Jump to content

Remembered one of the things I hate about CoH... its RNG.


Jeuraud

Recommended Posts

1 hour ago, Bopper said:

Can you confirm you had a 95% chance to hit? Base is 75%, but I don't know your targets or slotting.

When you are a LowB it's not hard to get to 95% because of Beginner's Luck, and the 1-miss Streakbreaker requires a min of 90.01%. Also your Combat Log tells you what your chance to hit was and what you rolled.

 

32 minutes ago, Generator said:

400 attacks an hour breaks down to only about 7 per minute.  The only way you're not making that many to-hit checks is if you're playing one of those "pure support" nitwits that think attack powers are the work of the devil.

Unless your a Soloing LowB, running Contact Missions. You know, run to your Contact, get a mission, and maybe even read what the mission is about. Run to the mission, do the mission, run back to the Contact and get a new mission, and maybe even read that mission info. Sometimes the new mission is where you were just at so you get to run back to the same spot. Eventually the Contact lets you call them, and eventually you finish with that Contact (Or kill them.), and go on to another Contact and start the whole process over. I use Ninja Run because I find it fun (Thank you CoHH Devs for giving it back to me with P2W.), and does not require a power choice. I'll not get a Travel Power until I start grouping, which I have yet to do in CoHH.

 

Now just because I'm not running my game your way, does not mean that I'm running my game the wrong way.

 

 

As for the rest of you I'm not sure why you feel you need to protect the CoH RNG-To Hit programing, I'm not asking you or even the Devs to do something. I'm asking if anybody had checked the code and found it to be clean. To be frank if I had remembered that to many the only Go-No Go gauge is statistics, at least here in the forums, I would not have asked my question.

  • Sad 1
Link to comment
Share on other sites

1 minute ago, Jeuraud said:

As for the rest of you I'm not sure why you feel you need to protect the CoH RNG-To Hit programing, I'm not asking you or even the Devs to do something. I'm asking if anybody had checked the code and found it to be clean. To be frank if I had remembered that to many the only Go-No Go gauge is statistics, at least here in the forums, I would not have asked my question.

So, because we disagree based on past experience, we are now somehow 'forum defenders'?

Holy Moly.

 

We presented the data we have - that Every Single Time this is brought up and someone has done _any_ research, code or otherwise, it has always seemed clean.

And this makes us somehow dismissive of your PoV?

Really?

Ugh. This poloraized 'agree or woe-is-me' is just too much these days...

 

All of the data, and tests, and code checks that we know of point the conclusion there is no problem.

Beyond that, what is it you expect to hear from the community, really?

  • Like 2
Link to comment
Share on other sites

19 minutes ago, Jeuraud said:

Now just because I'm not running my game your way, does not mean that I'm running my game the wrong way.

I'm neither telling you how to play the game, nor whether that's right or wrong. But if you're talking about attacks as number of rolls per hour, you can only run the clock on that hour when you're in combat.  Time spent interacting with the Contact, traveling to and from, fiddling with Wentworth, or getting up and going to the kitchen for Oreos shouldn't factor into that.

  • Like 2
  • Haha 1
Link to comment
Share on other sites

15 minutes ago, Jeuraud said:

As for the rest of you I'm not sure why you feel you need to protect the CoH RNG-To Hit programing, I'm not asking you or even the Devs to do something. I'm asking if anybody had checked the code and found it to be clean.

And we're telling you that yes it has, many many times over the last decade and a half, and it's always been clean, and nothing has changed recently that should have affected that. We're also telling you that the things you think seem wrong are in fact not wrong due to basic statistics. You're literally asking a question about statistics; of course the responses are going to discuss statistics.

 

We checked it by examining its behavior rather than literally reading the source code, because until recently, reading the source code wasn't even an option. If you feel that reading the source code might turn up something fishy that other analyses didn't, go for it. But if you want somebody else to do it for you, you're going to have to convince them it's worth doing. Right now, you've got a hunch and some objectively unremarkable anecdotes.

  • Like 7
  • Thanks 1
Link to comment
Share on other sites

28 minutes ago, Jeuraud said:

To be frank if I had remembered that to many the only Go-No Go gauge is statistics, at least here in the forums, I would not have asked my question.

How would you answer this question without statistics? Back on live, we didn't have the source code to look these things up, thus simple testing using statistics was required. We have the source code now, so you're welcome to look at it. But I suspect a simple RNG calculation would be hard to muck up. But even if it was, the streak breaker fixes it for you.


PPM Information Guide               Survivability Tool                  Interface DoT Procs Guide

Time Manipulation Guide             Bopper Builds                      +HP/+Regen Proc Cheat Sheet

Super Pack Drop Percentages       Recharge Guide                   Base Empowerment: Temp Powers


Bopper's Tools & Formulas                         Mids' Reborn                       

Link to comment
Share on other sites

23 minutes ago, jubakumbi said:

All of the data, and tests, and code checks that we know of point the conclusion there is no problem.

