Recipe Drops are screwed passed 99

[lemming]

Could it be something with parking in a dayjob that gives a better recipe drop rate?

As Faultline noted, it doesn't seem to be related to the vet level.  Could be some other variable that gets tweaked.

 

And one thing I have always felt that the RNG gets itself in ruts.   Doesn't seem to matter if it's hit/miss, opening hero packs, drop rates, etc...

Over a period of time, it seems to even out, but when you're on a streak of getting the worse rolls, it feels personal.

[UberGuy]

11 minutes ago, lemming said:

Over a period of time, it seems to even out, but when you're on a streak of getting the worse rolls, it feels personal.

 

That's just the nature of unbiased (pseudo) randomness. And by "unbiased" I don't mean it's not got a chip on its shoulder. Rather, I mean that the shape of the distribution of numbers over time looks like a nice Gaussian curve.

 

Streaks of similar outcomes are actually completely normal in the short term even for completely random systems. A lot of folks assume "fair" randomness won't do that, but it totally does.

 

And, yes, when that turns against you on conversions or missing a dangers foe a bunch or something, it really stinks.

 

Edited May 5, 2021 by UberGuy

[Kyksie]

11 hours ago, LadyDee said:

I will gladly speak to every player on every server and put together a consensus. I will do a video highlighting my drops over four hours (or more). I will stand “naked” in Atlas Park, asking people for information regarding the recipe shortage and will donate $1.00 to any charity for viable information.

 

 

So... uh... let me see if I get this straight. You're doing an MSR, and the words "Recipe Found!" don't appear in purple on your computer screen like they did on the previous 872 raids you did this week. You talk to a few 'friends' who mutter 'mmm yeah', and the obvious conclusion is that the devs have turned off purples after 99 as part of a malicious plot to, uh, deprive you of purples, and the GMs are all in on it except for one brave GM who dared speak the truth. And now you're going to devote your life to rooting out the evildoers by standing in Atlas Park "naked" and donating A WHOLE DOLLAR to Autism Speaks in exchange for information on this vast conspiracy, is that right?

 

The Veteran Level field is just a number that increments when you ding, and is not looked at by anything other than your bio. Saying that it affects your drop rates is exactly like saying that, when mommy drove you to McDonalds last week, you didn't get a Happy Meal toy because the last digit of the odometer was a 9.  I mean, I used to code a bit for fun, so I know how fixing one bug will cause seventeen new and exciting glitches to pop up, even in parts of the code which have no connection whatsoever to the part you chang- what? Another compiler error? I didn't even TOUCH that module, you flea-infested piece of... version mismatch??? AAAAUGHHHH(*@!#*!*#

 

(deep breath)

 

Look. You just got unlucky for a bit. There is no huge conspiracy between the devs, GMs, New York bankers, and the Thule Group to take away your purples. Let me remind you that those people are all volunteers, working without any compensation, so that you can have the pleasure of seeing "recipe found!" pop up on your computer screen 89324 times a day. (except for the New York bankers, they never work for free.)

 

Edited May 5, 2021 by Kyksie

[LadyDee]

Once again, never said it was thee cause of the drop rate and I cannot control how other people read. So if someone chooses to start a conspiracy theory (about a game none-the less) well, that is their issue not mine.

 

Suggesting it was past level 99 was the only conclusion I came to after exhausting all other possible reasons. Then when others seemed to have a lower drop rate past said level, it just added more information that this could be the reason but not 100% certain. I'm not a moron, just because something seems to be the cause does not mean it is, but at least it's a place to start.

 

I did offer data in my original post. Yet no one has offered an explanation. Maybe the drop rates in certain events or TFs/SFs are screwy. Who knows.

[arcane]

I just want to thank the OP for spiciest thread of the week.

[wjrasmussen]

15 hours ago, Faultline said:

What GM took your report? The issue as described is impossible. Veteran Levels are a single integer field (meaning it goes up to 4 billion) and there's absolutely nothing special about it going over 99. The rewards code is not aware of veteran levels at all. If this is really happening, there's another cause.

Goes up to 4 billions.  Wow, I am feeling like a slacker now.

[arcane]

5 hours ago, UberGuy said:

The reason for wanting to know the GM's name is very, very certainly so Faultline can chat with the GM about any additional details they may have. The devs and GMs talk to one another. There's nothing punitive going on.

That’s what they want you to think. The truth is these phat cats are going to squeeze the OP until she names names so that this rogue GM double agent can be shipped off to Siberia.

