Melding - and chance of luck

Hello,

I just bought 20 pieces of tenacity materia, and I melded and I melded, - there was a 17% chance of succes, but none of the 20 sucesseded.

Now, and because Ive melded a little before writing this post, Im suspecting Yoshi P to not being honest as when he says 17% succes.

- When patch 5.2 came out and I was a little lala I heard people saying, "wauv, penta-melding is easy".

And now I cant get a 2nd or 3rd on.

Have something being changed and the tooltips are not exact?

In Japan you can flip a coin 30 times in a row and it will always come up tails.

The game knows you're trying to meld tenacity and it's trying to stop you.

It's a 17% chance every time. Not 1/17. It doesn't keep track.

Since we are not in a cheap f2p p2w game which try to sell you overpriced "+xx% chances to succeed the upgrade of a piece of stuff" items in the cash shop, i dont see why SE would tell wrong numbers on purpose. This is just how luck and RNG are working, you are not safe even with a 98% chance. Sometimes i succeed a 5% meld in 3-4 trys, and sometimes it take me 50+ of them.

That or you didn't pray RNGesus enough.

You have to sacrifice one Lala for every slot you want to overmeld. But since you're a Lala maybe RNGesus didn't accept your offer?
You could try a virgin or maybe a newborn but no guarantees.

I've succeeded first try on the 5% slot and then I NQ crafts with 98% chance to HQ. Sometimes you get lucky, sometimes you don't.

You will rarely ever see actual success over a moderate period of time match the % chances.

You can replicate that by using the HQ% on gathered mats (the new relic ones are good for that since the nodes never have quality% bonuses on them).

Its only over a significant amount of attempts you'll see the actual be close to the predictive.


You did get very unlucky though, I usually assume ~2.5x as many materia as you'd expect, and it generally ends up being about 2x as many over the course of pentamelding a full set.

Like people have said it's a 17% every time. It doesn't keep track.
It's a known issue in game design. Percentages always feel like they are lower to the players than they really are. It is most definitely 17%. Some games are aware of this and make it so it DOES keep track and you're guaranteed a proc, say every 5 for 20%. But FF14 does not do that

Learn some probability. If you do an action with a 17% success rate 20 times, there is still a 2.4% chance that all of them will fail. No matter how many times you do it, it will not be a 100% success rate. So long as each attempt has a chance of failure, there is a chance that all of them will fail.

Quote:

---

Originally Posted by EaMett

Like people have said it's a 17% every time. It doesn't keep track.
It's a known issue in game design. Percentages always feel like they are lower to the players than they really are. It is most definitely 17%. Some games are aware of this and make it so it DOES keep track and you're guaranteed a proc, say every 5 for 20%. But FF14 does not do that

---

With how it is set up now, it could easily keep track adding onto the success rate with each failure.

However, to the frustrated player any success rate no matter how high or low is looked at as a flat 50%.

Some truly believe RNGesus is a true entity; out to bless or smite as they see fit. This is how things are explained such as like blowing up 100 materia when you really need that last meld, while melding the lowest chance with materia you don't need just for shizNgiggles pops on there each and every time.

Quote:

---

Originally Posted by EaMett

Like people have said it's a 17% every time. It doesn't keep track.
It's a known issue in game design.

---

No it's not. It's a known issue with peoples' understanding of statistics. Namely, they don't understand statistics.

The main problem is that whatever algorithm SE is using, tends to cluster the results.
Meaning, sometimes you get lucky in a row and sometimes you get unlucky ... A LOT and feel as if you're running against a wall.
IMHO that happens far too often and results should be more evenly distributed.

Still, statistics say that as players, our sample sizes are pathetically small. We will never feel the actual %age of success.

Quote:

---

Originally Posted by Granyala

The main problem is that whatever algorithm SE is using, tends to cluster the results.
Meaning, sometimes you get lucky in a row and sometimes you get unlucky ... A LOT and feel as if you're running against a wall.
IMHO that happens far too often and results should be more evenly distributed.

---

No, just no. Even if there were systemic clusters of random numbers a single player isn't going to see it.

Random is random. If there weren't streaks it would be less random. Streaks, even long ones, are expected in a truly random sample.

P.S. the op's situation of a 17% success chance failing 20 times in a row has roughly a 2.4% chance of happening. Unlucky, yes. Out of the realm of possibility, definitely not. Unintuitively higher than expected, absolutely. Humans are bad at judging probabilities.

rng in videogames is always interesting to me.

