding things. It should be hard to find things in space which is, after all, big.

I think engineers should know what they're doing and work to a specification though, and be able to replicate good modifications they've previously done on any given module type.