SPF Hall of Records

[ffs]

@ffs I hope you won't find any missed horks and other compromising things xD I think you have a really high chance of getting similar score too. You just need to get those good spawns and work a little bit on that s&e lag ^^ Your 1,441 score with pretty much no preparation was sick. I was struggling to beat it for a few sets.

I think we decided a few hork misclicks are acceptable.

Those loading times are crazy indeed. Seems you're the first player on Windows who appears to have solved the issue, at least for Pit running. I'm still finding improvements to my runs here and there, by now efficiency is constantly below 1.45 over 100+ runs even with low-ish density of 8.4 overall. But from this point onwards it seems the more time I shave off, it's really noticeable how S&E becomes a huge problem again. Think that's what @Gripphon meant when he talked about his past attempts to improve his score. My run times for first 100 MFO runs were in the 53.5-54s range, but S&E lag became really bad again, losing some 2-3 seconds every couple of runs, and something like 1s is normal unfortunately. Sometimes I look at that S&E display long enough until I get from resting heart rate to 200 bpm and wanting to scream at the screen. Overall I'm still seeing run time improvements despite S&E getting worse, but they are rather small now and won't be enough to get to 1.4 or below (even with awesome density), unless I can magically solve the loading times. It's a bit of a shame because our little competition would be more interesting otherwise. Our maps have pretty much the same quality, and I think me having better charms vs. you having lvl 99 Barb makes things even overall as well. So it would really came down to boss RNG and fewest mistakes in execution.

I'm considering switching back to craft ammy + Goldwrap. If I'm lucky it might get average times slightly higher, maybe just so much that I save overall time from having faster S&E again. But I don't have high hopes of improving the S&E issue enough to really challenge your score.

But since I came up with your setup I consider that score 50% mine anyway

 

[Kinkara]

Sometimes I look at that S&E display long enough until I get from resting heart rate to 200 bpm and wanting to scream at the screen.

I'm sure this is bad for you XD

 

[ResTTe]

@ffs You can check my s&e times in the video. I get like 2-3 seconds lag sometimes too. Don't think i saw anything above that though. I had pretty ok loading times before, but after switching to -ns they improved by a lot. I literally didn't do anything else. Hmmm... Maybe it has something to do with the map itself? As you probably know, my map has only 4 tiles. Your map and Gripphon's map both have 5 tiles, so it should take a little bit longer to load it or something? On the other hand, your second map has 4 tiles too and you are still getting some bad s&e times. Weird... Does recording change anything? Maybe it slows things down?

It was always a problem to me that loading times impact your performance. I know that we want to see how fast you can farm and exiting and creating a new game is a part of it, but i was always about improving your skills rather than getting loading times shorter ^^ I was forced to do it, cause i wanted to get some good score. I wouldn't play without sound otherwise. Don't think it is possible that someone will get below 1,4 with mediocre s&e times... Kinda sad.

Yeah, you are right. I just wanted to try it, cause it was the most reasonable build to play. Barb can get 1 mf bp on kill and 2 mf bps on hork and that's pretty much it. I am still not sure if merc's gear and even his type is the most optimal though ^^ Maybe it is possible to go for less health, more str and for example prayer merc? As it was mentioned before, might aura doesn't do that much and you have to wait for merc to proc it pretty often, so maybe it isn't that great? This is just theorycrafting, but i think it's quite interesting to test.

 

[ffs]

I'm sure this is bad for you XD

I agree.

You can check my s&e times in the video. I get like 2-3 seconds lag sometimes too. Don't think i saw anything above that though. I had pretty ok loading times before, but after switching to -ns they improved by a lot. I literally didn't do anything else. Hmmm... Maybe it has something to do with the map itself? As you probably know, my map has only 4 tiles. Your map and Gripphon's map both have 5 tiles, so it should take a little bit longer to load it or something? On the other hand, your second map has 4 tiles too and you are still getting some bad s&e times. Weird... Does recording change anything? Maybe it slows thigns down?

Yeah I see your S&E times, thanks for rubbing it in my face.

