Point I tried to make is there might be some magical runtime when s/e time goes out of control. Is it around 50 seconds or below, I can't say. Same issue as with Pindle or Travincal, although I know not all players have those. I have s/e issues regardless of amulet in pit if playing fast. Things are much worse when rolling for new pitmap if I do small teleporting looking for fast entrance to pit. After clicking save, I can comfortably go get some cup of coffee.
That s&e lag is weird. When i switched to -ns, the runs became way faster, because the lag is pretty much vanished. On the other hand, it is very unpredictable. It only last for like a second or 2 at most now, but it still happens with some randomly long runs. Sometimes i get no lag when i do 48s runs, sometimes i get like 1 second. Sometimes i get the lag with 56s runs, sometimes not. It's very uneven. You can see it in my HoR entry video i posted today.
Hi! One small question, if I have the same hork % on both kill and caster switch, is there a point in horking with caster whereas I don't get benefit from the 6 Ist PB?
Thx for all the information I collected here on the char!
The point is that run times > everything else. Horking on 105 FCR switch just results in faster runs. That's not only the faster hork cast animation, but also being able to switch much more rapidly between teleport, hork, howl and FI... whatever the situation demands. 180 MF sounds like a lot but (assuming reasonably fast runs) saving just 2-3 seconds overall by horking on caster switch is more efficient than horking with higher MF.
This ccl, is it true according to your experience even if you play on-line and the maps change everytime, hence, you loose some time not knowing where to go precisely, hence, seeing less uniques overall?
No that's assuming fixed maps in SP, but frankly most of the theorycrafting on this build is centered around SP. In fact most of the information on these forums is... as indicated by the forum name. When playing online it may be beneficial to pay more attention to MF, although that's purely speculative as I have no idea of how long-term seconds/boss kill numbers look like when playing on random maps.
Haha, yes, sure but since nearly 100% of the guides or variants of the pitzerker I saw come from here, all of them are just relaying the same info (based on SP so!). Hence, I don't have much info for BN players
Using a simple spreadsheet factoring mf and hork%, I found that with my current items the build generates 96.42% of the uniques of the HLW/arach meta build. To match the efficiency of the meta build, the 200 fcr variant would need 3.71% faster run times.
From 105 to 200 fcr, the barbarian's cast animation is reduced from 8 frames to 7 (320ms to 280 ms), which results in a 14.29% increase to cast speed.
In a pit run, the barbarian is either idle, running, attacking, looting or casting. With 25 attacks at 9 fpa, attacking takes 9 seconds (HLW would save about 0.85 seconds). With a generous approximation of 11 seconds spent idle, running and looting, only 20 seconds of the run are dedicated to actions other than casting. That means for a 65 seconds run, the barbarian would be casting for 45 seconds, which can fit in roughly 141 casts at 105 fcr. The same number of casts would require 39.5 seconds at 200 fcr, for a 5.5 second save. That's a 8.4% faster clear time, far above the 3.7% that was required to match the efficiency of the 105 fcr setup. With 35 seconds of casting and 30 for actions other than casting, we would still have a 4.4 second time save, or a 6.8% faster clear time.
These approximations suggests the 200 fcr setup could be optimal for pit zerking.
In addition to gaining 1 frame on cast switch, the barbarian gains 2 frames on kill switch for cleaner telestomps or the occasional howl/hork in the middle of combat.
I think the biggest downside is the fact that with 280 ms cast animations, you're getting dangerously close to a human's reaction time. This can lead to backtracking and inefficient teleports resulting in time loss. This issue can be overcome with practice, imo.
@suddenZenith it's an interesting idea, to me it seems very similar to the 105 FCR Grief switch approach (with Shako, HLW, Arach, Wizzy, FCR gloves and 2x FCR rings). This too results in faster run times and is very close in efficiency. During small sample sizes of ~50 runs or so I even surpassed it at times. However long term numbers always favored the typical setup with high MF Grief switch and 105 FCR cast switch.
I haven't tried 200 FCR on a Berserker, but did some messing around with a Singer on 200 FCR on a dedicated switch (whereas Singers are normally played on 1 switch only). Didn't outperform the single-switch 105 FCR setup either though.
