Jump to content
You need to play a total of 20 battles to post in this section.
ResonantOne

So is Patch 0.5.9 going to fix the broken RNG?

42 comments in this topic

Recommended Posts

Members
76 posts
3,240 battles

I didn't even notice that OP claimed he had a 90% hit rate. That's pretty hilarious. Sorry, no. The top players on the NA server have a hit rate in the upper 30s of percent, and you claim that you had more than double that? Riiight.

 

yes, I exaggerated there.  But it was higher than it is now.

Share this post


Link to post
Share on other sites
Members
76 posts
3,240 battles

 

Clustering illusion:

The tendency to erroneously consider the inevitable "streaks" or "clusters" arising in small samples from random distributions to be non-random. The illusion is caused by a human tendency to underpredict the amount of variability likely to appear in a small sample of random or semi-random data.

 

Starting to feel like the convenient use of the "No True Scotsman" argument is being applied.  But let's assume you're right and the problem is solely me and my ability to play the game. 

 

 For a point of comparison I looked at your stats, and how you do with the Kongo.  

Kongo 5 BB  Japan 21 52.38% 62,478 1,630 2.9 1.4 3.8 52% 27% 0% 1,615

Versus my performance 

Kongo 5 BB  Japan 53 41.51% 31,271 722 1.8 1.0 0.8 47% 22% 0% 1,024

How is it that there is such a significant difference in outcomes.  My accuracy is only 5% less than yours, but my KDR is around 60% of yours, as well as damage and XP being less than half.

 

Or something more comparable: St. Louis.

 

Your performance:

St. Louis 3 CA/CL  USA 23 52.17% 39,274 1,035 2.6 1.6 0.0 39% 39% 0% 1,431

My Performance:

St. Louis 3 CA/CL  USA 22 40.91% 19,490 488 1.1 0.7 0.0 36% 38% 0% 916

 

I do find this comparison interesting though in our Myokos:

Japan 20 50.00% 57,570 1,785 1.4 1.0 2.9 30% 33% 4%

1,453

Versus my stats:

Japan 21 61.90% 28,291 885 1.4 0.5 0.5 62% 27% 11% 852

Of course that goes to show W/L has little to do with your own performance, and more to do with the other players.

 

So, what is it that you are doing so radically different?  Are you getting citadels with nearly ever shot?  Sacrificing digital goats to RNG?

 

I know I'm not a zero rank player, but I can consistantly get to at least Tier 14, so I'm not a complete slouch.  The question is, what changed since that patch?  I cannot conceive a radical drop in my skill either particularly since the game feels different with some ships.  Perhaps the external variables are just too much to figure out (just enough bad players on my team or theirs, a fluke bad choice, or lucky shot) but to just chalk it up to just being a statistical anomaly, sorry, I have a hard time with that, particularly when games like this are subject to all sorts of other interference.  The point is I want to improve, and it pisses me off when suddenly things that are tried and true stop working with the same ship, and the only change has been the update.

Edited by ResonantOne

Share this post


Link to post
Share on other sites
21
[G0ATS]
Beta Testers
295 posts
6,805 battles

Of course that goes to show W/L has little to do with your own performance, and more to do with the other players.

 

So, what is it that you are doing so radically different?  Are you getting citadels with nearly ever shot?  Sacrificing digital goats to RNG?

 

I know I'm not a zero rank player, but I can consistantly get to at least Tier 14, so I'm not a complete slouch.  The question is, what changed since that patch?  I cannot conceive a radical drop in my skill either particularly since the game feels different with some ships.  Perhaps the external variables are just too much to figure out (just enough bad players on my team or theirs, a fluke bad choice, or lucky shot) but to just chalk it up to just being a statistical anomaly, sorry, I have a hard time with that, particularly when games like this are subject to all sorts of other interference.  The point is I want to improve, and it pisses me off when suddenly things that are tried and true stop working with the same ship, and the only change has been the update.

 

Its not just about hitting - its about knowing where to hit.

Its also knowing when to shoot and when to hold of for 3-5 seconds and get a higher % for a citadel.

It comes with time and experience

  • Cool 1

Share this post


Link to post
Share on other sites
105
[_TWP_]
Members
611 posts
6,767 battles

I know I'm not a zero rank player, but I can consistantly get to at least Tier 14, so I'm not a complete slouch.  The question is, what changed since that patch?  I cannot conceive a radical drop in my skill either particularly since the game feels different with some ships.  Perhaps the external variables are just too much to figure out (just enough bad players on my team or theirs, a fluke bad choice, or lucky shot) but to just chalk it up to just being a statistical anomaly, sorry, I have a hard time with that, particularly when games like this are subject to all sorts of other interference.  The point is I want to improve, and it pisses me off when suddenly things that are tried and true stop working with the same ship, and the only change has been the update.