I tend to get the bad luck streaks. sorta like ruby weapon a week or so ago. We ran 12 runs, out of those 12 runs 6 ninja weapon drops dropped (no one needed ninja). Was becoming a meme. the chance of 6 ninja weapons dropping out of 12 runs with each fight having a pool of 17 weapons is astronomically low.

But that stuff happens all the time across multiple different mmos I feel. I mean even the retrieve materia function that is 40% I fail 5-6 times in a row all the time.

I know with large samples it will average out but still interesting how things do tend to cluster in video games. Either get good luck streaks or bad luck streaks. Never actually get the odds in 1 session.

The issue with true randomness, where each result is independent of past and future results is that it's incomprehensible to the human brain. The brain is so obsessed with patterns that it sees them even in places where there simply isn't any. This leads to issues such as gambler's fallacy and refusal to believe getting the same result twice in a row at low probability is actually just randomness.

In a lottery, it is just as likely for the numbers 1,2,3,4,5,6,7 to be drawn as it is for 3,12,14,21,33,35,36. The brain will however see a pattern with the former and cry foul. It is also just as likely for the same numbers to be drawn the next draw as it is for any other combination, but the brain will cry foul at the former.

When games start adjusting their RNG in a way the brain believes RNG works, the devs are actually making it less random.

Quote:

---

Originally Posted by Kukulumi

Hello,

I just bought 20 pieces of tenacity materia, and I melded and I melded, - there was a 17% chance of succes, but none of the 20 sucesseded.

Now, and because Ive melded a little before writing this post, Im suspecting Yoshi P to not being honest as when he says 17% succes.

- When patch 5.2 came out and I was a little lala I heard people saying, "wauv, penta-melding is easy".

And now I cant get a 2nd or 3rd on.

Have something being changed and the tooltips are not exact?

---

Did you try sacrificing some aldgoats? That may increase your luck with the RNG gods.

seriously as people stated its a 17% chance of success on each attempt. there is no tracking. It's just sheer luck with RNGesus. unfortunately RNGesus is not a kind and forgiving god.

Quote:

---

I tend to get the bad luck streaks. sorta like ruby weapon a week or so ago. We ran 12 runs, out of those 12 runs 6 ninja weapon drops dropped (no one needed ninja). Was becoming a meme. the chance of 6 ninja weapons dropping out of 12 runs with each fight having a pool of 17 weapons is astronomically low.

But that stuff happens all the time across multiple different mmos I feel.

---

Varies by the MMO on exactly how they do it, but!:

Each MMO uses different formulas for mob drops.

EQ rolls for every possible item that can drop from a trash mob, weighted from impossibly low (expansion-wide rare drops) to relatively common (species specific vendor loot and zone-specific vendor loot) and uses that to determine the actual loot.

For raid group nameds, it rolls once for each guaranteed loot spot, on a table, with ranges based on how rare the item is supposed to be. Most nameds will have table drops one spell turnin possibility, which is weighted slightly towards the ranks melees get the most abilities, one visible item possibility, one non visible item possibility and one ornamentation possibility as a bonus. Sometimes you'll get 2, sometimes 4, usually 3.

For raid encounters, the boss drops and chests will roll on a table for pools each time for each loot space they have. Some of those encounters have multiple pools (Vulak, for instance, in old school velious will drop 3 items, and each item is pulled from a pool of 8 items on his table. Other raids will roll X times on the full table and can have duplicates. In Plane of Time, Quarm will drop 3 normal items, each item coming from its own pool [1 melee, 1 caster, 1 hybrid], and then have a chance to drop a 4th item that will only be limited to one of the 4 chest slot molds).

WoW uses a similar system, but limits the possibility of world-epics to one drop per mob.


The seed for the rng also varies by game. The lazier they are, the more often you see streaks.

It's 17% each time. Trust me rng can be mean at times. I have failed gathers with 99% all 4-5 gathers on a single node 37 times. Just because you failed 20 times doesn't make the next one any more likely and the game isn't going to petty you either. That's why for mount drops trials they added token incase you just are that unlucky.

I used to take a lot of issue with the melding rates, but now that DoL/H materia are coming out of our ears due to the spiritbonding changes the rates feel more reasonable now.

Still wish you could choose to burn more materia for a guarnteed meld, though.

Quote:

---

Originally Posted by Valkyrie_Lenneth

It's a 17% chance every time. Not 1/17. It doesn't keep track.

---

Well no it's not 1/17, it's more like 3.4/20. :p