Jump to content
Search In
  • More options...
Find results that contain...
Find results in...

Harvesting and crafting potions


Angelhearth
 Share

Recommended Posts

As they aren't going to make it to the final I can't see the importance in that. If you just want to check how many times the players use them, have a counter.

I'd be quite happy having them automatically, but it still provides valuable testing data.

 

Right now they want as much load on the crafting part of the database as possible and they build quickly.

 

It also lets them test the add recipe system by adding and then removing them.

 

Lastly, they would need to take steps to make them unlootable or to replace them if they were given freely.

 

Inventory protection is not in the current build and getting items for free is not something which would be in the final game so it would be extra work for little payoff.

Edited by ledeir
Link to comment
Share on other sites

Lets talk about the math a bit, and what I have learned regarding the shown odds of 77%

 

With odds of 1/4.347 I should be able to make a potion once every 5 attempts.

The odds of getting 6 fails in a row, should be 1/6755 tries

 

I have experienced in the making of less than 100 potions, 2 occurrences of 6 failures in a row.  That puts the odds of getting a failure on any six consecutive attempts at 1/50

 

Current experience tells me that the randomize function is WAY off, because there is currently an apparent 135 times greater likelihood for failing consecutively than true random math would indicate.Right now I would conclude that it's wonky as hell.  Maybe the built in Unity randomize function is not very well built, or the seed used is not actually a random seed.

 

Now, if we were only crafting final goods, there is no way I would have sample size high enough to even guess at this problem.  The randomize algorithm would "feel" off, but without enough data points that's all there would be, feelings.  

Edited by KrakkenSmacken
Link to comment
Share on other sites

If they gives unlimited potions we wouldnt harvest as much, i mean the potions make us 5-10 times more efficient? The crafting recipes dont need to be tested individually too much, if we make a basic weapon or a end-game one provide them the same results...

 

They gives us options to decrease the process time but we dont stop testing the game because of that. It is basically the same reason the armor decay is so big... and perhaps why it is full loot...

Link to comment
Share on other sites

Lets talk about the math a bit, and what I have learned regarding the shown odds of 77%

 

With odds of 1/4.347 I should be able to make a potion once every 5 attempts.

The odds of getting 6 fails in a row, should be 1/6755 tries

 

I have experienced in the making of less than 100 potions, 2 occurrences of 6 failures in a row.  That puts the odds of getting a failure on any six consecutive attempts at 1/50

 

Current experience tells me that the randomize function is WAY off, because there is currently an apparent 135 times greater likelihood for failing consecutively than true random math would indicate.Right now I would conclude that it's wonky as hell.  Maybe the built in Unity randomize function is not very well built, or the seed used is not actually a random seed.

 

Now, if we were only crafting final goods, there is no way I would have sample size high enough to even guess at this problem.  The randomize algorithm would "feel" off, but without enough data points that's all there would be, feelings.  

Well as you said under a 100 made, and well 100 is the VERY LEAST you need to begin making conclusions about the RNG function, also truly random will mean that streaks will happen. If you know enough about statistics you can in fact spot from a large enough sample size if something is truly random, by the fact that the truly random is NOT well distributed but have streaks in it, that is simply how statistics work. As a matter of fact the pots I been making seem to kinda gravitate towards its failing way less than 25% of the time when it comes to pots

Link to comment
Share on other sites

i have to say the random elements feel indeed very off atm.

Crafted a dozen equipments with potions running. 98% succes changse. Still failed a good 10% of them. 

 

It could indeed be bad luck and most likely was. What they could do is a fail safe mechanic.

If my changse is 98% and i failed 2 times then the other 98 times can not fail. So that per statistic i cant fail more then 2 times per 100.

 

But do i want such a fail safe? iNot really.

 

I played ragnarok online in the past where some drops where 0.01% changse. Per statistic i need to kill 10000 mobs to obtain a drop.

I killed around 70000 to obtain it while my friend killed 20. RNGesus hates me sometimes and loves me other times.

And that in itself is fun to.

o8WHnLc.png

THE most active European guild. Join us now!