[Faultline]

3 hours ago, LadyDee said:

I did offer data in my original post. Yet no one has offered an explanation. Maybe the drop rates in certain events or TFs/SFs are screwy. Who knows.

 

The only data that would be useful is the character name and server so it can be copied off the database and tested to see what drop rates it gets. I did chat to the GM involved; they had their difficulty set to -1/x6 with no bosses, so going a mission without a non-common recipe drop under those circumstances is not unusual.

 

Let's say the mission was composed of 100 minions and 30 lieutenants - that's 2.66 recipes from the minions and 1.6 from the lieutenants for a total of 4.26 recipes, each of which has a 75% of chance of being common. That doesn't mean you're guaranteed 1 in 4, it means you have 4 separate 25% chances for a non-common; the probability of getting one non-common is 68%, the probability of getting all commons is 32%. So running a mission with no bosses and getting zero non-common recipes is perfectly within bounds.

 

I copied the GM's level 126 Mind/Kin that was used for that mission to a dev instance, repeated my 10K critter test, and these are the results:

 

Minions: 104c 24u 5r 2vr; total 135/5000 = 2.7%

Lieutenants: 139c 34u 8r; total 181/3000 = 6.0%

Bosses: 126c 32u 9r 1vr; total 168/2000 = 8.4%

 

Everything is within bounds. Rarity distribution: 76.2% common, 18.6% uncommon, 4.5% rare, 0.6% very rare.

 

If you want to provide your character I'll repeat the 10K critter test with it, but after running this test three times, I am not seeing anything to indicate a problem with the drop rates. 

[LadyDee]

1 minute ago, Faultline said:

 

The only data that would be useful is the character name and server so it can be copied off the database and tested to see what drop rates it gets. I did chat to the GM involved - they had their difficulty set to -1/x6 with no bosses, so going a mission without a (non-white) recipe drop under those circumstances is not unusual.

 

Let's say the mission was composed of 100 minions and 30 lieutenants - that's 2.66 recipes from the minions and 1.6 from the lieutenants for a total of 4.26 recipes, each of which has a 75% of chance of being common. That doesn't mean you're guaranteed 1 in 4, it means you have 4 separate 25% chances for a non-common; the probability of getting one non-common is 68%, the probability of getting all commons is 32%. So running a mission with no bosses and getting zero non-common recipes is perfectly within bounds.

 

I copied the GM's level 126 Mind/Kin that was used for that mission to a dev instance, repeated my 10K critter test, and these are the results:

 

Minions: 104c 24u 5r 2vr; total 135/5000 = 2.7%

Lieutenants: 139c 34u 8r; toral 181/3000 = 6.0%

Bosses: 126c 32u 9r 1vr; total 168/2000 = 8.4%

 

Everything is within bounds. Rarity distribution: 76.2% common, 18.6% uncommon, 4.5% rare, 0.6% very rare.

 

If you want to provide your character I'll repeat the 10K critter test with it, but after running this test three times, I am not seeing anything to indicate a problem with the drop rates. 

 

You are speaking and testing in missions.

Ship Raid--30+mins.....ZERO drops

TinPex TF----About 45mins to a 1hr --ZERO drops

Miss Lib TF---about 40 minutes--- ZERO drops.

By your assessment, it is perfectly within bounds to not get a single recipe drop within oh about 2-21/2 hrs. Even though on every other TF, SF, Raid I have done, normally I'd walk away with a handful even if they are crappy.

[Arbegla]

10 minutes ago, LadyDee said:

 

You are speaking and testing in missions.

Ship Raid--30+mins.....ZERO drops

TinPex TF----About 45mins to a 1hr --ZERO drops

Miss Lib TF---about 40 minutes--- ZERO drops.

By your assessment, it is perfectly within bounds to not get a single recipe drop within oh about 2-21/2 hrs. Even though on every other TF, SF, Raid I have done, normally I'd walk away with a handful even if they are crappy.

 

Fundamentally speaking, missions are no different then TFs, as those are just a tightly bound group of missions with a specific story.

 

The MSR might be slightly different, but for the sake of testing, its also probably fundamentally the same. Could you maybe, I don't know, provide the actual information he asked for? Character name and server?

 

For the sake of testing, his parameters seem accurate, as the other 'variables' shouldn't have an impact on drop rates whats-so-ever.