I've seen you getting delays every once in a while, didn't watch the whole set yet but it seems rare and most of the time pretty much instantly. Reminds me of Mac D2 gameplay where I had the same S&E. Not missing it though, overall my run times improved a lot from the switch to Windows, D2 just runs way smoother.

I can't really tell how much of an impact recording has, because currently I'm simply always recording during MFO. (Reasons is just when I felt had a good session which I want to run the numbers for, I always have a recording I can quickly check.)

I'm gettingthose S&E issues also when not recording. Could be a little better maybe, didn't observe this a lot yet. But I didn't feel like it made a difference tbh. By now I tried all of this:

Spoiler

• -ns
• RWM disabled
• with glide wrapper, and trying various glide wrapper settings
• without glide wrapper and only -ns
• without glide wrapper, -w and -ns
• without -w but using windowed mode via glide wrapper
• only one character in save folder
• no screenshots in D2 nor Runewords folder (disabled anyway)
• not running any other apps apart from D2 and OBS when recording
• updated from the 1.14b as provided by the installer to 1.14d

By now I don't have any further ideas.

-ns made the biggest difference for me, too. But not enough to solve it as well as it apparently did for you.

Using -w did a small thing: If runs were not super fast, the "normal" S&E was a little faster. I mean those situations where also in fullscreen S&E speed was okay. I was back in the main menu maybe 0.2 seconds faster or so in those cases. But somehow it seemed to make things actually worse with the really annoying "long" S&E lags that occur every couple of runs, at least it definitely didn't cause any improvement there. I felt it was better to use glide wrapper, and be it only to avoid misclicks when teleporting on the edge of the screen.

Windowed mode from glide wrapper didn't seem to do anything at all, except screwing up the mouse cursor big time.

Another thing why I kinda need to use glide wrapper is the game is too dark otherwise, and I cannot adjust brightness in game. It's not just visually ugly -- it also makes it harder to catch the subtle coloration differences between normal monsters and some bosses.

I'm running on a MacBook with onboard GPU which allows basically no setup/customization. Also running on a Win 10 partition with "only" 40 GB. It's an SSD but the graphics unit requires available space, so I guess maybe hardware plays a role here, too.. I don't know. Not really a strong gaming computer anyway.

Maybe it has something to do with the map itself? As you probably know, my map has only 4 tiles. Your map and Gripphon's map both have 5 tiles, so it should take a little bit longer to load it or something? On the other hand, your second map has 4 tiles too and you are still getting some bad s&e times. Weird...

I run the smaller 4 tile map at the moment. On the 5 tile map, S&E lag was not as bad as it is for me now, but I think it's really just about run times rather than the map layout or so. My first runs that I posted on the bigger map were a bit slower, around 56-57s. I think my 4 tile map is better in the long run because density is a bit more consistent. Maybe I should try the other map anyway, and see S&E loading times behave there when improving run times. But I'm afraid once I get below 53-54s on that map I'll run into the same issues..

Yeah, you are right. I just wanted to try it, cause it was the most reasonable build to play. Barb can get 1 mf bp on kill and 2 mf bps on hork and that's pretty much it. I am still not sure if merc's gear and even his type is the most optimal though ^^ Maybe it is possible to go for less health, more str and for example prayer merc? As it was mentioned before, might aura doesn't do that much and you have to wait for merc to proc it pretty often, so maybe it isn't that great? This is just theorycrafting, but i think it's quite interesting to test.

At this point I don't see a better setup, both in terms of theoretical efficiency as well as comfortable playing. Bigger life pool = more aggressive play = faster runs. I'm convinced adopting such a playstyle saves way more time than saving like 0.5 - 1 swings per run from dumping 100 points into strength. Jumping into Fana/Might packs while cursed is no biggie with >3k life, but with >2.4k I don't feel quite as comfortable. As for Merc, I don't really see a better option either. I considered trying another aura, but I think one of the remaining situations where I can still improve my gameplay is creating situations where you start horking packs while Merc finishes off some champion (or second/third boss in same location, after Barb killed one). For those situations Might is still best choice because it boosts Merc damage a lot (even though it doesn't do that much regarding hits required by the Barb to kill a boss). CNBF seems ideal basically for the same reason IMO, and then 60 IAS armor is the logical remaining choice along with GT/T Insight base. At least I haven't discovered a better combination of Merc setup, considering aura triggering speed, DPS and assistance vs. those champion packs in particular.