Link to comment
Share on other sites

Well as you said under a 100 made, and well 100 is the VERY LEAST you need to begin making conclusions about the RNG function, also truly random will mean that streaks will happen. If you know enough about statistics you can in fact spot from a large enough sample size if something is truly random, by the fact that the truly random is NOT well distributed but have streaks in it, that is simply how statistics work. As a matter of fact the pots I been making seem to kinda gravitate towards its failing way less than 25% of the time when it comes to pots

 

Yes random streaks will happen. For example one site reports that the world record for roulette single number sequence from a reliable record was 6 times (1/3,010,936,384) in a row, while the record for hitting the same color was 32 (in 1943)  (1/10,321,314,387).

 

I would agree that 100 was in fact far too small a sample to determine the true randomness of the system, and is basically anecdotal evidence, however due to the appearance of numerical favoritism and patterns in what is a very small sample size, and frequent reports of it "feeling" wrong, I would say that a proper examination of the RNG behavior is in order now.

 

I say that because moving forward everything is going to be balanced based on that behavior. If balance is attempted and then a problem with the RNG is discovered, they will have to go back and re-balance everything again once it is behaving as it is expected.

 

If they however elect to live with a pattern bias, it opens up all sorts of possibilities to exploit what could very well be predictable behavior and results.

Link to comment
Share on other sites

Having played a crafter in many games I have found that the best way to calculate randomness is to use the percentage on the tries combined. If you have a 50% chance, that's what you have on yoyr first try, but then you get a 50% chance on the 50% left, ie. 75% chance, and so on. That way, you can still fail indefinately, but the chances of failing will be so small after a couple of tries that it's close to negible. If it's a 50% chance for every successive try the chance of failing indefinately is actually quite possible. I know, after a few of those failing streaks in WoW. It's so disenheartening I quit playing for several days.

Link to comment
Share on other sites

Lets talk about the math a bit, and what I have learned regarding the shown odds of 77%

 

With odds of 1/4.347 I should be able to make a potion once every 5 attempts.

The odds of getting 6 fails in a row, should be 1/6755 tries

 

I have experienced in the making of less than 100 potions, 2 occurrences of 6 failures in a row.  That puts the odds of getting a failure on any six consecutive attempts at 1/50

 

Current experience tells me that the randomize function is WAY off, because there is currently an apparent 135 times greater likelihood for failing consecutively than true random math would indicate.Right now I would conclude that it's wonky as hell.  Maybe the built in Unity randomize function is not very well built, or the seed used is not actually a random seed.

 

Now, if we were only crafting final goods, there is no way I would have sample size high enough to even guess at this problem.  The randomize algorithm would "feel" off, but without enough data points that's all there would be, feelings.  

 

I also note that if feels like fails beget more fails...   in other words bad luck is streaky. 

6FUI4Mk.jpg

                                                        Sugoi - Senpai

Link to comment
Share on other sites

I also note that if feels like fails beget more fails...   in other words bad luck is streaky. 

 

Yes. No hard numbers but changing recipes after a fail produced many fewer consecutive failures. Changing recipes and changing back without crafting something else seemed to still cause consecutive failures.  But the sample size is low on that.

Link to comment
Share on other sites

Yes. No hard numbers but changing recipes after a fail produced many fewer consecutive failures. Changing recipes and changing back without crafting something else seemed to still cause consecutive failures.  But the sample size is low on that.

I can see a few reasons for this, and have in other things I built in the past.  Usually it revolved around poor seed management.

 

Quick lesson on RNG's.  Usually RNG's are a series of 0's and 1's with a random distribution in a limited universe or period. These are usually incredibly large universes. Mersenne Twister is 219937 − 1.  

 

Think of a clock as a simple example of a very small universe.  If you randomize the numbers, then for example 7 will always follow 10, and 9 will follow 7, etc.  If you were to ask for Odd/Even RNG from the "clock" you would see the exact same pattern emerge after only 12 requests.

 

The seed tells you where start at, for example the 7.

 

Now depending on when you initialize the seed in the process, and what you use, the function will return the exact same results, the same as the clock that is always asked to re-seed at 7.  

 

So in our case, if the seed was determined (say a time value) when the recipe is selected, and re-seeded with the same time value every time you added new items to the formula, every result would be identical, because you kept telling it to restart at 7.

 

I would also be worried about seeds being client side, as those could be manipulated at the RAM level to gain complete predictability of events.  The main server should keep all randomized functions internally, and crafting request results, if tampering is to be prevented.

Edited by KrakkenSmacken
Link to comment
Share on other sites

 Share

  • Recently Browsing   0 members

    No registered users viewing this page.

×
×
  • Create New...