VL Adv Search – Alarming Change In Drop Rates?

[Aurboth]

I wanted to collect some data before I posted again on this subject. Here is the table of Very Long Searches:



The two groups of searches had explorers booked to maximize speed, with the old difference being that S2 was also given a 20% chance of getting 2 adventures per search. In the data sampling, I actually got a 22% occurrence of double adventures.

While I expressed the returns with percentages to 1 significant digit, as there were still only 760 searches in total, it does not have the power to express accuracy to that level. The two groups individually still show a fairly significant range between individual returns, as high as 3% in some cases. This would indicate the need for much larger sample groups to get a useable table.

[NeoOne]

-_-

lol - yes, even first year Statistics courses reveal that for any distribution, the larger the sample, the better the results.

[Perwyn]

Excellent!! I'm glad to finally see someone besides myself posting search results.

Especially since Aurboths results reinforce what I've already said earlier when posting examples of a guildmates searches:

If someone is sending out 18 explorers at a time, they're getting results that are more or less balanced, with a decent distribution of results.

If, on the other hand, someone, like me, is sending out only a few explorers at a time, the results went from more or less balanced, with decent distribution as in my data from 12/6 thru 3/15, to completely whacked with Secluded Experiments becoming a dominant adventure returned.

There's no denying the fact that my small number of results obtained prior to the Intrepid are similar in composition to the large number of results Aurboth has managed to gather in a short time using a small army of explorers. Young Wood Cutter, VLT and Storm Recovery are the dominant adventures returned with Whirlwind coming in at 4th most returned, Secluded Experiments is a low number and all others are somewhere in between. Map frags are in the high 30% range. I don't believe my pre-Intrepid results are a fluke, I believe it was the search mechanic playing out over the course of time, regardless of data sample size.

Sending out a small army of explorers and compiling hundreds of search results, while proving a point about statistics the compiler wishes to make, fails to address the change in the game that has happened to searches done on a smaller scale.

For that reason I will continue to track and post my results. At the end of the 42 weeks it will take to compile 760 results at the pace I'm searching, perhaps our numbers will be similar, perhaps not.

So for me the question still remains:

Why did the distribution of search results change suddenly and drastically with the release of the intrepid explorer for someone sending out a small number of explorers?

Below is the latest screen shot of my results.

[Perwyn]

I may not be a trained statistician, but I have a pretty good grasp of how TSO loot drops work from the time I've spent tracking results of various things in game during the 4+ years I've been playing.

I know with certainty that if I do X number of artifact searches per day, at the end of 4 weeks I will get Y number of Rainbow Snows, give or take 1 or 2. Within those 4 weeks there will be fluctuations, days when I get no Rainbow Snows at all, other days when I get many at a time. But always at the end of a 4 week period, the results are predictable as long as I send out the explorers. That's the way the game works, it's quite reliable that way.

I've logged medium treasure searches, tracking granite returns with the Lucky Detour skill booked. The game is again quite reliable in the amounts returned. The skill says a 24% chance of granite and it delivers.

Then there are adventure loot drops. If, for example, I do 4 Outlaws a day for 3 days running or, do 12 Outlaws over the course of weeks, the loot dropped will still be pretty consistent within the results. I might get a lot of E-logs in a row, but over the course of time, the ratio of e-logs to granite dropped always evens out.

I also know beyond a shadow of doubt from my own experience that when adding one more explorer to artifact searches the amount of Rainbow Snows returned will increase not decrease. And while the number of cheese sandwiches will undoubtedly go up, I won't be getting 3 times the number of cheese sandwiches that I was prior to the addition of 1 explorer.

Looking at my results to date for VL Adventures searches, I still think something is amiss here. I'd like to thank Aurboth for reinforcing my certainty that the numbers in the Prior to Intrepid column are a fairly accurate representation of expected search results. Comparing those results to his, it's quite obvious that Young Woodcutter, VLT, Storm Recovery and Whirlwind are the 4 most returned adventures, Secluded Experiments is a low number of returns and everything else is somewhere in between with map frags being in the high 30% range. The distribution is very similar even tho the percentages vary due to volume.

