28 replies to this topic

[WolfTri]

Saw the previous thread here, making a separate thread to specify we're only talking about enriched rates. References have been provided by others in the previous thread. If you have more evidence please post.

12 Elder Wands (Level 3 Weapon)

1 broke at +5->+6

4 broke at +6->+7

5 broke at +7->+8

2 broke at +8->+9

 

7 Headgears, not a single one made it to +7.

2 of them broke at +4->+5

 

The average chance of a level 3 weapon going from +5 to +9 is 31% or so. That's one in 3. When all 12 of them break, either my luck is really horribly bad or the rates are broken. 

Other people also had similar results with enriched. One person being horribly unlucky is understandable. However, everyone posting having terrible luck makes a fair case for an investigation.

 

More results were linked in the previous thread, not a single person could confirm enriched rates were working properly (as of this week, at least)

 

Also keep in mind the first time this event was released, we had a bug where the enriched rates were not implemented. So we have all the more reason to suspect. 

This requires at the very least a test by the GMs, I don't have enough enriched ores to happily waste to prove a point.

 

 

 

 

 

Edited by WolfTri, 28 June 2017 - 07:09 AM.

[LordYggdrasill]

umm this. 1 broke from +5 to +6 is perfectly normal, thats 95% success rate and the player had 11 out of 12(91.6% success rate), so its not too bad. 4 broke from +6 to +7 smells fishy here though, it could be RNG but I feel very sorry for the player if this is the case. But anyway, this gives more reasons not to refine without protection, since you never know when and if the refine rate will stop working suddenly.

[ChakriGuard]

There have been many threads concerning about the enriched not working. Unfortunately people were quick to blame the RNG and ignored it.. So again maybe you and many other people just had bad luck xD

I also believed Enriched is not working. I broke all 14 mad bunny :3

[Ethanasael]

This things happens, totaly common for a human for missing. The problem is when there is no humilty for correcting the error, reward affected players and ask sorry for the problem.

[Latsabb]

I will do a very quick breakdown. You posted this:
 

1 broke at +5->+6

4 broke at +6->+7

5 broke at +7->+8

2 broke at +8->+9

Another way to say this:

 

1/12 broke going 5->6

4/11 broke going 6->7

5/7 broke going 7->8

2/2 broke going 8->9

 

So, we get your rates (in a very small sample set) as:

 

5->6 : 91.7%

6->7: 63.6%

7->8: 28.6%

8->9: 0%

 

So, for 5->6, you are MUCH closer to the event rate. Easily within error. You should have only had 60% without the event. 6->7 you are closer to the normal rates, but still pretty far from both. 7->8 you are even closer to the non-event rates. And last gives you no statistical information. For the 8->9, even with the increased rates, it is far from certain that you would get any of them. 60% chance, with only 2 tries really doesnt give you stellar odds. My take? You were very unlucky on 7->8, and the rest were spot on.

Edited by Latsabb, 28 June 2017 - 08:38 AM.

[Kusanagisama]

This things happens, totaly common for a human for missing. The problem is when there is no humilty for correcting the error, reward affected players and ask sorry for the problem.

 

I think the real problem is that you have no way to audit it. For example, it could display, for a 80% chance, something like this:

 

Event Active:

Your roll: 13, Required: 20. Upgrade has failed 

Your roll: 92, Required: 20. Upgrade successful!

 

Event not active (thus, 60% chance):

Your roll: 28, Required: 40. Upgrade has failed 

Your roll: 41, Required: 40. Upgrade successful!

 

 

Heck, even a "Refine Event Active!!" text like the Shadow refine girl has would be good enough.

Edited by Kusanagisama, 28 June 2017 - 09:57 AM.

[Latsabb]

No CHEAP way to audit it, you mean. It is very possible to check it. Start with 100 items, and you will get a much, much better idea of the distribution.

[Handersen]

This many ppl and they still blaming it on rng. Either they don't check or they just troll

[Kusanagisama]

No CHEAP way to audit it, you mean. It is very possible to check it. Start with 100 items, and you will get a much, much better idea of the distribution.

 

100 is not even close to a reasonable sample. You'd need at least 1000, preferably more, which is not viable, and if it was found that the event is actually NOT working, the loss to a lot of people will be enormous.

[Latsabb]

100 is not even close to a reasonable sample. You'd need at least 1000, preferably more, which is not viable, and if it was found that the event is actually NOT working, the loss to a lot of people will be enormous.

 

Considering the odds, 100 would be a very reasonable sample. You could easily get a 95% confidence level in the results considering that you only really need to test the 95%, 90%, 70%, and 60%. And the corresponding rates without the event are 60%, 50%, 20% and 20%.

Edited by Latsabb, 28 June 2017 - 10:23 AM.