[lemming]

4 hours ago, UberGuy said:

 

That's just the nature of unbiased (pseudo) randomness. And by "unbiased" I don't mean it's not got a chip on its shoulder. Rather, I mean that the shape of the distribution of numbers over time looks like a nice Gaussian curve.

 

Streaks of similar outcomes are actually completely normal in the short term even for completely random systems. A lot of folks assume "fair" randomness won't do that, but it totally does.

 

And, yes, when that turns against you on conversions or missing a dangers foe a bunch or something, it really stinks.

 

Exactly.

[Faultline]

9 minutes ago, LadyDee said:

You are speaking and testing in missions.

 

No, I'm not. This latest test was in Echo Plaza (no enemies around, easy to set up) using the same Rikti critters used in the RWZ raid. The rewards code doesn't know anything about how a critter was spawned, so the fact that I spawned them with commands rather than via script makes no difference.

 

13 minutes ago, LadyDee said:

By your assessment, it is perfectly within bounds to not get a single recipe drop within oh about 2-21/2 hrs. Even though on every other TF, SF, Raid I have done, normally I'd walk away with a handful even if they are crappy.

 

Correct, because the drops are based on how many enemies you kill and get credit for, not how much time you spend in-game. In a ship raid in particular, there are lots of critters, but there are also lots of players.

[overmanprime]

5 minutes ago, LadyDee said:

Ship Raid--30+mins.....ZERO drops

TinPex TF----About 45mins to a 1hr --ZERO drops

Miss Lib TF---about 40 minutes--- ZERO drops.
By your assessment, it is perfectly within bounds to not get a single recipe drop within oh about 2-21/2 hrs. Even though on every other TF, SF, Raid I have done, normally I'd walk away with a handful even if they are crappy.

*Sigh*
Speculation
Conjecture
Presumption
Not to mention a remarkably small sample size. Faultline and several other people have run mutliple tests, using far larger sample sizes, Faultline has taken a look at the code, spoken to the GM you reported it to. Run more tests. You've been asked, several times now, to provide your character details so Faultline can actually take a detailed look at your character under controlled conditions to see if there's anything weird.
So, one hypthesis is that you are either a troll or you have serious difficulty understanding even basic concepts. You have, after all, shown that your reading comprehension of what people have ACTUALLY SAID is a bit... creative to say the least.
So, do as the nice Dev has asked you to do (provide him with your character and server details so he can replicate) and this whole issue can be properly investigated and put to bed. Alternatively, if you don't want to do as you've been asked, and still want to contribute something more to the conversation, try contributing silence.
 

[Coyotedancer]

Having done ship raids and other things that resulted in jack-all as far as drops go, myself, I feel like it's safe to say what this really comes down to is the RNG being an absolute butt.

 

Runs of bad luck happen. It's not broken code... It's just the nature of the beast. I've had it happen on lowbies and on my "finished" 50s, so it's not a matter of level, veteran or otherwise. The problem... if it is one, which is debatable... is that the RNG can be streaky as all hell, and when you're on the low side of one of those streaks you really can go a ridiculously long time without getting bupkis. 

[UberGuy]

Every normal critter has a chance to drop one recipe on defeat. When you are on a team, which teammate gets that drop is random and evenly distributed*, and doesn't care what level the person is, how much damage they did, or whether they even have room for the drop. On a team of 8, you get 1/8 of the drops, on average, than you would if you defeated all that stuff yourself. Of course, it's very likely you defeat it all faster, but maybe not always 8x faster.

 

In a raid, its much more complex. Which team is given the chance to get the drop from a critter does depend on how much total damage was done by each team that damaged it. (A team is considered to have damaged a critter if any member of the team did damage to it.) My experience is that ship raids give me jack squat for drops. I consider myself lucky if I get 1 drop. 3 is outstanding. (The point at ship raids is to damage as many things as possible before they die rather than focusing fire on single foes, because every Rikti that dies that anyone on your team damaged at all gives everyone in reward leash range of it a VM on defeat.)

 

The point is teaming is almost always butt for drops compared to solo, unless you're soloing on something that kills stuff slowly.

 

* There are caveats here - you have to be on the map and, if it's in an open zone, you have to be within a certain distance of the critter when it dies, unless you personally damaged it, in which case you get your (random) share of the rewards at any range.

Edited May 6, 2021 by UberGuy

[Arbegla]

11 hours ago, UberGuy said:

Every normal critter has a chance to drop one recipe on defeat.