Beyond that, what is it you expect to hear from the community, really?

I asked this community a specific question, and I have yet to read a reply to my question. What I have read so far is peoples opinion about there not being a problem, because Statistically, there is no problem. Statistics has never been my Go-No Go gauge, for anything.

 

1 minute ago, Hopeling said:

because until recently, reading the source code wasn't even an option.

And this is why I asked my question, of this community, because I know there are people in this community who have parsed the code, so I was wondering if there was anyone here who had checked the RNG-To Hit section of the program. Now if someone in this community replies that they checked the code and could find nothing wrong,  then I'm going to have to do what I've been doing since 2004, shrug and accept that Murphy is going to royally shaft me every so often, and still not give a fuck that Statistically there is nothing wrong.

Link to comment
Share on other sites

You just had a short run of bad luck. I do a lot of testing on the game. I have done 1000s of attacks on a test dummy with a 95% chance to hit. Everytime I check my hit probability it hovers right at 95% (a hair more, thanks to streak breaker). Like I said earlier, getting 4 misses out of 7 roles will happen 1 out of 5,332. It is rare, but it's not that rare.

  • Like 2

PPM Information Guide               Survivability Tool                  Interface DoT Procs Guide

Time Manipulation Guide             Bopper Builds                      +HP/+Regen Proc Cheat Sheet

Super Pack Drop Percentages       Recharge Guide                   Base Empowerment: Temp Powers


Bopper's Tools & Formulas                         Mids' Reborn                       

Link to comment
Share on other sites

21 minutes ago, Jeuraud said:

I asked this community a specific question, and I have yet to read a reply to my question. What I have read so far is peoples opinion about there not being a problem, because Statistically, there is no problem. Statistics has never been my Go-No Go gauge, for anything.

 

And this is why I asked my question, of this community, because I know there are people in this community who have parsed the code, so I was wondering if there was anyone here who had checked the RNG-To Hit section of the program. Now if someone in this community replies that they checked the code and could find nothing wrong,  then I'm going to have to do what I've been doing since 2004, shrug and accept that Murphy is going to royally shaft me every so often, and still not give a fuck that Statistically there is nothing wrong.

The only way to prove the system works is to use it and test it and look at the statistics - that's how testing works.

No matter how much you look at the code, testing it still based in confirming that statistically it is doing what is intended.

The system is WAI from every test the community can find or has done.

 

Honestly, It seems like until someone you trust as a code auditor walks you through the code line by line and tells you it's "OK", you will not be 'satisfied'...good luck with that...

 

  • Like 1
Link to comment
Share on other sites

30 minutes ago, Jeuraud said:

Statistics has never been my Go-No Go gauge, for anything.

Identification = Not an Engineer

 

I don't mean to be insulting, but statistical evaluation is exactly how you determine whether an RNG is functioning properly.

  • Like 1
  • Thanks 1
Link to comment
Share on other sites

3 hours ago, GM Fiddleback said:

I just put this here for reference, not as commentary.: https://paragonwiki.com/wiki/Attack_Mechanics

To Start with, I've got Murphy (yes, That Murphy...according to family legend) in my family tree, so my whole life is a cold streak. I've had more DnD characters die from random encounters and nat 1's than anyone I know. But...

I'm actually making my luck worse by continuing to use Brawl and beginner ranged attacks. Might explain why 9 times out of 10, the first attack after popping a yellow is a miss. My natural bad luck, plus nerfing the streakbreaker by using Brawl (with just 1 Acc in the native slot) in my attack chain. I usually take Boxing, Tough, and Weave on my melee toons, but not until the 30's when I can put 4 slots in to use 2 End Red along with Acc, Dmg Res, or Def (respectively). Might have to grab Boxing earlier just to slot it with  Acc and drop Brawl from my chain sooner.

 

Thanks Fiddleback for a very useful link. In all these years I never thought to actually look up how the mechanics work. Now I can return to collecting Debt Badges knowing it's just my bad luck and not poor attack chain choices. :-D

Once you turn your Path to the Darkside, forever will it dominate your Fate... and you get Dental. - Dark Acolyte, Warcraft III

Link to comment
Share on other sites

Fun fact. Mike Trout, easily the greatest baseball player of this decade, has a lifetime 0.306 batting average. But for a 21 at bat stretch in 2018, Mike Trout went hitless. The odds of a .306 hitter to have that run of bad luck? 1 in 2,141. Sometimes rare things happen. 