 

Your stats on Warships.Today suggest that things have not changed for you since the 5.7 patch, though.

 

From 2016.05.13 to 2016.06.26, here are your battleships:

 

New York (37 battles): 56.76% win rate, 33864 average damage, 837 average XP, 2.8 KDR, 23% main battery hit rate

Kongō (24 battles): 54.17% win rate, 33240 average damage, 863 average XP, 3.0 KDR, 20% main battery hit rate

ARP Kirishima (1 battle): 0.00% win rate, 5,680 average damage, 98 average XP, 0.0 KDR, 16% main battery hit rate

ARP Haruna (1 battle): 100.00% win rate, 9486 average damage, 622 average XP, 0.0 KDR, 15% main battery hit rate

 

Patch 5.7 rolled out on 2016.06.22, so it may have some bearing on those numbers as four days under that patch are included. However, look at your numbers for 2016.06.26 to 2016.07.26:

 

New York (4 battles): 50.00% win rate, 45972 average damage, 814 average XP, 2.5 KDR, 24% main battery hit rate

Kongō (25 battles): 32.00% win rate, 31639 average damage, 619 average XP, 1.3 KDR, 23% main battery hit rate

ARP Haruna (2 battles): 50.00% win rate, 49496 average damage, 770 average XP, 1.5 KDR, 27% main battery hit rate

 

There's a big disparity in the number of battles in New York in the two intervals, so I don't think it's a very useful comparison.

However, in each interval, you've got a similar number of battles in Kongō or a Kongō clone, so let's look at these more closely. Factoring the battles in Kirishima and Haruna into the numbers, we get these:

 

Kongō (all, first interval): 53.85% win rate, 31266.38 average damage, 824 average XP, 2.77 KDR, 19.7% main battery hit rate

Kongō (all, second interval): 33.33% win rate, 32961.74 average damage, 630 average XP, 1.31 KDR, 23.3% main battery hit rate

 

Your average damage and your main battery hit rate for this family of ships have gone up since the 5.7 patch, not down.

 

Your main battery hit rates with DDs have also improved. Your main battery hit rates with CAs and CLs is a mixed bag: Königsberg shows a small improvement, Kirov shows a small dip, Cleveland and Emden both show sharp decline. Aoba appears in one set, Myōkō in the other - I wouldn't put too much stock into comparing different ships, but they do have similar artillery, so you'd probably expect fairly similar accuracy; Aoba (pre-patch interval) had a 24% main battery hit rate, Myōkō (post-patch interval) had a 26% hit rate.

 

I think what's happening is that you're more vividly noticing those times when a perfect shot brackets the target - those are a lot more memorable than the perfectly-average shots where two of eight shells hit, so it's completely normal for that to stand out, and that's coloring your perception of what's going on. The numbers reported on Warships.Today don't show any particularly notable changes for you.

Edited by BGolightly
  • Cool 2

Share this post


Link to post
Share on other sites
Members
3,959 posts
7,738 battles

                             
Of course that goes to show W/L has little to do with your own performance, and more to do with the other players.

So, what is it that you are doing so radically different?  Are you getting citadels with nearly ever shot?  Sacrificing digital goats to RNG?

 

 

On the contrary, W/L has a lot to do with your own performance. While midnight goat sacrifices are definitely important, it also helps to know where to sail and what to shoot (i.e. killing a DD that is trying to cap is much more beneficial than killing an afk DD). It also helps a lot to know which parts of which target you should aim for and what ammo to use. Notice how your hit rate is not much lower than Vakhnenko's, but your damage done is almost half of his?

 

Share this post


Link to post
Share on other sites
Members
5,202 posts
3,461 battles

But Derpitz... that thing has like.. 1.5 sigma, which makes the guns derp hard beyond 12km.

 

Then why are idiots firing it at 20km and praying they will hit something more than 1/2 the time? That right there drives me nuts. Sit too far away in your BBB, lob shots at an impossible distance then flood the forums with "wthdispersion/RNG" threads.

Share this post


Link to post
Share on other sites
3,828
Beta Testers, In AlfaTesters
8,146 posts
10,605 battles

let's assume you're right and the problem is solely me and my ability to play the game. 

 

How is it that there is such a significant difference in outcomes.  My accuracy is only 5% less than yours, but my KDR is around 60% of yours, as well as damage and XP being less than half.

 

I do find this comparison interesting though in our Myokos:

Of course that goes to show W/L has little to do with your own performance, and more to do with the other players.

