Guild Wars 2

Hello. My name is Rain Spell and some of y’all may have seen me running around doing fractals all day. This is a long post. Tl:dr at the bottom, but it’s better to read this whole thing.Dropbox link to data.

In early October, I began keeping track of rare drops from fractal end chests. I began with the intention of proving to a friend that fractals could be worth running, in comparison to a dungeon (via ectos). I noticed some numbers were coming up strange, so I switched from collecting data per tier to a more specific, data per level per tier. I do not have the expertise to fully compile and analyze the data I’ve collected, but I did notice a not-so-fun trend.

For our purposes, “Rare drops” are rare, salvageable equipment gained from end chests at the end of each fractal. The final number of rares at the end of the set is taken, to help eliminate some of the variation from different fractals. (I considered throwing out fractals that included the Molten Testing Facility, because of its notoriously awful chest, but decided that that was too much manipulation.) I did not count exotics, because I did not want to mess with a loot table that was so divided. (There are 3 possible exotic types you can get from chests, Tribal, Karma, and Standard, as far as I know.)

Assumptions:

• Magic Find does not affect end chests (I believe this is true, even if some claim otherwise. I noticed no major variations between friends who had less than 100 and those that had more than 200 when it came to end chests.)
• Reward level does not vastly effect end chest rewards. While I’ve heard suspicion that reward level affects end chests, I believe it only affects the last shaky chest.
• Participants were truthful and watchful of which rares were from mobs and from endchests. They did not salvage their rares until after they had given me their count.
• Primary assumption: Any human errors are within the realm of reasonable deviation, since the sample size is so large.
• A sample size of 50-70+ per tier is large enough to see clear trends. (In fact, it was possible to see the trends at less than 20)
• Fractal rewards increase per fractal tier/level being completed, and are affected by both, positively.
• “Luck stealing” does not exist. One person in the party receiving 8 rares has no effect on the other party members.

Continued.

Guidelines:
Participants who submitted data to me were given the following guidelines. When there was doubt on whether their numbers were correct (did I get that rare from a mob or a chest?), that particular sample was thrown out. Cores and Lodestones do not count. Aetherized scraps and similar do not count. Only salvageable rare equipment. (Please remember, I started this with ectos in mind.) If they received the rare from a mob, and not an end chest, they were to either mark it down, tell me to subtract it, and/or salvage it immediately so it would not affect the end count.
If there is anything unclear, please feel free to ask. I will do my best to justify my thought process and my procedures.

Limitations:
General unwillingness for certain fractal levels via painful mistlocks, so there is some lack of data on a few higher level fractals. Results may be especially skewed toward higher levels in the 31-40 tier and 41-50 tiers. Also, I am an amateur. I am not a statistician. This may not be the best method to compile this data. I also have no idea which tests to use to check for errors.
My guild is relatively small, and while several of us love fractals, we are not a fractal guild. I would be interested if someone else wanted to continue this particular thread of data collection on a similar or larger scale.

Findings:
Together, my fractal running friends, cooperative pugs, and my guildmates have put together 675 data points. We did our best to spread out our data throughout the tiers, though we ran into some roadblocks when it came to the more tedious mistlocks. Pug responses were used, though only when it was clear that they understood the guidelines above. Otherwise, their responses were scrapped. (I’d say less than 20 samples needed to be scrapped, not counting those from disconnect/no reconnects.)

Our results were almost as expected, a general trend going upward from around 2.5 rares at the lowest tier (1-10) to 4 rares at the final tier (41-50). The anomalies that led us to attempt more thorough investigation and collection were the 2nd and 3rd tiers. Their “rare” loot tables appeared to have been swapped back when we only had about 20 samples per tier. While there was a general positive trend, the rare count decreased significantly enough to warrant a second glance when you reached tier 3. So we increased our sample size. With the loot table as it is, you are rewarded slightly less for doing a 3rd tier than a 2nd tier fractal. We did not collect enough individual data to make a conclusive decision on whether higher level fractals in the same tier truly give better loot than lower level fractals, but there are some interesting splits worth looking at. (One friend insists her best rewards come from level 22s, though the data suggests 23s as much more profitable than 30s.) Please see the provided data to draw your own conclusions.

Continued.

Data: How to read this spreadsheet.

Sheet1: Contains my initial Tier-specific data at the top. The lower columns are the compiled data points from the whole project (second phase level specific data included). There are small notes to remind myself of what goes where, and some random dates when I was still marking what day I had done a fractal on. The lower right section contains the master chart with the final averages and some useless data I thought would be more useful when I first began. I’m leaving this in such a way so that if I missed some useful opportunity, someone else might be able to spot and utilize it.

