[Research] Map increased item quantity (IIQ% ) and its role in map sustainability

It is a nice data set and a really solid effort on your part to give some method to the madness of maps. POE has a great community and these contributions are a big part of it.

A) You cannot disregard lower level maps. Maps are a spectrum and whether they are useful to you or not the low maps are apart of this spectrum. Even if it just increased the amount of lower maps you found IIQ would still have an impact on the amount of maps you are finding. Analyzing your results while ignoring relevant data points is at best a very skewed analysis. At worst it provides a completely incorrect conclusion. As there are a finite number of map tiers, there is really no reason to disregard any of them. The tedium of the data log lies in actually making the log. If you are already logging it, there is no reason to not make it as accurate as possible.

B) Your test is fundamentally flawed. The inherent map seeds are unknown to the player base but we can infer, based on GGG's cryptic information, that the higher tier maps have a lower chance to drop maps when compared to lower tier maps. If the goal is to strictly analyze the effect of IIQ on map drops a good test would compare only same level maps with different IIQ's. The "best" test would compare only identical maps with different IIQ's. Simply put: different maps, different level maps and different IIQ's is not a solvable equation. Drawing a conclusion on IIQ would require making at least the map level a fixed variable.

There is good data here, it just does not really show anything in its current state. I suspect there is in fact a pattern to all of this (IIQ/Map level/Map Drops) but the data and analysis are rigorous enough to draw the conclusions you have made here.

Anecdotal evidence: My own anecdotal evidence has shown me that chiseled blue maps are very sustainable in mid-tier (74+) maps. That being said, the additional currency/items found in higher IIQ maps is well worth the added risk and investment. This would obviously only matter to players interested in more than just leveling their characters and maintaining a map pool.

ScrotieMcB did an interesting thread a while back about streakiness in map drops (PSA: Understanding Map Drop Streaks) which somewhat agrees with the data you've provided here. According to their model (which, disclaimer, was only a model) quantity has little effect on whether a map "breaks even" and drops a map to replace itself. Instead, the gains go toward getting multiple drops per map, but even then, there's only a (calculated, not empirical) 20% difference or so between a 20% quant map and a 100% quant map. Given your small sample size compared to the likely drop rate of maps, it's not surprising that you didn't detect much of a difference.

Their model didn't take Pack Size into account, but assuming Pack Size acts as a separate More multiplier on map drops, it's not too hard to do the math. For example, a map with 50% IIQ and 30% Pack Size should drop about as many maps as one with 100% IIQ (1.5*1.3=2). A map with 100% IIQ and 15% Pack Size should drop as many as a map with 130% IIQ (2*1.15=2.3), or, alternately, it should drop about as much as a map with 75% IIQ and 30% Pack Size (1.75*1.3=2.3). In general, since mods that give Pack Size also give IIQ, it makes sense to prioritize Pack Size.


However, neither of these analyses take Bloodlines or Nemesis mods into account. I'd be very curious to see what effect those mods have, since they both increase drops (likely as a separate More multiplier) AND they increase the chance for a +1 drop or +2 drop. Since each +1 drop is worth 3 +0 drops for map progression purposes, it might make sense to roll for those mods.

"

Dos_Fafner wrote:

B) Your test is fundamentally flawed. The inherent map seeds are unknown to the player base but we can infer, based on GGG's cryptic information, that the higher tier maps have a lower chance to drop maps when compared to lower tier maps. If the goal is to strictly analyze the effect of IIQ on map drops a good test would compare only same level maps with different IIQ's. The "best" test would compare only identical maps with different IIQ's. Simply put: different maps, different level maps and different IIQ's is not a solvable equation. Drawing a conclusion on IIQ would require making at least the map level a fixed variable.

i don't think this is true.
ppl were complaining that high tier maps didn't drop enough maps in the range which probably has more to do with more potential maps being able to drop (# of various maps/tiers) the higher you get.
meaning: in a 68 map you have a 50/50 (%) chance of dropping either a 68 or 69. in a 78 map your chances are 1/12 to drop a specific tier.

i THINK what ggg did was to increase the chance of getting maps in your range (or lessen the chance of getting way too low maps. whatever).


