2017 Winter RFL Signup and Running Thread: Grande Finale

T72on1

Diabloii.Net Member
I finally got around to doing 10 test runs with my Fire Sorc. According to runcounter advanced, I averaged 3:50, on /p5. While slow compared to others, that's far better than I expected after my first few runs. Still plenty of room for improvement, as I'm still not sure what routine is best with her. Static or not, Meteor and FireBall or only the latter, ...

Btw, after round 2 there was a mention (I think by @Gripphon) about runcounter not being accurate, and for example total time not being correct? What's with that? Because I just stopped when the runcounter got to 20h. I know you have to S/E the first run when starting, because runcounter won't count that, but other than that ... ?
 

Gripphon

Diabloii.Net Member
Ah yes, runcounter issues. This post will be about handling runcounter and will be kind of lengthy.

There are 2 ways to handle runcounter, depending on what do you prefer.

Way 1:
have correct total playing time as your total elapsed time from run counter and nearly correct average run time (not 100% correct), but totally incorrect number of runs done.
Way 2: have correct average run time and correct number of runs done, but total elapsed time on run counter will differ considerably from your total playing time.

You can't have both correct total playing time and at the same time correct average time with number of runs. You have to choose in between way 1 or way 2 which do you prefer. In my thinking way 2 is superior and more precise on all fronts, even total time played which is easily calculated from average runtime multiplied with total runs done. Though, I've seen players prefer way 1. So, allow me to explain why the issue arises. In fact, let me name all the issues of runcounter and how to deal with them to get as precise data as possible.

Issue 1: first run doesn't count toward average time nor runs done, but still counts toward total elapsed time
From this first issue it is immediately clear why we can't at the same time have correct total playing time (in elapsed time) and number of runs done with correct average time. When you start your run counter and try to do that first run, after you finish you will see you did 0 runs with 0 average time, but still total playing time is not 0, but is whatever it took you to complete that run.

When does it happen:
- if runcounter was inactive and then you start it
- if you pause runcounter and then resume it

What to do if you prefer Way 2 of running:
- simply s/e first run after you do any of those, it won't save towards average runtime nor toward runs done, however it WILL count toward total elapsed time normally. Imagine you forgot to pause your counter and it works for 3 hours, then you s/e, do 1 run. Runcounter will show total elapsed time of 3 hours 20 seconds and that you did 1 run with average time of 20 seconds at the same time.

When does it NOT happen, aka first run DOES count normally:
- whenever you perform s/e and continue running
- whenever you visit gomule and then close character there after mulling items off, first run after that counts normally
- if those two things happen, DON'T s/e first run, you will totally screw your average time and runs done count

Issue 2: literally first session of the runcounter after you start fresh from the run zero, no matter what you do, that first run will register in fastest run done.
If you wanna correct information about fastest run done, DON'T s/e first run when you start new session in runcounter (if count of runs is zero). Simply either run normally or wait some time then s/e. This DOES NOT happen if your runcounter session is not at zero runs, aka you continue from run 300 or whatever, then this issue is not active for whatever reason.

Issue 3: if you die, your fastest run in session could be screwed
Not important if you don't care about that info.


Problem with a total playing time

As explained in issue 1, whenever you start runcounter or unpause it, that time WILL count toward total playing time even if that run does not register toward runs done or average time as also explained under issue 1. This means when you try to calculate your total playing time via

average runtime * number of runs done = total playing time

result you get will NEVER be the same as total playing time displayed on runcounter due to issues explained. Let's see from example of my RFL round 2 set 2. Have in mind I prefer to do the WAY 2 with runcounter to have correct number of runs done and correct average time per run.



My runcounter shows I have spend total time playing of 20 hours, 6 minutes and 10 seconds. Did I overreach 20 hours mark allowed by the rules? Let's do some math. Lets calculate total playing time from my average time and total runs done instead.

17.307*4160 = 71997.12 seconds (20 hours is 72000 seconds)
Total time displayed by runcounter: 72370 seconds

Okay, so there are two total playing times. Which one is correct, did I play 71997 seconds or 72370 seconds? Both can't be correct at the same time, right? Which one is correct depends entirely on which way did you decide to play: way 1 or way 2. Since I prefer way 2 to have correct average time and total runs done, for me correct playing time is 71997 seconds while this displayed with runcounter is wrong.

Also, let's calculate my average runtime from total time elapsed knowing runs done.
72370 / 4160 = 17.397 s.
So, runcounter does not use total time elapsed to calculate average time. My guess is it keeps tracks of two independent times: 1 is total elapsed time, other is total playing time used to count average run time from. I'm not sure though. Reason why this second time needed to get average running time from is not needed to be displayed is because we can calculate it by simply multiplying runs done with average run time.