Sheet2: Rows1-51 are all collected, Level-specific data, where row 2 is level 1, and row 51 is level 50. Below that are averages associated with their respective levels. Averages are highlighted green if above 3 (what I arbitrarily decided was a nice number of rares to get from a fractal) and yellow if they are less than or equal to 3. The color blocks were to make it easier to differentiate the tiers. Each Tier has its accompanying sample size. Each Level lists its sample size. The final column of data in that subsection is the Standard deviation, which I have no idea how to use. If the Std. Dev. was above 1.5, it was highlighted red. If it was below 1, it was highlighted green. I considered not bothering, but I figured someone would be interested.

The 2 graphs to the bottom right are the average loot table currently, and the average loot table with the 2nd and 3rd tier tables swapped.

Graphical Analysis and Conclusion:
It looks like keeping track of these fractals has given a small difference of less than .2 rares between the 3rd and 2nd tiers. Still, it does show that there is something strange with the current loot tables, and they are not as straightforward as higher level=higher reward. Swapping the two tables creates what I could see as a more reasonable curve. The slight spike between Tier 1 and Tier 2 could be attributed to giving higher rewards on purpose due to AR checks and other prerequisites. Rewards do get better the higher your tier, but not necessarily by level, and not for tiers 2 and 3. The standard deviation between levels is seldom below 1 and usually around 1.5-1.8.

Completing all tiers of fractals in a day would net you almost 17 rares. My groups tended to run 35-45 minutes high fractals, (21-50) while running 25-30 minute low fractals. (1-20) So total estimated time is: (45+45+45+30+30)/60=3.25 hours at the most. If you estimate 1.5 ecto per rare, you make 25.5 ectos per 3.25 hours or around 7 ectos per hour. Gwtp.net places this at 2.40G per hour (after tax)+the 1g to 1.5g base end fractal reward. To be around 3.4 to 3.9g and hour. This is still not taking into account the high chance of exotics at later levels, nor the chance at ascended chests. Still, while the gold per hour is not as high as a standard pug dungeon speed group, there is one thing fractals have that dungeons don’t: a massive (up to 10k) karma bonus. In the end, I would place fractal rewards as on par with dungeon rewards, just because of the karma.

Thank you [vVv], my wonderful guild, for being awesome and supportive and contributing the lion’s share of this data. Thank you, my fractal friends, who groaned whenever I asked for another rare count. Thank you Kendall, for keeping me motivated and on track in the initial phases, when we realized something was up. All of y’all made this worth it. And hey, it’s done! Woohoo!

Perhaps I should have kept a separate set of statistics for exotic drops,as they may have an effect on getting other drops, but I don’t see this as likely. Maybe next year.

Tl;Dr: Something’s strange with the Tier 2 and 3 fractal tiers. Namely, it’s seems they’ve been swapped. But it’s not much to worry about in the long run. Probably?

Hey, I remember you! Helped out with one of your runs. I’m excited to see this data. After we got molten facility you made me realize for the first time how bad that chest was. As an addendum, the bonus chest in the shaman fractal seems to be linked to other chests in game and even in game events. For example it was dropping winter’s day items recently. Thanks for the data and I hope something comes of this!

Thank you for this, I knew the rewards werent THAT bad

On a somewhat related note, are people still getting bonus items for completing a set of 4 Fractals? I haven’t received any for the last 5 or 6 runs (just the Pristine Relic, Empyreal and gold); I’m unsure if there’s been a stealth nerf or whether I’m just having a really bad streak.

Zaxares.5419:

On a somewhat related note, are people still getting bonus items for completing a set of 4 Fractals? I haven’t received any for the last 5 or 6 runs (just the Pristine Relic, Empyreal and gold); I’m unsure if there’s been a stealth nerf or whether I’m just having a really bad streak.

Bad streaks happen man. Remember, you don’t get rings/ascended equipment until at least level 11. I’ve gotten nothing from the shakies at least 4 times in a row before. Same night too.

Yeah, those were all from the 11-19 range. I know that <10 no bonus items drop.

There is 1 thing that don’t work in your stuff. If you are using mystic/master salvage kit on your rare. You shouldn’t get 1.5 ecto per rare, but 0.875.

That drop the gold/hour to 2.7-3.2 if i’m generous. But let say that you have the right numbers. That’s still 2-3 gold less than dungeon with decent pugs, 5 gold less than with a organized group.

If someone like doing fractals or need Karma, fractal is a good place. But it’s still lacking a lot in term of reward since Karma drop a lot and don’t have that much use except if you want to do a lot of legendaries. Its even worth now, since there is a lot of alternate source of Obsidian Shard other than karma.

The drop of ascended chest/weapons and fractals weapons is a nice addition that could make fractal as rewarding as other stuff in the game. But the fact that you most likely will get useless stats or fractal weapons that you already have, don’t make it that rewarding. If they could change it so we change receive ascended chest and fractal weapons chest and you could choose what stats and piece you want, now fractal could be rewarding enough. A bit like these two ascended armor chest that you receive after the ambrite and luminescent collection.