Starman complain about "Portalarium Crafting Math" #6 with data.

We've been saying there's problems with RNG and the endless time vs reward for weeks. Nothing's acknowledged by the Devs other than 'it works the way it's supposed to'.

So they'll learn, ultimately. I have the sense that a lot of the people playing at this point in time aren't going to be playing by the time launch comes. It's a lesson learned for me; I won't participate nor contribute to any other Early Access games in the future.

 

It isn't completely random. I always get 2 exceptional in a row, or 1,0,1 in a line.

Craft parts, non-batch, until you see that series come up again. Then pop over for a combine.

This would not work if it was true statistics. This is more like the fake RNG for VLTs

 

It's exactly what it is.

 

Game the system so the odds are in your favour... Then they will fix it

The probability of my 3 consecutive fails at 93% success = 0.0343 % (unless my math is off)

This also occurs very frequently while playing.

 

Where to start...

So I hope everyone here knows how RNG works at least to some extent. The simplest way I look at it is like rolling a 1000 sided dice. If I have a 33.33 percent chance, if I roll less than or equal to 3333 then I have success, if not I fail.

So with this method unfortunately there is no stored data from your last attempt to make adjustments. So each attempt you roll the dice and each attempt has a 33.33% chance of success. This process overall works and like the earlier posters mentioned, with large amounts of rolls 1000's you will end up with an overall percent of success close to 33.33%. Unfortunately, this method leaves the actual success of each attempt more up to luck than skill. Even with a 50 percent chance you could end up failing 10-15 times in a row just from being unlucky, or succeeding 30-40 times in a row from luck.

So when I say this is broken I am mainly trying to say this system sucks without having to explain math that most will not understand. With the advancements in technology, coding languages and programmers knowledge they do have the ability to add in additional controls to minimize the lucky swings up and down. But unfortunately these changes would have some cost associated with them and can be exploited, but if it could prevent 15 fails in a row With a 75% chance. I am all for it.

I know it would be more complex than this because it would have to be for each skill / percent but something like:

Percent Chance = ((Percent Chance * Evaluation Attempts) - Current Successfull Attempts) / (Evaluation Attempts - Current Attempt)

What this does is set a Evaluation period of 10, 100, etc. Now it will look at your success in that period to determine what the current success chance should be to reach the desired percent by the end of the evaluation period. As your percent chance is now changing on each attempt based on your pass/fail rate through the period this will curb the lows and highs of a lucky roll.

 

Maybe instead of blowing up the item, they could make it take a permanent max durability hit on failure. That would be a lot better than straight up losing items you put a lot of work into gathering the materials for. They'd still be useful to someone. You just wouldn't be able to get as many enchants or masterwork onto them due to the reduced durability.

IDK how to fix the exceptional chance problem other than if the game guarenteed you that rate by tracking how many you have made and then forcing an exceptional upon so many failures. So if you made 4 pieces and the first 3 weren't exceptional, it would automatically make the 4th one exceptional.

 

Or you get: "Rescue item for 5 ingots"

 

Okay, time for some big boy math. Crafting is, at least in theory, a series of independent tests with a fixed (assuming no skill change, etc) target. In statstics, calculating the odds of a specific number of those tests passing out of a specific number of tests is calculated using the binomial distribution. While the formula is a bit hairy to calculate by hand for larger numbers that aren't trivial cases, most modern spreadsheet programs haVe a function built in for it.

For example, if you crafted 100 items with an exceptional chance of 25%, there is a 9.18% chance that exactly 25 of them will be exceptional. There is also a 1.00% chance that 16 will be, and 1.12% chance that 34 will be. In fact there is a bit better than a 99% chance that you will get between 14 and 36 exceptional items. However, there is less that a 0.01% chance of only getting 10. If you craft 100 each day everyday for 20 years you'd have about a 50% chance of seeing it happen at least once.

 

Well..for fun I kept track of some numbers.

I made 20o items at 98%. 4 failures.

I made 120 items at 95%. 13 failures.

Sometimes I think I see a pattern. It could be that they are using pseudo-random numbers. Or the seed resets to the same number at times.

 

The maths is mostly irrelevant to the issue I think. We don't need a crafting system that is mathematically perfect across hundreds of thousands of attempts, we need a less frustrating one across the tens or hundreds of attempts of a single player.

Unless you have one of these plugged into your computer all random number generators you have on it are pseudo-random.

 

I also don't think it is as much about the numbers as it is with the experience. So say the numbers become perfect, but oh perfect is a perception unique to each person. Still imo and perception there are things missing from crafting that would increase the enjoyment factor overall and they don't necessarily have to do with the numbers and don't get me wrong I do not say anyone is wrong or devalue what is being said here since the numbers are what we have to go by atm but if the fun and satisfying factors of crafting were not enclosed only in this vortex then what would we have? Perhaps we would have crafting that wasn't bashing keys and crafting that was more interactive and surprising. But that's another discussion of course and I do commend everyone for their hard work on trying to find balance with what we have so far.

 

I would say that the first is a prerequisite of the second. If you can't trust the number displayed, you can't plan out how much you need for a project or how much it would cost.

 

Well then Port needs to put much more gold into the system so we can afford to buy these expensive rare items.

Simple as that.

They'll just have to accept this is the system they've created.

 

No. There is no need for the crafting system to be based on independent tests.

You already can't plan out how much something would cost, you can only (if you run the numbers like you did above) work out how much mats you need to give you a 99% probability of getting at least a certain number critical successes (and then factor in masterworking and enchanting).

 

You'd prefer the XCom effect? Wikis full of complicated systems of combining materials and forcing failures to maximize the chance of a success despite what the number shows? I'm more of a KISS guy that likes the system to be up front with the odds and modifiers.

 

I'd prefer a crafting system which didn't shaft me after spending a week of playtime gathering mats.

 

Well then you shouldn't have picked a game where "punishing the player for trying to excel" isn't a primary design element

 

Huzzah for dev griefing players!

 

As long as we have a crafting system based on RNG/Gambling rather than player skill it will be unfair.

 

Who says we don't have a robust DvP system?!