Also tried 200 FCR on a bunch of different Sorcs, and I found they didn't actually get that much faster run times with 200 FCR, even though they do nothing but casting.
The practical improvement in run times was always way below the theoretically expected improvement (as in saved frames * number of casts).
I think the biggest downside is the fact that with 280 ms cast animations, you're getting dangerously close to a human's reaction time. This can lead to backtracking and inefficient teleports resulting in time loss. This issue can be overcome with practice, imo.
That was exactly my observation whenever I ran something with 200 FCR: Any time saved from faster casting was then lost again by having to make 1-2 extra teleports after flying by something too quickly. Maybe someone with "young people reaction times" can handle it better though.
Every single time I did the math I realized that uping the breakpoint is "correct" thing you want to do in terms of efficiency, but every single time I tried it in practice it ended up same as lower breakpoint. I tried going for 105 fcr breakpoint on AT sorc, realizied that with 63 I actually have better performance overall and higher MF meant that that I couldn't blame lack of MF when that yellow Unearthed Wand dropped.
I tried 200 fcr pit zerker ages ago and it felt far too hectic, but then again I was never even remotely good on that spec, 105 or 200 fcr.
It's not too hard to get sub 280 ms consistently but doing it in variable environment with more than just ultra focus on single task just isn't doable, at least for me. Probably has something to do with bad case of old.
I think one more elephant in the room is items. I don't know a single person on SPF that has jewelery that wouldn't be bleh for 200 fcr build and zerker is one of more gear dependant MFers out there. Like 2/20 amulet with MF probably doesn't exist on these forums? Also fcr rings which you need two and are much harder to get. I think best place to get proper fcr rings is pre LoD Classic patches but they are somehow discouraged around here for god know what reason. HLW and decent Nagels are extremely hard to beat with rares/crafts and super easy to get.
And last thing are those loading screens. The faster the runs the worse are times on saving the game. This is probably the worst thing as there is upper limit on optimizing for speed. I think someone managed to get by save times by running on Linux or Mac, but at that point we're approaching speedrunner tier effort (which would be interesting to see, but I think this community is far too casual for something like that nowdays).
If you read it on the SPF that was probably me, I used to play natively on a Mac before macOS stopped supporting it. That being said, D2 generally runs smoother on Windows. For Pit runs specifically, my run times actually improved despite the terrible S&E delay issues I'm getting there.
Indeed the gearing for 200 FCR MF Barb is challenging as well, the drop in MF probably will add up quite a bit unless you have 2/20 circlets and amulets with at least some MF. In all likelihood, it's not like swapping in HLW/LoH etc. for negligible MF losses.
But as a general found I always found it hard to capitalize on cast speed improvements in general for MF runs (as opposed to for example Travincal or Cows). Just because they also reduce the error margin.
A misclick/bad decision that costs you 1s with 55s average runs is already more painful than the same misclick/decision with 60s average runs... and that effect then slightly increases due to MF (will still be lower no matter how good the FCR gear).
With the aforementioned 105 FCR Grief switch I noticed that clearly. You do away with weapon swapping, which improves run times quite a bit and really does make up for the MF loss if execution is spot on. But I wasn't able to capitalize on the faster runs enough. Small inaccuracies here and there over the course of many runs made me lose the runtime advantage -- just because such inaccuracies have more of an impact the faster the runs are. Overall the typical Grief + 105 FCR switch combination always came out on top in my case.
Still, I'd be curious to see how a 200 FCR Berserker plays out in practice in D2L. Atm I don't have much time to play and if I do I'm mostly playing D2R which is just utter garbage for such max efficiency-related testing...
Interesting link frozzz - I got 237ms but there is no way I could sustain that.
The theorycrafting is interesting, but luck is always the biggest factor. The odds are so large in D2, sustainability is the biggest factor for me - You have to have something that you can maintain over 100s of hours.
And you can always offset your efficiency by staying up for one minute and doing an extra run.