  • Thanks 1
  • Haha 1

PPM Information Guide               Survivability Tool                  Interface DoT Procs Guide

Time Manipulation Guide             Bopper Builds                      +HP/+Regen Proc Cheat Sheet

Super Pack Drop Percentages       Recharge Guide                   Base Empowerment: Temp Powers


Bopper's Tools & Formulas                         Mids' Reborn                       

Link to comment
Share on other sites

I've seen quite a lot of odd behavior out of the RNG. Not odd as in "bugged", just odd as in "wow, wtf". Like having 75% to hit and landing 3 swings out of 10, and only then because the streak breaker forced them to hit. Nearly losing a fight to a white-con lieutenant from full health was making me quite peeved, I'll tell you what. :)

 

Or fighting a couple of arachnos flunkies at 8th level or so, and their shotguns with like 30% to knockdown on each attack knocked me down every. Single. Hit. I was on my butt for like 70% of the fight, and nearly lost because of it.

 

Or jumping into a pack of mobs on my brute with 50% defense (and no debuffs), and instantly face-planting, because every mob rolled 5% or less.

 

Or my bot/ff mastermind having the bots running around at capped defense, and having a mob obliterate all my tier-1's one after the other because it rolled 5% or less with each attack in succession.

 

Stuff like that isn't common, but it makes for some serious "wtf?!" moments.

Edited by WumpusRat
  • Like 1
Link to comment
Share on other sites

1 hour ago, biostem said:

The problem, IMO, is that people interpret things like a 75% chance to hit as "I WILL hit 3 out of every 4 attacks", not "over the span of 10000 trials things will work out to a 75% hit/25% miss ratio".

Except you don't need 10000 attempts to work out the odds of an event occurring. 

If you have a 50% chance to hit, then the odds are 50/50 that you hit or miss.

Getting the same result twice in a row (hit or miss) is 25%.

Getting the same result three times in a row is 12.5%

And so on.

 

Is it statistically possible to miss 10 times in a row with a 75% to hit? Of course. The odds of it happening, however, is .0000953674%

 

That's why people notice extreme outliers.

Link to comment
Share on other sites

3 minutes ago, WumpusRat said:

Except you don't need 10000 attempts to work out the odds of an event occurring. 

If you have a 50% chance to hit, then the odds are 50/50 that you hit or miss.

Getting the same result twice in a row (hit or miss) is 25%.

Getting the same result three times in a row is 12.5%

And so on.

 

Is it statistically possible to miss 10 times in a row with a 75% to hit? Of course. The odds of it happening, however, is .0000953674%

 

That's why people notice extreme outliers.

Indeed, which is why you need as large a sample as possible, so as to rule out any anecdotal feelings you may have or just happening to hit a bad/miss streak.  Low probability != impossibility, so we need a large sample to make sure.

  • Like 1
Link to comment
Share on other sites

Just now, biostem said:

Indeed, which is why you need as large a sample as possible, so as to rule out any anecdotal feelings you may have or just happening to hit a bad/miss streak.  Low probability != impossibility, so we need a large sample to make sure.

You only need a large sample size if you're trying to prove that, on average, it works out that way. This is simply pointing out a rare outlier. Not saying "it's broken because X happened", but simply noticing a rare occurrence. Same as noticing you crit five times in a row as a scrapper. "Wow, lucky rng."

Link to comment
Share on other sites

5 minutes ago, WumpusRat said:

You only need a large sample size if you're trying to prove that, on average, it works out that way. This is simply pointing out a rare outlier. Not saying "it's broken because X happened", but simply noticing a rare occurrence. Same as noticing you crit five times in a row as a scrapper. "Wow, lucky rng."

If, as you stated, all you're doing is pointing out a lucky or unlucky streak, with no implication that something is wrong/broken, then go for it!

Link to comment
Share on other sites

13 hours ago, Jeuraud said:

This morning I was trying a couple of Corrupter ideas, and I was having to take a knee, or even dying... a lot. I started paying closer attention and noticed I was missing a lot. This was especially noticeable with /Dark's TG, which I was missing 2 and even 3 times in a row. I checked my combat record and was getting a lot of this,

 

Hit

Hit

Miss

Streakbreaker

Miss

Streakbreaker

Hit

Miss

Streakbreaker

Miss

Streakbreaker

 

Murphy was having fun with me this morning.

I have never understood why the CoH Devs took their RNG to the 100th place; what I do understand is how frustrating it is to roll a 95 and still have 99 chances to miss, and there is absolutely nothing I can do about it.

 

One of the things I've also noticed is that the MOBs consistently hit with their first two attacks, and then will start missing. I've also noticed that MOBs consistently roll lower numbers, and I roll higher numbers. I've always wondered if there is something in the Attack formulas that causes this, or is it just viewer bias, because I'm only looking at the combat record because I'm frustrated.

 

So now that the code is out I'm wondering if any of you code people have looked at the Attack formulas?