One phrase I've used repeatedly in these posts is “over the course of time”. My pre-Intrepid results from 12/6 thru 3/15 took 100 days to compile. For the sake of making a suitable comparison, it will take until 6/24 to have 100 days worth of data post-Intrepid. I'll continue to post results about every 14 days until then and let's see where this ends up, using a small number of explorers over the course of time. I also began tracking the Intrepid and the Tavern Explorers separately a few weeks ago, at some point I'll be showing those results as well.

Below is my most recent screen shot:

[Perwyn]

Here's a screen shot of my wifes Intrepid, tracked since the beginning. Results look a whole lot better than what I've been getting with 3 tavs and the Intrepid, well, aside from Lakeside Treasure:

Wait a minute…….. is it possible that since the application of the Wild Determination skill is kinda wonky on the Intrepid, that the RNG for bonus adventures could also be kinda wonky, especially when Trouble Seeker is added into the mix? That would explain why the effect is buried when a large number of explorers is used. Numbers may be changing within the distribution in that case as well, but because of the volume of explorers going out, the effect isn't seen as vividly. Just a thought.

[Perwyn]

Here's the latest screen shot of results. Happy to see I only got 2 more SE's and finally got more Woodcutters. The ratio of SE to Woodcutter is still really off and it's going to take a very long time before SE comes back where it belongs, if it ever does. Time will tell. Will post again around the 1st of June and see where we're at then. (No more SE! No more SE! No more SE! Everybody! Sing along!)

[KingHomer]

Only two of your last 56 searches have resulted in SE. I am sensing an alarming change in the SE drop rate.

Oddly 'nough, your results are starting to look a lot like the pre-intrepid results and normalizing over time, as expected.

Looking forward to future posts... your SE rate should never drop all the way down, as each explorer search is an independent event, but it should continue to move to the middle a little. Of course, there's always the chance you get a few more in a row again.

Good luck with your continued efforts to track the BB rng.

[Perwyn]

I am sensing an alarming change in the SE drop rate.

LOL.... I truly hope so. I won't be happy till the number of Woodcutters is more than double the number of SE's.

Good luck with your continued efforts to track the BB rng.

Thanks. It'll be interesting to see how it looks around the end of June.

[Eriond]

Well, it might just be me, but I have a fleet of explorers, and I send them all out on very long searches (except my lucky's).

I generally averaged around 1 Woodcutter/day doing this. In the last week (maybe since the Crisis event started?), I haven't gotten any.

[Perwyn]

OK sports fans, here's the latest edition of “a funny thing happened to the drop rate when we were out looking for eggs”.

I was getting hopeful that SE's were finally going to be overtaken by either VLT or WC but no such luck. SE continues to be the dominant adventure found overall. Out of 44 additional results since last post, SE dropped 4 times, VLT 3 times and WC dropped a paltry 2 times. I continue to be disappointed in these results and to suspect something's amiss. My results were satisfactory with 3 tavern explorers. The Intrepid was added and all of my explorers results went into the toilet.

It took 127 search results to get 13 Woodcutters prior to Intrepid. It took 208 search results to get 13 Woodcutters after Intrepid. That's the same number of WC's for 81 more or 1.6 times the number of search results. Even if 13 out of 127 was a little on the high side, 13 out of 208 is much too much on the low side.

Out of 127 search results prior there were 5 SE's. Out of 208 search results after Intrepid there are 20 SE's. That's 4 times as many SE's for 1.6 times the number of search results. What?!?!?!?!?

Numbers don't lie, something is wrong here. We shouldn't need to send 18 explorers out at a time to get a decent return on an expensive, not to mention time consuming search. I'll post again on or after June 24th, when I'll be comparing 100 days worth of data to 100 days worth of data.