I tried 200 FCR when i was actively playing the game. It was fast as heck, but you can definitely get used to that speed. Don't remember any numbers, but the loss in mf territory was massive, that's for sure. Maybe if you get some godly FCR/MF gear, you can make it happen, but those items are too hard to find honestly. Even trading for them would require a lot of effort.
The other thing is that S&E stuff. You can't really compete with it or go around it. When i was going for pitzerker score, i had godly S&E times already and my runs where like 52s on average iirc. I have no idea how i got them, i just played without sound and i was able to exit the game instantly most of the time. I don't think i got a single run below 50s without S&E lag though and we are talking about runs way below that. Every single run that i did in 46-47s territory had S&E time so terrible that it automatically made it ~50s run. There were some runs that could have been like 42 or 43s, but they were instantly getting a few seconds longer just because of the lag.
The run couter we were using ends the run when you press S&E and adds S&E times to the next run and also those runtimes i mentioned here can differ from the actual ones i was getting, cause i don't remember it that precisely, but you get the point.
Great points! I hope I don't come across as rude, I've learned a ton from the SPF and will keep doing so. Here are a few things I wanted to bring up after reading the discussion:
Gear:
- The gear I use is all found on single player lod (no d2r, bnet, trading, etc.).
- A 2/20 magical circlet with 2 os from larzuk isn't that hard to gamble, I even got some duplicates for other classes. I keep gambling circlets so I might even get extra barb 2/20s in the future.
- If you have a 2/20 circlet you don't need a 2/20 barb amulet. If you do have one you can switch one of the wizspikes to suicide branch for 56% hork. I use a 2/20 assassin amulet with 46 life and 9 mf.
- I don't use crazy rings either, they are simple rings with decent mods and and 12 mf each, they could use a lot of improvement.
- With that gear I still reach 607/486 mf breakpoints with 55% hork, compared to 712/573 with 56% hork using the HLW/arach setup.
Theoretical efficiency:
- As I mentioned before, even with my items the build generates 96.42% of the uniques of the HLW setup.
- This means it only needs to run 3.71% faster to break even in efficiency.
- 3.71% of a 60 second run is 2.2 seconds. Adding 0.85s from the loss of kill speed from HLW, we get a total of roughly 3 seconds.
- To break even you'd only need about 75 casts, which is a very conservative estimate for a 60 seconds run.
- Those points are not really affected by loading times, they would even hold true in D2R. Even a very slow loading time would only slightly increase the time save required to break even.
Practical efficiency:
- I think our discussion covered pretty well the reasons why the build would lose some practical efficiency.
- However, I don't think they are necessarily as impacful as one could think:
a. Mistakes already happen at 105 fcr. While 200 fcr will definitely increase their occurence, it's hard to gauge by how much, and it varies per player.
b. With the 200 fcr variant, you can fix some player errors more quicky. That includes not only the mistakes that happen because of the faster cast speed but also those that were already occuring. Here's a list of some of them:
I. Inaccurate teleport: with faster teleport speed, you can fix your position more quickly.
II. Unique/champion skip: if you notice you skipped a pack, you can backtrack more quickly.
III. Cast on kill switch: If you accidentally cast on kill switch, you lose 2 frames less with 63 fcr than 20 fcr.
IV. Minion hork: horking minions by accident doesn't cost you as much time with 200 fcr.
- Fixing those player errors more quickly would at the very least mitigate the efficiency loss caused by their increased occurence at 200 fcr.
I don't know for sure if the build is more or less efficient than the standard setup, or by how much, but I think it's close enough to warrant some testing. It's also easier to gear for than you'd expect. As a bonus, it's a good training tool for 105 fcr. I noticed my accuracy on 105 fcr was improved after playing a lot of 200 fcr.
My prime D2 days are far behind me, so i won't test it. Maybe i will have some sort of renaissance in the future, who knows, but at the moment it looks like this to me:
If you wanna go for top scores, you will loose too much time due to S&E lag.
If you wanna just farm some items, going for 200 fcr is not optimal, cause it requires too much focus.
Although it seems appealing and 200 FCR build just sounds cool, i think it won't surpass the existing build. Maybe someone wanna prove me wrong though? xD
Google