Why is wrong and where does that extra time come from? Since I also wanted correct value for fastest run done, after starting runcounter from zero, I simply waited ingame for 30 seconds repairing gear etc. and waited for runcounter to cross at least value of 14 seconds which I knew my fastest runtimes will be about. In fact my first run lasted 30 seconds while NOT running at all, I simply waited for runcounter to get some time (14+ seconds) to have correct fastest runtime laster. If I s/e immediately, my fastest runtime would be displayed like 3 or 4 seconds... Read issue 2 again.

Every time I had to pay attention to issue 1 - runs that are not registered by runcounter. This means every time I started it, I had to s/e first run (which otherwise would not be registered toward average time or runs done screwing precision of my count). Since s/e takes some second or 2, those 2 seconds DO COUNT toward total time elapsed, but NOT toward runs done or average run time. With time and continuous repeating, every time there were 2 more and 2 more and 2 more seconds that made a difference, so in the end I ended up 6 minutes off in my time elapsed on runcounter than what my real playing time was. Actually being off by only 6 minutes in 20 hours is pretty darn precise, I'll explain how I managed that. Otherwise my total elapsed time would be like 21 hours and I would have to explain to everyone that I for real played only 20 hours as per rules, and not 21 hours... Last time I used my runcounter time to also cube gems and then I s/e that run (not counting toward average nor runcount, but still screwing my total elapsed time to differ a lot from my actual playing time, so my runcounter showed 22 hours of elapsed time while I still didn't exceed 20 hours of playing time. This time I decided to minimalize that offset and ended only 6 minutes off thanks to that.


How to do WAY 1 of running

Way 1 means you care about total elapsed time on your counter to be correct. Downside is your average time won't be 100% correct and your runcount will be totally wrong and you won't have a perfect way to estimate did you do 3605 or 3608 runs since your average time is also precise enough, but not entirely to every millisecond.

To do this way of running, forget about doing s/e or anything like that. You want your total elapsed time as precise as possible, right? This means once you start runcounter, you simply go run like a maniac and simply embrace the fact you won't have correct number of runs done and will have to calculate it by yourself.

Problem: running this way, runcounter will actually steal the time from your total playing time. Why? Because you will have the same issue as I had when I ran - it will take you a second to start running after you unpause or start your counter (second wasted) and another second to pause it when you are done (another second wasted). In the end you might end up like 10-20 minutes in your 20 hours session doing nothing but pausing/unpausing your runcounter. If you are not perfectionist, you shouldn't be bothered. But to me, those 20 minutes easily mean extra 60-70 runs...

Secondly, average run time won't be 100% correct, but still nearly as precise. Why? Remember, runcounter never counts first runs of the session toward runs done or average time. Maybe that first run of yours was terribly long, but it didn't register at all nor did it harm your average in any way. Or maybe it was the fastest run done ever, still won't register or help your runtimes. So, you won't be sure is your average time 20.035 seconds or 19.978 seconds in the end. But, doesn't matter much if you don't wanna 100% precise numbers, but something approximative and close enough. Not everyone cares to have fully precise data on their running, especially if difference is so small.

Thirdly, displayed number of runs will be totally incorrect (read issue 1 again why). You will have to calculate it yourself easily by

Total elapsed time / average time = number of runs done

In the end, this method doesn't provide you with ANY number with 100% precision, even that total elapsed time unless you wanna get your 20 minutes stolen due to your pausing/unpausing counter.



How to do WAY 2 of running

This method is somewhat bothersome to do, but it will provide you with as precise data as possibly can be. Only issue is total time elapsed displayed on runcounter WILL DIFFER from your total playing time. What I'm saying is, you can expect your total elapsed time to easily exceed 20 hours of running by at least few minutes and up to an hour depending how much care do you pay to your runcounter.

Now, get back ti issue 1 to make sure you understand when will runcounter miss your first run and when it won't. It is simple really. My playing rountine was this:

1. When I started runcounter, I simply s/e first run to start running since that run would not count toward runs done or average time, only towards total time elapsed we are completely uninterested in with WAY 2 of running.

2. After character got full, I paused counter and went to gomule, dump items. Upon returning to running, as explained, once you close your character in gomule, it is saved and runcounter WILL count first run after unpausing it toward average time and runs done. You can know that whenever runcounter starts from 00:00 of current run time elapsed, ti will always count it toward all relevant things (except if you started it when it was inactive, it will start from 00:00, but WON'T count first run toward average time or runs done). If not, then it won't count it. So, after gomule I didn't bother with s/e of first run, I simply continued running normally after unpausing counter.