 

Not to derail this too much, but what is the definition of 'normal'? I've seen AVs for example give multiple people drops when they died, and GM code seems weird in that sometimes everyone gets something, and sometimes it doesn't.

 

For example, I got the merit and Temp Power reward from the new Paladin in Kings Rows, but I didn't get the badge. So somewhere in the reward code, I was tagged to receive something, but not all of the things.

[UberGuy]

I don't know the formal definition for it, but non-unique minion/LT/boss/EB mobs don't give everyone a drop. They drop one thing from each reward category (enhancement, recipe, salvage, etc.) and each only goes to one person.

 

AVs and GMs normally only have special drop rules for SOs, in which they give everyone who damaged them (or perhaps met some damage threshold). DE Monsters have special rules where everyone on a team that damaged them is eligible for an EoE drop.

 

Incarnate rewards are different. I forget the specifics. I believe everyone on a league can get Threads from an iTrial critter defeat, for example.

 

I'm not aware of anything that drops a recipe or invention salvage per person. Only special rewards, like EoEs or Ambrosia, or SOs from AVs.

 

Edit: Doing some digging, it seems like there are two places unusual rewards can be attached to a critter. Either in its EntityDef in the "additional_rewards" field, which is something you can see on City of Data's raw data view, or in the spawndef that sets up the enemy for an encounter. For example, EoEs and Ambrosia drop assignments come from the spawndefs in The Hive or Eden Trial. "Normal" rewards seem to be specific to a critter's rank (really, its archetype, which is correlated with rank), and unless overridden in one of the two ways mentioned, every critter of a given AT does actually have a standard, "normal" reward distribution. Invention salvage is a bit different, as it's defined on a per-villain-group basis, because different groups drop tech, magic or both.

Edited May 6, 2021 by UberGuy

[Ura Hero]

I feel the need to repeat myself here.

 

RNG is RNG. 

 

It is totally possible to go on a streak with no drops for hours on end.  I have literally played all day and not gotten anything for recipes.  Other days I have gotten 10 purple drops.   Random is random.  Is it unlikely?  Yes.  Is it possible?  Yes.  The only way to prove something is broken is by producing and crunching data.  

[TheZag]

If im understanding correctly,  the amounts of drops stays the same as the team size increases and the drops are distributed randomly to team members.  So the larger the team,  the less items you will get.  And on the 48 man raid when you see 2 people link the purple they got,  those were the only 2 drops for the whole raid - not that 45 others didnt link theirs and you got left out.

[UberGuy]

While I agree that "impossible" needs to be used with care, I think we need to be careful not to assume that someone with access to the source code can't know how something works well enough to know that certain correlations cannot be causation. What we can present to devs is evidence, not proof.

 

It's worth remembering that "veteran levels" were added by this team. They actually do know how that code works. If they didn't add a check against vet level to the pre-existing rewards code, it really is impossible that it's checking that.

 

If we use our imaginations, we could perhaps envision a situation where there's an out of bounds memory access going on, and code that should be looking at something else is looking at the vet level integer. However, in such a situation, it's just as likely that the misbehaving code would look anywhere else, would be looking at that integer as if it were something else entirely and do really weird things as result (including crash).

 

In a situation where statistics are involved, the best tests are ones that involve gathering lots of appropriate data. The lower the probability of a given thing (like a purple drop), the more samples you need. The player can do that in a farm or something. Or the dev can do it in a test environment, which gives them access to spawn rates and kill rates players can't achieve. (Having a 0.5s recharge fireball that does 40,000 unresisted damage isn't going to invalidate a drop rate test.)

 

The request for character information is so that the devs can clone the character exactly. Same AT and powersets and badges and temp powers and you name it, and make sure it's not something specific to the character that's tripping up some obscure branch of code that causes an issue.

Edited May 6, 2021 by UberGuy

[Faultline]

1 hour ago, RMS33 said:

For future reference, I was completely floored by hearing the word "impossible" from a programmer in this context. It brought back many memories of company infighting, usually between the coders and some other part of the company. In two words?  Spaghetti Code.

 

In this case, it really is impossible. Veteran levels are a badgestat - it's the same thing that tracks how many times you've been to the top of the Atlas statue, or how many Paragon Protectors you have defeated. There's literally two places in the code where that badgestat is referenced:

• The level up function, which already grants you all the inspirations when you level up, checks if you're level 50 and increases the "overleveled" badgestat by 1.
• The "display player's infobox" that queries it to show the number.

