Hiatus Update
Hours in: 98
Level: 98.07
Hi everyone. Wanted to give an update on this project even though I haven't finished a 15 hour stint to get things up to current. Unfortunately Lindon's runefinding hasn't been contributing to the runeword grail progress nearly as much as I hoped (projected to find a bit over 2 "cubed jahs" by 98 hours in, and got to exactly 1). Then during rfl, i was reminded how different the rate of rune progress feels with a "real" runefinding run
. That, coupled with the vast chasm ahead that is lvl 98 (not to mention first baby due in a couple of weeks
) has just left me totally uninterested in this project. I very well may come back to it at some point, but have no plans to soon. So the project is on hiatus.
However, I did collect some very interesting exp efficiency data before I lost interest, and even had some nice finds. Also wanted to show off his gear for this level, and show a video of what the runs look like.
Finds
Runes: gul
Noteable s/u: azurewrath, boneflame, boneshade, bk ring, eth reapers 208/15!
That eth reapers is my second ever and was quite fun to find!
Gear
As I mentioned in the most recent post, I went with a 402mf setup to make finding stormlash more likely. The best available way that works was with using most of the usual gear and stacking a bunch of mf scs. Well, while it worked ok, my idea of getting 80 fire res through points in resist fire just wasn't quite getting the job done against infector. I would often lose more than half my hp fighting his pack even when they had no dangerous mods (especially after I switched to doing him on p7, which I will discuss below). Which wasn't too dangerous, but it meant my potion usage was far too frequent for my taste. So, as you can probably guess, I went back to rising sun. Hurts the hammer dmg a good bit, but the runs are still almost as fast and the comfort is much, much better. See the video below for the full gear. Same mf and similar life. I didnt respec to remove the points from resist fire, so the defense could be higher.
Maybe a bit rusty in that video, and the game doesn't usually lag like that nor have such long load times. I guess OBS was to blame? Idk. But that's about how they look anyways. The shrine checking and playersetting will be discussed below.
Exp Efficiency
Shrines
So a few interesting things to discuss. First is my shrine checking behavior. I have the 3 shrines "on the way" in chaos, which I would always check, and which have an 8.2% chance of being an xp shrine. Then I have up to 5 a1 shrines I can check on a somewhat reasonable route, all of which have an 18% chance to be an exp shrine (see
here for shrine chances). But it takes a fair bit of time to check them all, so I wondered if it was worth the time, or if only checking a couple was best or what. So what I did to analyze that was time myself doing runs checking 0, 1, and all 5 act 1 shrines. Did about 20 runs of each of those, and then I counted the number of extra teles (as compared to checking 1 a1 shrine) for checking 2-4 a1 shrines and calc'd the projected time from that. Ended up looking like this.
Code:
a1 shrines runtime
---------- -------
0 82.80
1 85.90
2 89.14
3 90.94
4 92.74
5 92.94
So then with that data, its a bit complicated, but we can calc the overall chance to get an xp shrine, and adjust the runtime based on which shrine you check is xp (as in, if the first a1 shrine is xp, you won't take the fill time of checking 5 shrines). Looks like this for the case of checking 4 a1 shrines for example.
And then finally, here is all options summarized
So as you can see, it is indeed worth it to take all that extra time to check all of the a1 shrines, despite it adding more than 10 seconds to the run! That was a pretty fun calc!
Playersetting
Now, even more interestingly. The above numbers are all for p1/p7 runs. Where I hit infector and de seis on p1 and then switch to p7 before hitting vizier and ofc diablo. I guess that is slightly less efficiency than doing vizier on p1, but I just can't stomach losing out on extra rune drops from his minions to save 0.5s per run. Just for comparison, and just to see "how much worse" it is, I decided to also try just doing the whole run on p7. Imagine my surprise when it turned out I was getting better exp/hr doing that! This definitely was unexpected.
Those minions give so little exp compared to diablo that I thought they were irrelevant. But after I actually calc'd it out, I figured out your runtime has to go up by at least 10% when you do all p7 runs for p1/p7 to be better. It varies a bit based on how many shrines you check, but p7 is even a little better compared to p1 if you check less than me (for example would have to be 12% faster on p1/p7 if I only checked the three a4 shrines). In my case, I only add about 5-6s to my runtime from doing everything on p7.
So all p7 is definitely better for me, and I suspect would actually be better for most everyone running hdin. I guess p8/p7 would also need to be tested for completeness
. But meh.
Now to be fair, I am much more likely to die from p7 die seis than p1 de seis, so thats a factor in favor of p1/p7, but whatever, Id rather have better exp when not dying, better rune drops, and zero changing of psetting for convenience.
Here is the projected exp/hr chart per shrine when my runs are adjusted up 5-6s (based on runtime testing).
Much better! And lastly, I did track this method, while running with a lot of focus for an hour and got this result.
Code:
all p7, 3xa4, 5xa1 shrines
---------------------------
start runs 4,886
start timer 93.47
start exp 3,241,266,520
end runs 4,929
end timer 94.59
end exp 3,244,672,486
runtime 93.77
xp/run 79,209
xp/s 845
xp/hr 3,041,041
So broke 3million as predicted
. I thought getting >3m would be easier tbh, and also thought that I would be able to get faster p1/p7 times, but I'll take it. Still though, not neeeearly as nice the exp/hr I was getting with my sorc from nihlathak (was 3.7-3.8mill). I'm sure I could make it slightly faster by running with more dmg and less mf, but I doubt it would make too much of a difference.
@ffs @PhineasB @Neksja did you guys ever compare p1/p7 vs all p7 for lvl 98 chaos hdin? And have any of you seen anyone get significantly more than 3m xp/hr with one? I think Corrupted had something similar to me, but not sure how many have tried to push it.
Ok that's all from me. Hope everybody is doing well and has a nice weekend!
.
Google