 

So, what is it that you are doing so radically different?  Are you getting citadels with nearly ever shot?

 

Just to be clear: I wasn't critiquing your ability, but rather your understanding of probability and statistics. Clustering illusion and confirmation bias can easily create a perception that "something must have changed in the past patch", when that is not the case. To truly prove that something has indeed changed (assuming that WG cannot be trusted and you can't access their code) you'd need to collect enough data (while carefully controlling the variables) and do statistical analysis to demonstrate a statistically significant difference between dispersion from patch to patch (with a good p-value too, 0.05 wouldn't really cut it here).

 

Now for differences in performance. First, few comments:

- XP is not directly comparable, unless you always run premium (which I do). Unfortunately, premium XP bonus gets counted by the stats aggregators. It's very stupid, but it is what it is. 

- for a small sample (and n~=20 is a small sample indeed) win rate is fairly meaningless, so comparing our Myoko W/L rate provides little useful information. Win rates start becoming a good indicator when you hit a ~few hundred games, or so.

- don't forget, damage isn't everything either. 10k damage on a DD and on a BB are two completely different animals.

 

Now as to what I do differently. Well, I don't know, since I don't know how you play :)

 

But generally, in a BB I normally try to get into a medium range (~10-13km, maybe even closer), since that's where citadels become easy to get -- less effects of RNG, and you don't have to aim that well (I am not very good at leading, to be honest). And yes, I do that even in IJN BBs -- sniping from a distance might work for others, but not for me. The key here is to not get killed yourself -- you can't do damage once you die. Don't get in a position where you'll get focused by multiple ships, angle and dodge carefully, anticipate enemy DD and CV actions. Don't tunnel vision on someone, constantly look around to see if some enemy is presenting a juice broadside. 

 

In a IJN cruiser, such as Myoko, don't forget to use AP when there is a decent chance of citadeling somebody (mostly broadising enemy cruisers inside 10-12km, but you can also citadel some BBs with relatively thin armor from close up). Plan ahead -- loading guns takes 15 seconds, so start loading AP ahead of time when you expect the enemy cruiser to turn or show up from behind an island). Light BBs on fire when they blow their repair, IJN 203mm HE is good at that. Again, don't get killed -- don't broadside, don't get focused, don't outrun your team and get in trouble. That sort of thing.

Share this post


Link to post
Share on other sites
2,079
Alpha Tester
6,683 posts
3,338 battles

 

They all kind of have derp guns. Warspite seems to be one of the worst though. I've had games where every salvo got me a citadel, then the next game I get 40 hits, all overpens. She seems to be one of the most reliant on RNGsus' favour. Tirpitz too. 

 

I know exactly what you mean. Sometimes I'll have game where I do 115k damage and get 6-7 citadels and other times I'll have 30k games where I overpen and bounce everything. It's infuriating.

 

 

Warspite has 2.0 sigma, which makes her pretty accurate at all distances.

 

But Derpitz... that thing has like.. 1.5 sigma, which makes the guns derp hard beyond 12km.

 

Sigma only affects vertical dispersion, not horizontal. It also doesn't account for ships that are sailing at very hard-to-see angles and ships that turn.

Share this post


Link to post
Share on other sites
272
[-AGW-]
Members
1,154 posts
20,927 battles

 

 

I do find this comparison interesting though in our Myokos:

Japan 20 50.00% 57,570 1,785 1.4 1.0 2.9 30% 33% 4%

1,453

Versus my stats:

Japan 21 61.90% 28,291 885 1.4 0.5 0.5 62% 27% 11% 852

Of course that goes to show W/L has little to do with your own performance, and more to do with the other players.

 

So, what is it that you are doing so radically different?  Are you getting citadels with nearly ever shot?  Sacrificing digital goats to RNG?

 

 

I wonder if it might be a connectivity issue and dropped packets. I've had similar days where nothing seems to hit. Then I switch to a different ISP and my damage rate seems to improve significantly.

 

Doesn't explain how I got stuck with so many bad teams yesterday, but FWIW :)

Edited by NashW8

Share this post


Link to post
Share on other sites
Members
76 posts
3,240 battles

First off, thank you guys who are showing me things I have not noticed in my own statistics.  Seriously, it is appreciated.  :honoring: BGolightly & Vahknenko.

 

Secondly, Some of these points might help others and not just myself.  I'm a pretty intense person about these games and I get very frustrated when things appear to do random things when I perform what I think is the same action time and time again.  :red_button:  I've had some matches that are just headbangingly bad in the last few days of play as my stats show.  But last night, teaming up with a better person has turned that around.