You could, after visiting gomule, simply unpause-pause-unpause counter, then you go s/e first run normally as you would as if you do thing explained under 3.)

3. When I needed pause in the middle of the running, upon returning I had to s/e first run after unpausing since in that case runcounter DOES NOT count first run done toward average time or runs done.

Overall, quite simple with some practice. As a result, you will have correct data about:
- total runs done
- real average time per run
- total playing time once you multiply total runs done with your average run time

So, WAY 2 offers more precise numbers on all fronts at the same time, only difference is in WAY 1 you pretend your total elapsed time is the same as your total playing time while having slightly off value in average time and completely wrong number of runs done, while in WAY 2 you embrace the fact your total playing time won't be the same as total elapsed time, but you will calculate it yourself instead.


Hopefully handling runcounter with precision will be easier for everyone now. I'm gonna stick with Way 2 of running style since it is more precise on all fronts, but as a result my total elapsed time will always exceed 20 hours and will never be the same as my total playing time as explained above. I want precise numbers, not nearly precise as what WAY 1 offers.
 
Last edited:

Gripphon

Diabloii.Net Member
No hype from anyone?

I thought about starting to play at midnight, but since I'm really unlikely to go for a 2 sets, I kinda wanna do other things. 20h of cows is good enough considering what kind of area that is. I plan on testing 5 characters - nec, trapper, javazon, lite sorc and fire sorc. Bad thing is I have to roll maps for them. Good thing is I have plenty of time to do that with 1 set in mind. I assume cows will be somewhat interesting with changing of characters. Though, I have no idea how will sorcs turn out to be. Perhaps I'll stick more to some characters than others.

Good luck to everyone and have fun.
 
Last edited:

pharphis

Diabloii.Net Site Pal
I'm hyped, but I still have to level my frenzy barb a bit before I can run with it. The necro at low lvl works fine but probably is less efficient than simply the /p1 sorc. Oh well, something new to try
 

ffs

Diabloii.Net Member
Too much work this week to hype more :) but will go for it on the weekend for a couple of hours at least. Obviously, the prospect of killing thousands of cows for 20 hours is not very exciting either haha...

But: You gotta do what you gotta do.

Going to go with Javazon first obviously. I think I've figured out a nice set up that is very capable, but needs better map as well to really shine. I'm convinced it's going to deliver times under 3:30 to get leg and clear cow level on /p7 with her without much trouble, just needs a decent map. Somebody needs to push her further, so let's see. :)

Still convinced I'll be bored with her at one point though. Then I'm going to mix it up with Blizzard and Lightning Sorcs, Trapper, and/or maybe some funky less efficient build that's different, like some Druid variant or Hammerdin. As for Necro, I probably just suck at it ;) at least it's slower than it should be. With some practice it should improve for sure but I don't feel the desire to do that at the moment cause I enjoy playing other characters for now. That might change quickly though if Java gets too boring and others don't have enough potential. :)
 

Noodle

Moderator
<Enters door with paper sign taped to the nearby wall reading "Newbs Anonymous.">

"Hi, I'm Noodle, and I'm a newb."

<In unison> "Hi, Noodle!"

"Hi, everyone. Thanks for the warm welcome. Oh, hey there Pharphis, good to see you acknowledging your problem."

"If you tell anyone I'm here, I'm going to revoke your 1.07 license."

"Pretty sure that's not a thing, but anyway, I'm here for another reason."

----------------------------------------------------------------------------------------------------------------
So, I've been spending some testing out some builds for cows. Nothing new there. Over and over again, I found myself unable to open the portal. Character after character, I even checked it out in Normal and Nightmare, and nothing. I started thinking there was something corrupted in my install, and maybe I should restart.

Then I thought maybe I'd try an Tome of Town Portal instead of a Tome of Identify. Were I a clickbait site, here is where I'd say, "The results will astonish you!!!"

In my defense, I've gone after cows maybe a dozen times since leaving Battle.net after the first season of 1.10. But yeah, a total newb maneuver.
 

Gripphon

Diabloii.Net Member
Don't wanna spoil it, but found upgrade to trapper, she will likely beat even nec at cow running.
jk, but did find potential upgrade, however I will be sure once I test it.

@Brak
Gogogogogogogogogogogogogogogogogogogogogogo

@Pb_pal
Regarding your fire druid, perhaps you could try this:
Once you get the leg and cast town portal, cast Armageddon, then get into portal to open a new one for cow level. Once you get there, timer of using fire skills could be off and you could cast Fissure normally. If this works like I think it does, you will also have 40-45 seconds benefit of pretty solid Armageddon damage without any penalty at all.
 