 

[ResTTe]

What is weird is the fact that my s&e times got slightly worse like 2 days ago. At least it felt like that. I think they were better when i first used -ns. Maybe it was just the contrast between bad loading times with sound and good without it and i got used to that after a few sets. Hard to tell. I don't remember changing anything. The only thing i can think of was some win10 update iirc and checking LK map which should be a pretty obvious reason why the lag got worse. Having bigger map files = worse s&e times.

I personally got very small upgrades if any from doing all that stuff you mentioned. The only worth using trick is the -ns command. If you already have good loading times and you want to go for that top tier godly score, you can probably test those small things. The whole thing is probably some weird combo of the map, runtimes, your computer spec, operating system and some drivers. That's why it is so hard to do anything with it and different things worked for different people etc.

What i observed is the fact that mobs still put you in hit recovery or interrupt your berserk or whatever you wanna call it, when you teleport into the middle of the pack. You can see this in my last set i think. More life makes a big difference, but i think i can go with a little bit less health and more str and be good with such setup. What i would really want is more light and cold res. You said you are comfortable with your res, but i am not with mine. I loose quite a lot of hp to some light/cold enchanted bosses. That frost nova hurts a lot sometimes. I would like to get rid of that completely. I still kite some elites when there are many of them. It's just impossible to kill them when you have like 20 minions around you... That's why i think less life and more str can work, cause i am luring those elites no matter what. The downside of that is the fact that every single hit by some random minion and elites themselves will hurt a lot more.

All the stuff we are talking about here doesn't make a big difference imo. It's not like you will loose 2 seconds per run, because you have 0 res instead of 50 or something. You can stack those little improvements, but all that really matters is getting that good RNG if you want to go for an entry.

 

[ffs]

Having bigger map files = worse s&e times.

This made me realize I never cleaned the map files after leveling this Barb to 95 at Baal. I did that straight away – and then was disappointed when it didn't help at all.

Some good news at least: Today I did ~2 hours of MFO running and got a batch with 1.422 score. Got very nice boss density of 8.64 and 53.94 seconds average run time. In light of your new top score I can't help but wonder how it would have been with a better S&E situation... but it is how it is. Still happy to achieve the goal of getting past Fabian's long standing record as well.

Can't upload the video during MFO because two qualifiers dropped in it. I'm only ~200 runs into MFO so not unlikely I'll get a better set before MFO ends, but in case I don't I'll upload the vid after MFO.

Maybe tomorrow or next week I might go back to the caster ammy + Goldwrap setup, and/or try the Barb with bigger map, in hopes of getting a better grip on S&E. I have doubts either will help enough to get a much better score, but maybe there are some improvements to be gained. Pretty sure I'll need out-of-this-world density to get even close to 1.4 either way.

 

[Albatross]

MacBook with onboard GPU

This might actually have an impact. Haven't run tests but playing on my MBP (win7) does feel slower compared to the PC with the same processor and similar ssd, but with a dedicated graphics card.

There's one more improvement I don't remember I wrote about, but I'm afraid it only potentially impacts load times.

When using glide, there's an option to turn on VSYNC, which should cap the game at the same number of frames your screen refreshes (usually 60). I have this turned off and wouldn't recommend setting an FPS limit either. Although the game is locked at 25 FPS, load screens don't appear to be. It would be best to run a side by side comparison with recordings to prove or disprove this, but the load screen takes noticeably longer to animate the light coming through the door at 30 FPS. Weather this actually slows down the load is hard to tell.