Also OP said "sustaining" so i think the range he applied for his test is appropriate.

another thing is: you assumed that the same map always has the same amount of mobs. this is most likely not true. also we dont know if different mobs have different chances of dropping maps (spawn of seawitches for example). you also can't influence the amount of strongboxes and shrines that randomly appear. so OP did the correct thing and used the average which also reflects the "reality of mapping".
also if you can't solve an equation you use data points to approximate the solution/function. another thing the OP did right here.

there is no need to whiteknight map drops. the system is flawed. but it seems to be a good way for ggg to force players to spend all their currency on it.

PS:
a fixed variable is called constant.

I know this is a lot for me to ask; I think you should log as many data points as possible. Maybe some data is irrelevant to your personal experiment, but maybe someone can find patterns in your excluded data.

That said, maybe we can use mods to help you with data collection? We can keep each individuals data separate from each other. But, if we can establish a reasonable method for data collection your sample size will grow much faster than if you are lone wolfing.

"What this research shows is that GGG is likely to have a major bug in their programming of the IIQ in connection to map drops. Rather than relying on the programming yielding the intended results (which can be hard to even verify for GGG if the code is very large), they should do tests in practice and then check whether or not the results match their actual intent."

Well, I do not know about ma dro that much, bur most of the time I get gems with quality - the are 6% ... I feel cheated like about 75% of the time. I do not know if this is wanted, but I have a strange feeling that either it is a bug or it is intended... :(

I'm currently working on an .xlsx model that would be fairly easy to fill out for an average PoE player. This model is likely to feature:
- all map tiers as opposed to mid-tier that I implemented for my personal research,
- all map drops as opposed to the -2 to +2 range,
- certain modifiers like magic monsters,
- stats reflecting map returns and sustainability per quantity, pack size, magic monsters, map tiers and tier brackets.
And probably many other things depending on the number of samples I'll get. If I manage to convince a few players to fill out the map drp form on a regular basis I might end up with a few thousands samples which will make the research more valid.

The issue is that IIQ in the short term is subject to a lot of rng and players have very selective memories.

When you say gains are linear, Are you saying a map with 100 Q has twice the chance to drop a map as opposed to a 60 Q map and so forth?

"

Chris wrote:

This thread has been linked to us a lot of times today and we're worried that it might be spreading misinformation.

You said "Not all map drops were being included in the research data", and then conclude that higher IIQ doesn't help much. Can you post your raw sample data that includes all map drops? IIQ is linear on items and maps. Note that the map boss has a chance of an extra map, which can bias the data if you don't take it into account.

Our concern is that people will read the thread and think that IIQ doesn't help (when it significantly does, as your data should show when you look at all the drops) and then receive fewer maps, further increasing the complaints about drop rate.

"Nothing is more damaging than ignorance wielding statistics"

I will clarify what I mean by this.

How this post is being used by the community in general is very much wrong and incorrect. Or what I have seen on reddit specifically, (some comments here as well)

The post itself is very well done (although academically it should be more clear in its "selected" findings IE it should talk more about exclusionary factors)

I agree with the dev's comments about posting the full data so that a more proper review (vetting) of the findings can be done.


This is why I love POE :) lots of data and "research" :)

"

Weißenberg wrote:

I'm currently working on an .xlsx model that would be fairly easy to fill out for an average PoE player. This model is likely to feature:
- all map tiers as opposed to mid-tier that I implemented for my personal research,
- all map drops as opposed to the -2 to +2 range,
- certain modifiers like magic monsters,
- stats reflecting map returns and sustainability per quantity, pack size, magic monsters, map tiers and tier brackets.
And probably many other things depending on the number of samples I'll get. If I manage to convince a few players to fill out the map drp form on a regular basis I might end up with a few thousands samples which will make the research more valid.

I would personally LOVE to be apart of any data you collect. If that be by form or plugin.

If you ever put up a google doc or submission form Please post it in the thread! :) I would love to be apart of it.