Gripphon

Diabloii.Net Member
Upon some more thinking, I'm willing to try more builds to see what they can do. Adding these to my testpartyballz:

Nec CE/nova - basically to confirm his Opness. I'd make Beast golem if I knew I won't die to random teleport into cows
Javazon - gotta make this chick work, no excuses. p5 vs p7 test for her as well as 35 vs 55 ias
Trapper - new map and new upgrade, be afraid etc.
Lite sorc - in theory awesome, in practice electrifying to the bone
Fire sorc - should be awesome
Fire druid - planned one, looks good, should be good enough to smack sum cows
Hammerdonk - 5 hammers per cow doesn't sound that terrible, we shall see
Blizzard sorc - honestly, not into it really, but... see below




Also rolling all those maps, can't wait...

All tests will be done on p5, however I will do p5 vs p7 test for javazon.

And to answer in advance to obvious question, NO, I won't test FoH paladin for cows.
 

PhineasB

Diabloii.Net Member
Best of luck to all the runners for this round. I look forward to seeing the results from the different builds -- it seems like there is more diversity of builds for this round (?), which is pretty cool. I have limited play time the next two weeks and will probably focus time on my 99ers (in places they can gain XP).

So no entry for me as of yet. I may play around with the trapper or a poison necro for a break and will start a timer on the first run just in case it's too much fun to resist trying to squeeze in a full set.

Question: does hammerdin + cows = lots of cheeseburgers?
 

Gripphon

Diabloii.Net Member
Also, some prediction how I think testing could turn out. Some of those characters I never played at cows, so... should be interesting to compare with real results in the end. I think it will be something like this:

1. Nec
2. Trapper
3. Javazon
4. Fireball
5. Lightning
6. Hammerdonk
7. Blizzard
8. Fire druid

Possible in theory: javazon beats trapper
Don't think so, but wouldn't be massively surprised: firesorc or lightning sorc beat javazon
Dark horse that could turn better than predicted: lightning sorc
Potential massive surprise: hammer much better than expected, trapper gets as good as nec
Unlikely surprise: fire druid
Mega jumbo surprise: lightning sorc kills King... somehow
 

NanoMist

Diabloii.Net Member
Make nova great again...?


Recorded this one before the start of my set. As I learn the map and cow spawns more, I should be able to get better positioning for nova to do its work. Most runs are under 3 minutes, depending on how clumped up the cows spawn. Mana burn is annoying, but not deadly. It's still the might/fanat/extra strong packs that are most dangerous, as it is with most characters. Overall a fun build for cows, no aiming needed. :D

Late edit: Realized the video was cut short. 3rd run ended at 3:10.
 
Last edited:

T72on1

Diabloii.Net Member
I'm totally not hyped for Cows at the moment to be honest. I just don't enjoy running them. So maybe I'll get a few hours in, maybe even a full set just because (if I find the time). Or maybe nothing at all.

Good luck to everyone though !!!
 

T72on1

Diabloii.Net Member
Ok, ok, I already like Cows a little bit more:

Screenshot137.jpg

Dropped on my second run.

First runs were horribly slow. Something in the 5:10 range. The run times I posted earlier were definitely the result of issues with RunCounter, as mentioned by @Gripphon. Thanks for the detailed analysis on RunCounter btw. I was losing a lot of time due to needing to drink mana pots all the time, so after watching @NanoMist's video of his Nova Sorc, and seeing he had 2000+ mana, I decided to switch some stuff around, ending up with some 1500+ mana and 5mpk from my boots. More than enough to keep my mana full at all times.

That way the next few runs went a lot faster. Not in the least because watching Nanomist's video taught me something else (which I already knew, but always keep neglecting): proper herding is key for Cows. I always started to shoot fireballs the moment I saw a group. Now I teleport past, herd a few groups, and then start to attack. This way dropping Meteors is far more efficient, and a combo of those and FB ususally does the trick.
 

Yann

Diabloii.Net Member
I remember I had a level 99 nova sorc which I played mostly in cows in 1.09 HC on the realms. I deeded her because I didn't think I needed the full rejuv, still one of my favourite characters.
Yours wins style points as well for wearing a scythe infinity
 

Grape

Diabloii.Net Member
Great job @NanoMist !

Code:
Sparking Grand Charm of Sustenance
Grand Charm
Required Level: 42
Fingerprint: 0x4e8c728b
Item Level: 81
Version: Expansion 1.10+
+1 to Lightning Skills (Sorceress Only)
+29 to Life
Maybe I'll test some lightning sorc too!

Some brief tests: 99 FCR wizzy & no BO variant was too fragile, even maybe little faster. I'll leave the pre-buff hassle to someone else and stick to my CtA & spirit switch.

P5 3:20 avg now after 15 runs.
 
Last edited:
Top