2021 Spring RFL Results thread: Moo Moo (Round One)

[Luhkoh]

@D2DC lol that would have been embarrassing, but since you won, looks like it was just the correct strategy .

As for map drop tendencies, I'm not someone who buys into that kind of thing easily. But I'm just gonna come out and say it. I believe it too for sure haha (plz dont tell fabian ). Not positive its specifically tied to map seed, and can't say i've ever seen it with a specific rune type. My observation has mostly been the same boss type in the same location dropping the same base item. But who the heck knows?

@Nano Awesome thanks so much for all that nice data! And awesome drops. I'll have to give necro a more serious go next year. 400+ cows per run sounds so nice.

 

[maxicek]

Well in 1.00 Classic it very much is a thing with drops and seeds. But did they actually fix it properly or just mask it and make it less visible?

 

[ffs]

I for one think it’s all cognitive bias never noticed anything unusual, it’s just that you pay much more attention to those potentially rare drop bases. I always see these theories when it comes to sacred armors, unearthed wands and whatnot... but for some reason nobody seems to be suspicious of GMBs or Shako or scarab husk bases dropping frequently.

 

[Luhkoh]

@ffs for me my observation has been more like, when over the course of lets say 50 AT runs, a certain specific boss spawn (location and monster type) drops something like a colossus blade base maybe like 6-7 times, and you dont see any other boss pack drop one. That's just too unlikely to be a coincidence imo, especially when I've seen it many times. But that being said, it never seems to last for a long time that I've observed, so I'm not making claims like I've seen elsewhere that you can get a "dimensional shard map". Though I'm not ready to dismiss that possibility either.

 

[D2DC]

I have the same experience @Luhkoh. Certain boss spawn points tend to drop the item, but nowhere else. Do monsters have their drop items (or at least TC) determined when they are spawned, or when they're killed? Maybe the kill only rolls for S/U/Magic/Rare?

Anyway, I and Charles are convinced. He is one of my oldest friends. Ive known him since we were roommates at Harvard. Infern ticked lvl 91 just before completing his second set. So, hopefully he has 4 more levels of Jahs and Ohms with the occasional Pul or Um sprinkled in..

 

[Pb_pal]

I am curious if anybody else notices any drop "tendencies" along these lines? Or, do I just get a bit loony when doing target running? Perhaps, I am the only one that can see it because of a chemical that was released into my bloodstream when my radium diode implant dissolved?

I have waffled back and forth on this one for years, borderline convincing myself of it many times. I think it's *probably* just that we only remember the memorable patterns, but I just can't shake the feeling that map seed does correspond to increased chances of certain drops.

Our RNG overlord certainly works in mysterious ways.

 

[saracen85]

^^ I'm sorta convinced about this "mapseeding" too. I had a mausoleum map where a Gheeds Charm dropped about 5 different times, but always at the same bosspack spawn point.

 

[Nano]

@Nano Interesting finds, because while I didn't use both methods simultaneously, I came to the opposite conclusion. Pretty sure that in my case drop method estimates from previous years rather over-estimated actual kills. I only used the drop method with runes, though, so somewhat less precise. But for me, run counter numbers seemed to be much more in line with the XP methods estimates I've had in the past. Those (run counter & XP method) were around 150-170 P5 kills/min (for good cow builds), whereas with rune counting in recent years I arrived at some 180-200 (and occasionally above 200 with smaller sample sizes). Seems way too high, at least for a long-term average.

You sure you picked up something like 95% of those item types? Based on ~2.5 minute run times I'd have assumed that you only picked them when they dropped close by (that's how I approach it most of the time in any case), in which case the difference makes more sense as well.

I use the SC+GC+Jewel drop method since I can gather that data over the entire 20 hour set and have a large sample size to calculate from. I use drop method mostly for post-round analysis. The difference is I generally don't do much testing prior to the round (10 runs or less), so I don't count runes or anything to assess builds early in the round- I feel I wouldn't have sufficient sample size. My early assessments were based on time only or overall feel of the game play.

Not sure why your rune drop method estimates were consistently over. With drop RNG, there should be some underestimating along with overestimating. Then there is the issue of not seeing drops which will skew the overall data slightly to underestimating.