There's simply no way that the "overleveled" badgestat could do anything at all to interfere with rewards; the rewards code is just not aware of it. Spaghetti Code only applies to systems that actually interact with each other, and these don't.

 

I know the next question: "but veteran levels give you rewards". Well, strictly speaking, no, they don't. Those rewards are handled by badges. Every 8 seconds, every badge in the game checks its own requirements to see if it should award itself. The Veteran badges just have a requirement that the overleveled stat be >= their given number in order to award. Badges can then trigger a reward table to grant you other stuff, which is how you get the Empyreans in veteran badges, but also respecs and costume changes in level up badges and a bunch of others. The veteran badges stopping at 99 is irrelevant, it just means that there's no badge that checks if the overleveled badgestat is higher than that.

 

If there was anything that caused badgestats to affect reward tables, we would be seeing all sorts of problems everywhere, because there's almost 3000 badgestats that track different things; overleveled is the badgestat number 2824, sandwiched between 2823 visits.Cimerora2 (for visits to the Battle Hardened badge) and 2825 kills.TheFamily.Lieutenant (for killing Lieutenants of The Family faction). There's nothing special about it.

[overmanprime]

5 minutes ago, Faultline said:

 

In this case, it really is impossible. Veteran levels are a badgestat - it's the same thing that tracks how many times you've been to the top of the Atlas statue, or how many Paragon Protectors you have defeated. There's literally two places in the code where that badgestat is referenced:

• The level up function, which already grants you all the inspirations when you level up, checks if you're level 50 and then rewards you +1 to the "overleveled" badgestat.
• The "display player's infobox" that queries it to show the number.

There's simply no way that the "overleveled" badgestat could do anything at all to interfere with rewards; the rewards code is simply not aware of it. Spaghetti Code only applies to systems that actually interact with each other, and these don't.

 

I know the next question: "but veteran levels give you rewards". Well, strictly speaking, no, they don't. Those rewards are handled by badges. Every 8 seconds, every badge in the game checks its own requirements to see if it should award itself. The Veteran badges just have a requirement that the overleveled stat be >= their given number in order to award. Badges can then trigger a reward table to grant you other stuff, which is how you get the Empyreans in veteran badges, but also respecs and costume changes in level up badges and a bunch of others.

 

If there was anything that caused badgestats to affect reward tables, we would be seeing all sorts of problems everywhere, because there's almost 3000 badgestats that track different things; overleveled is the badgestat number 2824, sandwiched between 2823 visits.Cimerora2 (for visits to the Battle Hardened badge) and 2825 kills.TheFamily.Lieutenant (for killing Lieutenants of The Family faction). There's nothing special about it.

Welp! Well, that's about as definitive an answer as it's possible to get.
So the OP's issue has nothing to do with Veteran Levels. Most likely a run of bad RNG (Occam's Razor and all that) but in the off chance there's something else going on, then the OP should DM Faultline with his character and server details (assuming this has not already been done) and let him take a look to see if there's any other bug or unintended issue causing the perceived problem.
It is gratifying to know that Dev's will taike the (considerable) time to look into issues like this. Big props to Faultline!

[Hedgefund]

People should be forced to cite the RNGods prayer before ever complaining about drop rates changing.  It's far more likely they simply have displeased RNGesus and need to thank him and the others for all the purple bounty they have ever received and beg for forgiveness to atone for cursing them for rolling 95.2 on chance to-hit of 95.

Edited May 6, 2021 by Hedgefund

[Kyksie]

9 hours ago, Faultline said:

 

 Every 8 seconds, every badge in the game checks its own requirements to see if it should award itself.

 

For each and every logged in player, the game checks all badges every 8 seconds? Doesn't this eat up CPU?

[tidge]

On 5/5/2021 at 4:26 PM, LadyDee said:

Once again, never said it was thee cause of the drop rate and I cannot control how other people read. So if someone chooses to start a conspiracy theory (about a game none-the less) well, that is their issue not mine.

 

Suggesting it was past level 99 was the only conclusion I came to after exhausting all other possible reasons. Then when others seemed to have a lower drop rate past said level, it just added more information that this could be the reason but not 100% certain. I'm not a moron, just because something seems to be the cause does not mean it is, but at least it's a place to start.

 

Had you recently changed your costume?

 

Was the arrangement of inspirations in your Insp trays different?