 

[ResTTe]

I switched back to playing with sound and RWM and the difference is huge. That exit lag is now way longer than it was for me when i was going for the entry. It's more or less 5 seconds. It doesn't really matter how long are the runs, unless they are like 70s or something. The shortest s&e time is like 1-2s which is basically the same as my longest s&e time with -ns. I do chill farming for the MFO at the moment, so it doesn't matter for me, but it definitely proves that -ns is crucial if you want to go for a good score. Wouldn't think about even trying without it enabled.

 

[Luhkoh]

@ffs absolutely huge congrats on beating fabians record. I know you two are fighting to top spot which is just awesome, but passing fab when you both were at ~1.6 adjeff very recently is just incredible. Great work!

 

[ffs]

Well, my old score was from when I last really played Berserker ~2 years ago and with not much practice back then, so getting closer now is not that surprising I think. Coming back now it took some time to iron out those misclicks and getting good density. I hope we'll find a way to get a better grip on S&E/loading times so that we can have some more sub 1.4 scores, but doesn't look too good it seems. It still took 7 or 8 serious attempts now I think for me to get past his score. Fabian once told me his score was only the 3rd set on his map. Actually I'm rather sure he could have done better with more tries as well tbh – just watch his old vids to see how flawless his gameplay really was; coupled with that sick gear and assuming good S&E times I still think he would have the best shot at top score.

This might actually have an impact. Haven't run tests but playing on my MBP (win7) does feel slower compared to the PC with the same processor and similar ssd, but with a dedicated graphics card.

There's one more improvement I don't remember I wrote about, but I'm afraid it only potentially impacts load times.

When using glide, there's an option to turn on VSYNC, which should cap the game at the same number of frames your screen refreshes (usually 60). I have this turned off and wouldn't recommend setting an FPS limit either. Although the game is locked at 25 FPS, load screens don't appear to be. It would be best to run a side by side comparison with recordings to prove or disprove this, but the load screen takes noticeably longer to animate the light coming through the door at 30 FPS. Weather this actually slows down the load is hard to tell.

I wondered if onboard GPU makes a difference, but so far I thought it didn't because it seemed everybody complained about S&E times irrespective of their GPU. But in your case it seems to be a bit slower indeed. I had that VSYNC disabled already, but thanks for pointing it out anyway.

 

[Luhkoh]

@ffs i recently figured out to run d2 on my laptop's gpu (it didnt default to that), and not only did that make it to where I could record with obs while playing wihtout blowing up my s/e times, but it also made countess s/e times quite noticeably faster. I think it's likely that's just because countess s/e is so terrible that it's just a place where it's easy to notice and that it makes a less noticeable difference on other runs too. Made me wonder how much time I could've saved in 98-99 with that simple setting change .

 

[Luhkoh]

SC Berserk Barbarian
Category: P1 Pit
Average run time: 69.986 seconds
Average boss number: 7.38 (369 total)
MF: 727/577
Find Item: 55%
Seconds/Bosspack: 9.483
Efficiency: 6.118
Adjusted efficiency: 2.172

With restte breaking 1.4, and ffs about to beat fabian's long-standing record as well, I thought I'd get my first official HoR entry in here, and make everyone else feel better . I was feeling like I was going fairly fast with my zerker during mfo and recently figured out how to record without slowing my game down drastically, so I recorded 50 runs. Huge thanks to @ResTTe for helping me count the bosses afterwards. You're the man!

The Bad: I have a goal this year to get a sub 2 adjeff alvl 85 entry in this thread. This was not quite there. I miss some packs in pits due to red/green colorblindness. Most notably, red devilkins without an aura are just a complete loss. Makes me want to stick to AT a little bit, since I think I miss fewer bosses there. Unfortunately, my map is not as good as I thought (it kinda sucks) and should be rerolled soon.

The Good: Now that I have an official entry up, I feel more motivated than ever to get that sub 2. And if I roll a non-crap map on this barb, I expect significant improvements immediately.

@ffs are we supposed to be replying with the tables and adding ourselves or is it something you just update? Just wasn't sure how you wanted to run it. Thanks!