Go play another game then IF YOU CAN'T DEAL WITH THIS.  Play chess or checkers or go.   I hate people like you.

I went to Ouroboros all i got was this lousy secret!

 

COH bomp bomp: 

 

 

Link to comment
Share on other sites

1 hour ago, wjrasmussen said:

Go play another game then IF YOU CAN'T DEAL WITH THIS.  Play chess or checkers or go.   I hate people like you.

Look, I find the OP's logic wonky and his demeanor somewhat demanding, but sheesh...hating someone for that?  I'd hate every boss, every significant other, every child, and most of my friends...

 

I'm really not a fan of people who say - "If you don't like something just leave," because it suggests that things shouldn't be changed or examined...and the OP is doing exactly that...

 

My thought might be if his posting is getting you that aggravated, maybe just ignore the thread?

  • Like 2

"The opposite of a fact is falsehood, but the opposite of one profound truth may very well be another profound truth." - Niels Bohr

 

Global Handle: @JusticeBeliever ... Home servers on Live: Guardian ... Playing on: Everlasting

Link to comment
Share on other sites

5 hours ago, Jeuraud said:

I asked this community a specific question, and I have yet to read a reply to my question. What I have read so far is peoples opinion about there not being a problem, because Statistically, there is no problem. Statistics has never been my Go-No Go gauge, for anything.

Let's say your favorite coffee company says they have invented a new coffee with double the caffeine of their last product (I know I would buy it).  And they want to see if it's really true...How would said coffee company test this claim?  Taste test?  Does it taste like it has double the caffeine?  Or do they run a chemical test on their old and new product and see what the quantities of caffeine are...The feature they implemented was a chemical feature, so it requires a chemical test to verify.

 

Same thing here...they were coding for a statistical probability.  So the best evidence of whether it is working or not is statistical testing...Does X event happen Y many times in this circumstance...Statistics is the ONLY way to know it's working...

 

5 hours ago, Jeuraud said:

And this is why I asked my question, of this community, because I know there are people in this community who have parsed the code, so I was wondering if there was anyone here who had checked the RNG-To Hit section of the program. Now if someone in this community replies that they checked the code and could find nothing wrong,  then I'm going to have to do what I've been doing since 2004, shrug and accept that Murphy is going to royally shaft me every so often, and still not give a fuck that Statistically there is nothing wrong.

I think it's fair to ask the community if it's already been checked...but that's it...no one should need to go check for you, unless they a.) owe you a favor, b.) you are offering cookies, or c.) have made a compelling case for someone else to take time away from the things they normally do to check the code.  Since no one is jumping in with the level of specificity you desire, I think you can either a.) call in favors, b.) start baking, or c.) start tracking the data over a large period of time...

"The opposite of a fact is falsehood, but the opposite of one profound truth may very well be another profound truth." - Niels Bohr

 

Global Handle: @JusticeBeliever ... Home servers on Live: Guardian ... Playing on: Everlasting

Link to comment
Share on other sites

On 7/29/2019 at 11:10 AM, Bentley Berkeley said:

accuracy in general early on is abit rough

Beginners Luck went live at Issue 12. You have a 90% chance to hit even level minions at Level 1 and it creeps to 75 at Level 20. New toons actually have it easier. There is no need to slot for accuracy at Level 12. Endredux will serve better in attacks than accuracy until 20, then replace 1 endredux with 1 ACC. 

  • Like 3

________________

Freedom toons:

Illuminata

Phoebros

Mim

Ogrebane

Link to comment
Share on other sites

1 hour ago, cejmp said:

Beginners Luck went live at Issue 12. You have a 90% chance to hit even level minions at Level 1 and it creeps to 75 at Level 20. New toons actually have it easier. There is no need to slot for accuracy at Level 12. Endredux will serve better in attacks than accuracy until 20, then replace 1 endredux with 1 ACC. 

Who fights equal lvl mobs at low lvls? I exclusively am facing at min yellow if not orange and reds at those lvls for MOAR XP!

 

2nd a lvl 15 acc Io still gives a 19% boost more then BL. and that bonus from BL goes down  steadily as you lvl up from 1, By lvl 12 its nearly gone. I like lvl 15 Ios because their bonus is enough I can not need to worry about reslotting again until 17 when I start using a collection of variou sets for franken slotting to get usually around 40% or so to 4 aspects out of 4 slots at a lvl range when most players are getting much less effect out of each slot.

 

I only use attuned set IOs after lvl 17 for attacks and defenses. only things like hasten get common IOs after that.

 

I am also taking into account the fact skulls, a major enemy for that early lvl range uses dark, and will debuff the frell out of your to hit so having your ability to hit buffed up a fair bit is damn handy in KR during those skull story arcs.

 

Link to comment
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
×
×
  • Create New...