Starting from the 2018 RFL, I tried to be thorough in collecting SC, GC, and jewels in order to have more accurate data, even if it cost me a little bit of time to reposition and scan large item piles from bottom up. If I see the drop, I'll pick it up.

@Nano Awesome thanks so much for all that nice data! And awesome drops. I'll have to give necro a more serious go next year. 400+ cows per run sounds so nice.

Necro is a very thorough clearer with almost no stragglers left behind, can recommend if you like the play style. My cow map with King next to the portal helped as well.
Looking through the data, I even had 4 runs in the 600+ cows region, but those runs are also in the 3:00-3:20 time range.

Kills on y-axis and run number on x-axis. That run under 100 kills was probably an early death..

 

[ffs]

@Nano Loving the precise data. Interesting to see the peak numbers, too. I did see a few runs north of 580 kills which was way above average, and wondered what the limit would be.

Not sure why your rune drop method estimates were consistently over. With drop RNG, there should be some underestimating along with overestimating. Then there is the issue of not seeing drops which will skew the overall data slightly to underestimating.

What I did was just counting (not picking) all runes. Highly unlikely to miss them, and by not picking them up unless they were qualifiers, it didn't compromise run times. In 2020 and 2019 I used this method and ended up in the ~190 P5 kills/min neighborhood for the top builds (Java, Trapper, Necro). In reality I don't think any char actually does that, even on perfect maps and with perfect play.

As to why I got these numbers – no idea, other than just RNG/variance indeed. I did this for entire RFL sets, so the sample size was not an issue. But of course counting runes is less precise than picking all jewels/charms/runes. Thing is I always figured I'd either lose too much time in picking up everything, or leaving just a few too many of them on the ground so that the result would become less reliable. That's why I suspected the discrepancy in your case (having picked up tons of jewel/charms) was down to missing jewels/charms in an amount just sufficient enough to dilute the result a bit.

Anyway, fortunately thanks to the run counter we won't have such issues going forward.

 

[Luhkoh]

@Nano awesome, love the scatterplot. With regard to necro, I love the playstyle, but i HATE dying and losing the insight iron golem which happens to me more often than I'd like. I kinda want to see how close I can get to my javazon's number with an insight merc . Also did you take any videos of your runs this year? I'm always impressed how you seem to get such good times despite clearing thoroughly.

@D2DC who the heck is Charles? An SPF'er?

 

[Pb_pal]

Also did you take any videos of your runs this year? I'm always impressed how you seem to get such good times despite clearing thoroughly.

Such is the way of a machine.

 

[D2DC]

@Nano awesome, love the scatterplot. With regard to necro, I love the playstyle, but i @D2DC who the heck is Charles? An SPF'er?

Oops.. not Harvard. We went to Princeton together..

Thats Charles in the middle with my niece Marcee.

 

[art_vandelay]

I can't tell if you're trolling. Coming in here with your 4 jahs and fancy ivy league degree and mocking us mortals

 

[D2DC]

4 Jah and map seed affecting drops = real.

Charles and Princeton = imaginary.

 

[maxicek]

@Luhkoh No need for an insight IG, Mana per Kill works better. I used a 1.07 ring, but you can get it for Silkweaves and jewels instead.

 

[Babyhell]

Well on my side, results were packed at the beginning, in the first 29 runs:

Given my available time, I haven't time to edit more than that. Hopefully a full set incoming? #JustBelieve

 

[Nano]

@Luhkoh I didn't record anything in the 20 hour set but here's a 4 run demo video with gear/stats at the end.

Total kills: 1532 Cows
Time: 9:01
Cows/Run: 383
Cows/Min: 170

Below average cow density in these 4 runs but faster times make up for it. Cows/Min is only a few cows difference from my 20 hour average. Like maxicek mentioned, I use MPK for mana recovery.

 

[maxicek]

Very interesting Nano. You have a similar map layout to me. I didn't use Revives but I did use Attract. I must have put the point into Revive after seing your build last year

 

[CaseyJones]

Holy crap that's awesome. My nec is still under geared (no enigma, single facet) but never considered the revives....looks like they really take the heat off.

 

[Luhkoh]

@Nano thanks so much! I should have given mpk more of a chance. I thought it just didnt get the job done with no time travel gear, but never tested it thoroughly. That's a gamechanger for me. Glad I asked . And yeah revives rock!