 

[Albatross]

Hell yeah

Didn't know about your RG color blindness, I've rolled brown/green which I guess shouldn't be as much of an issue in the pit? Does tweaking contrast/gamma help?
Have some fond memories when bosses had more color, but don't remember which patch broke the palettes, probably 1.10...

Anyhow, congrats on your first pit entry! I've also found that adding my name to a popular table motivated me to improve.

replying with the tables and adding ourselves or is it something you just update

Either way, we could also link the post where the adjeff was reached. Makes it easier to find later.

 

[ResTTe]

Nice entry @Luhkoh! Glad you finally did it. You definitely need a better map, cause that density was terrible xD

I think we can have the tables updated after @ffs posts his 1,422 score. Prove me wrong if we need some validation or something similar.

 

[ffs]

@Luhkoh feel free to update the tables if you so wish I think people usually don't do it with every post because it makes them so long... more about not spamming the thread with every result rather than waiting for anything. No need to wait for my most recent set either, I will get around to posting it one of these days, still need to sort through my MFO recordings on the off chance I somehow got something better.

Btw I really like your Beast Singer idea. I still haven't gotten around to trying it, but I checked Owlie's Merc calc to see how Beast reduces sec/kill for Merc. First rough number crunch indicated Beast is likely better than 180 MF sword in theory, saving more time than needed to make up for those MF breakpoints. Hope I'll find the time to check it out soon.

 

[Luhkoh]

Thanks guys!

@Albatross what do you mean you've rolled green/brown? You mean color of zombie boss vs minions? That probably happens to me, and I'm not sure I don't miss packs in AT. But recognizing packs is definitely a multifaceted thing for me, including the color, the location, and the like shape/movement of the group of monsters, and i at least feel like those are easier to work with in AT. Smaller map, less cluttered with trash mobs, and devilkins just move weird etc. Haven't messed much with the settings, I definitely should! My laptop screen is also not particularly bright at its highest setting, so that probably hurts too. Much to mess with .

@ResTTe map is already gone and out of my life , but yet to be replaced with a better one.

@ffs ok sounds good to me. Not enough entrants here in general for spamming to be possible imo, but I hope we get some more participation! And it definitely would make for too-long posts. So I like the idea of you just updating the tables for sure. Thanks in advance, and good luck with the beast singer. Looking forward to seeing how it goes.

 

[Albatross]

I meant green/brown color blindness, sorry for being vague might be why my infinity runs are much slower, I find it hard to figure out what to kill with so much green.

Have the same concerns you mentioned about high eff pit running, but we'll l see when I get a zerker going (probably not until 2021 ).

 

[ResTTe]

I knew we don't wanna spam this thread with the table, but i also thought we need some validation of out scores ^^ I also wanted to wait till everybody sents a "good" score that won't be beaten for some time like all those entries i posted here xD

I wonder how color blindness looks like. It must be really hard to get good efficiency in some cases if you can't see elites.

@Albatross Go for that Pitzerker entry!

 

[Luhkoh]

Maybe albatross will start playing pitzerker and restte will start playing AT blizzy?

@Albatross makes sense, didn't know I had a a fellow colorblind d2 magic finder working on improving runs . I think that's usually the same type of red/green colorblindness, but sounds like you might have more issues on the green side of things, where I would say I have more issues with red. But I DEF have problems with green and brown in many situations too irl. If you figure out settings in particular that help the issue, definitely send them my way .

 

[Albatross]

I wonder how color blindness looks like

Many people aren't even aware they have it and there are different kinds. For me, it's either really hard or impossible to distinguish some shades of green and brown. And yes, Luhkoh, you're right, some reds as well.

It's easy to test by looking at color blindness images. I can see there are different colors in the circle, but for the majority of green/brown/red ones the number is indistinguishable. Even if I pick up on the pattern I still can't say which number it is, only that there is actually a number in there, for some you could tell me there's nothing written inside and I'd believe you.

As far as impact on gaming, hard to say because I could never see colors the way people without color blindness do I suppose you find ways to compensate if there are any. I usually run with high gamma and brightness, I think it helps, otherwise I wouldn't put the additional strain on my eyes.