[ChakriGuard]

Sometimes even if the success rate is 99% and the failure rate is 1% and you have failed 99 times out of 100, some people are more than willing to stand that, you just have bad RNG xD

[Latsabb]

Sometimes even if the success rate is 99% and the failure rate is 1% and you have failed 99 times out of 100, some people are more than willing to stand that, you just have bad RNG xD

 

Untrue, because the odds of that happening are so astronomically low. As you can see in my post, there is only one section, 7->8 which seems to be out of whack by any real margin, and the sample size is very low. If all 4 levels were low, it would be more alarming, but they arent. As a matter of face, 5->6 is well above what is to be expected if the rates werent turned up.

[Kusanagisama]

Considering the odds, 100 would be a very reasonable sample. You could easily get a 95% confidence level in the results considering that you only really need to test the 95%, 90%, 70%, and 60%. And the corresponding rates without the event are 60%, 50%, 20% and 20%.

 

The sample is still to low to have confidence in it not being only bad luck, and it reduces as you go to upper tier in upgrades, so, the confidence level naturally decreases with that. I know it's better than nothing, but the credibility of the server goes by the drain when you don't have a decent way to validate these events.

[Latsabb]

The sample is still to low to have confidence in it not being only bad luck, and it reduces as you go to upper tier in upgrades, so, the confidence level naturally decreases with that. I know it's better than nothing, but the credibility of the server goes by the drain when you don't have a decent way to validate these events.

 

You dont seem to understand. With 100 samples, especially working with three different levels, with odds this high, confidence levels would be well over 99%.

[LordYggdrasill]

Maybe all they need is to actually show the refine rate when you speak to the NPC. For instance, mighty hammer will tell you the rate is X% before you confirm to upgrade, similarly for other NPCs. In this case, players can monitor the refine rate and check if the event is working or not. Otherwise, the only way to confirm this information is for a GM/staff to double-check on their end.

[Latsabb]

What an NPC says, and what the mechanics are, are two different things, though. They can change the dialog to say 100%, but if they dont change the mechanic, it doesnt matter what the text says.

[Ethanasael]

Maybe all they need is to actually show the refine rate when you speak to the NPC. For instance, mighty hammer will tell you the rate is X% before you confirm to upgrade, similarly for other NPCs. In this case, players can monitor the refine rate and check if the event is working or not. Otherwise, the only way to confirm this information is for a GM/staff to double-check on their end.

I, Realy, would like this to be added, but via SCRIPT, variables, so there is no HUMAN HAND to modify, only enable the event and the refine rates at the npc will show how much they are now, not what they want the numbers to show, but the real rates, understand?

[Latsabb]

Now, what WOULD be useful, is having a dialog option with a loop in it. You ask the NPC to run some sort of diagnostic, and then specify the number of loops. For obvious reasons, you keep the max loop level low, like say 500. So, you tell the NPC to loop an upgrade that many times as a simulation, and then he tells you the number of failures that occurred in the loop. The loop calls the refinement function, but doesnt refine anything, just simulates. If you ask it to loop 500 times, and it fails 243, you can safely assume that the odds are around 50%, so if it should be 90, it is way off.

[LordYggdrasill]

What an NPC says, and what the mechanics are, are two different things, though. They can change the dialog to say 100%, but if they dont change the mechanic, it doesnt matter what the text says.

 

Well like Ethanasael said, this could be achieved through scripting, with the refine rate coming from a variable that they can alter during refine event. In this case, the dialog will be guaranteed to show exactly what the current success rate is.
 

[Latsabb]

It likely doesnt work like that, though. I would be surprised if the function was using numbers that would be easily used for dynamic text like that. Not to mention it seems like iRO doesnt have anyone that actually does any real coding, so keeping it to an NPC script is more likely to be feasible for them.

[ChakriGuard]

Untrue, because the odds of that happening are so astronomically low. As you can see in my post, there is only one section, 7->8 which seems to be out of whack by any real margin, and the sample size is very low. If all 4 levels were low, it would be more alarming, but they arent. As a matter of face, 5->6 is well above what is to be expected if the rates werent turned up.

 

Very true because the person's RNG is just bad xD

[Campitor]

The Zone servers all report that the refine rate is active, The boot up process also reports that it assigned the proper refine rates to the zone server. The event IS active.

[Kusanagisama]

Thank you, Camp. Will the event end tomorrow, or will it be extended until july 4th's week?

[WolfTri]

The Zone servers all report that the refine rate is active, The boot up process also reports that it assigned the proper refine rates to the zone server. The event IS active.

Thanks!

[thesifer]

Thank you, Camp. Will the event end tomorrow, or will it be extended until july 4th's week?

 

The event will end on July 6th. xD