 

Third, you pointed out something that I had forgotten: Premium time plus premium ships equals a huge difference in results of a match from the perspective of rewards.  That would explain much in that regard.  I about never run premium anything unless it was dropped into my account as a reward, or I am playing lowest tier with new players in my Emden.  (I've been playing a bunch with newer players, helping them learn the game, and so I know playing that low tier ship has affected my aggregate total damage and XP, but it seemed to extend to my higher tier ships.)

 

Plus there are a lot of pro tips which will help.  Thankfully I do many of those.  Like I said, I'm not a total slouch, but I'm nothing spectacular.  As for gameplay I had a match this morning playing Co-Op of all things that infuriated me with how smart the bots seemed to have gotten today.  What was more amazing is I ended up with 6 kills (I know, Co-Op big whoop) but... these were the smartest bots I ever fought!  I had to pull about every DD trick I could think of to both survive, and get all the breaks.  I'll post a video once I process it.  I had to copy it twice from Replay to get all what I wanted to see, and will composite it later.  Generally I wouldn't do it for Co-Op, but wow.

 

So, yes, ultimately, thank you for the statistical analysis and advice.  I now I have to look more carefully at my stats and understand better what I'm seeing, but it probably not help mah feelz when a game goes sour. :playing:

 

 

Share this post


Link to post
Share on other sites
Members
76 posts
3,240 battles

 

I wonder if it might be a connectivity issue and dropped packets. I've had similar days where nothing seems to hit. Then I switch to a different ISP and my damage rate seems to improve significantly.

 

Doesn't explain how I got stuck with so many bad teams yesterday, but FWIW :)

 

I thought about that too, but actually I have an extremely stable connection here in the midwest.  I constantly run a ping between 30-40 and a FPS of about 50-60.  That said, I have had times where I push the trigger, or try to turn, and nothing or a hangfire where it can take up to a full 2 seconds to fire or activate damage control.  On a couple cases, I looked to make sure it just wasn't reloading or on cooldown and saw it was clear.  So I dunno.  Client glitch, internet burp, I don't know.

 

And yeah, some days no amount of carrying will save a team determined to lose.

Share this post


Link to post
Share on other sites
2,201
[SYN]
Members
5,882 posts
13,316 battles

 

Yes I know that is part of my perception of the issue.  It might be something else, but I do know that my performance, with same ships has plummeted between the patches.  Ergo, something has changed.  The question is what?

 

Do you really want us to answer that for you?  :teethhappy:

Share this post


Link to post
Share on other sites
4,482
[WOLF1]
Beta Testers
14,425 posts
20,578 battles

There is RNG there is no " Broken RNG"

RNG is a necessary part of a game like this. It adds tension suspense and variety. Game would get stale very quickly with out it.

Share this post


Link to post
Share on other sites
Members
76 posts
3,240 battles

There is RNG there is no " Broken RNG"

RNG is a necessary part of a game like this. It adds tension suspense and variety. Game would get stale very quickly with out it.

 

I am not arguing against RNG.  I've played in another game that has perfect convergence and no RNG at all, save for one weapon and you see how unfair and boring that is.  Particularly in a game with instant kill functionality like it had.  I am not arguing against RNG, but I definitely wanted to understand what the heck is going on when there seemed to be significant changes in experience post 5.7.  The discussion about this has grown and is showing perceptions are definitely affecting game enjoyment more than statistics show.  Something that isn't bad for any gamer frustrated with their results or misunderstanding their own stats.

 

Share this post


Link to post
Share on other sites
272
[-AGW-]
Members
1,154 posts
20,927 battles

That said, I have had times where I push the trigger, or try to turn, and nothing or a hangfire where it can take up to a full 2 seconds to fire or activate damage control.  On a couple cases, I looked to make sure it just wasn't reloading or on cooldown and saw it was clear. 

 

Yes. That's precisely what I see with lag. It seems to take forever for the guns to fire. Keep en eye on the top left hand screen ping and lag numbers. I do believe even on a good connect you can get bad periods. That said, sometimes the server might be to blame, but that's not something we can measure.

  • Cool 1

Share this post


Link to post
Share on other sites
Members
76 posts
3,240 battles

 

Yes. That's precisely what I see with lag. It seems to take forever for the guns to fire. Keep en eye on the top left hand screen ping and lag numbers. I do believe even on a good connect you can get bad periods. That said, sometimes the server might be to blame, but that's not something we can measure.

 

Thx for the tip.  

 

BTW, full thundering typhoons of blistering blue barnacles for your avatar.

Share this post


Link to post
Share on other sites

  • Recently Browsing   0 members

    No registered users viewing this page.

×