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

[Snakebit]

>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!

 

You still have higher accuracy.

 

>By lvl 12 its nearly gone.

 

No it isn't. 

 

>skulls,

 

You still have a higher accuracy.

 

 

[Jeuraud]

On 7/29/2019 at 9:36 PM, justicebeliever said:

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,

And where in this question,

On 7/29/2019 at 6:36 AM, Jeuraud said:

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

am I making any demands, hells this is not even the Suggestions Forum; you all are reading into my question, and that is not on me.

Worse, many of the posters in this thread are saying that because Statistically there is no problem with the CoH Attack-RNG, my feelings of frustration are invalid... that's just fucked.

 

On 7/29/2019 at 8:04 PM, 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.

I'm a box holder for CoH, CoV, and maybe Going Rogue (Cant remember if programs still came in boxes at that time, and the boxes came up missing on my last move.). I've been playing CoH since 2004, and have no fricken clue how much money I dropped on CoH over the years. I was frustrated with the CoH Attack-RNG right out of the CoH box, when there was not even a Streakbreaker, but I still payed my monthly subscription and bought new content.  I helped keep CoHV running with my hard earned cash... until a bunch of Beancounters decided it was not cost effective to keep it running, which leads to my final statement.

 

I've been around almost 60 years, and a Maintenance Tech for a good part of that time, I have seen Management decide that it's not cost effective to 'truly' fix an issue many times, so I've never trusted the CoHV Devs answers about the CoH Attack-RNG. Now if someone from this community tells me that they have looked at the code and could not find anything wrong then I can accept their answer, because their time is not money, and they don't have Management standing over them. Also this code is not theirs so they have no reason to defend it; I've had to deal with this from Engineers, as an MT, as well.

 

 

 

 

 

 

Edited July 31, 2019 by Jeuraud

[retiarius]

I fail to understand what would close examination of the code would reveal that repeated testing (experience) hasn't already revealed.

 

If the presented numbers are borne out by those who've done long, arduous testing and recording of the results, and those results match the expressed percentages, then how would looking at the code net anything of additional value with respect to the expressed percentages?

When I feel frustrated by this game, I play a different game or do something else.  I've found this to be an effective way of mitigating the frustration.

[Xeres]

Back in live I slotted 2 accuracy in every attack to mitigate this a bit.  It seems like at times there's the hit drought even though the numbers will break even after the whole gaming session is tallied.  Doesn't make it less annoying.

 

Heck the other night I left a mission just before the end to upgrade my accuracy IOs because at the time it seemed like I was missing a lot,

[Llewellyn Blackwell]

12 hours ago, cejmp said:

>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!

 

You still have higher accuracy.

 

>By lvl 12 its nearly gone.

 

No it isn't. 

 

>skulls,

 

You still have a higher accuracy.

 

 

 

having a bit of trouble with quoting on the forums eh bud?;)

 

While its been a long time my memory for such discussions on the forums isnt to shabby. I seem to recall BL being expalined as a base like 15% bonus, that drops by 1% each lvl after lvl 1. Thats why its called beginners luck after all.  Your not really a beginner anymore by the teens.

 

A lvl 15 acc Io is 19% that is actually a really healthy boost. and since it wont go bad can be left till a  major Set twink out usually done by most at 22 when sets like lotg become available. keep in mind we are discussing lvls the vas majority of players are largely zooming through now days. I mean even street sweeping Ill go from 1-22 in a casual play session or two. So its basically a non issue.

 

 

[justicebeliever]

9 hours ago, Jeuraud said:

And where in this question,

am I making any demands, hells this is not even the Suggestions Forum; you all are reading into my question, and that is not on me.

Worse, many of the posters in this thread are saying that because Statistically there is no problem with the CoH Attack-RNG, my feelings of frustration are invalid... that's just fucked.

I don't know you that you were making demands, but you have asked several times and no one has stepped up to your satisfaction...That's fair that you aren't satisfied, the point is just that other people don't need to take time out of their schedules to satisfy you either...

 

And I'm certainly not suggesting your feelings are invalid...just that software code doesn't work on feelings but rather fact and logic.  You need evidence to sign off on code in the first place (unit/beta testing), and you need evidence to suggest it's not working.  Feeling's aren't evidence...So have your feelings, but without evidence, it's unlikely that anyone will pay much attention...

 

9 hours ago, Jeuraud said:

Now if someone from this community tells me that they have looked at the code and could not find anything wrong then I can accept their answer, because their time is not money, and they don't have Management standing over them. Also this code is not theirs so they have no reason to defend it; I've had to deal with this from Engineers, as an MT, as well.

Except that for everyone in this community time is their money...I think that's the point.  Without evidence you keep asking (I didn't say demanding) people to take their time away to look at something that just "feels" wrong, despite the evidence of people who are suggesting otherwise...

 

I'm not attacking you here...Randomness and Coincidence are fickle friends and they are frustrating, more so in real life than in gaming even...I am however, just responding to what you are saying...

[jubakumbi]

13 minutes ago, justicebeliever said:

Without evidence you keep asking (I didn't say demanding) people to take their time away to look at something that just "feels" wrong, despite the evidence of people who are suggesting otherwise...

 

This is the crux of the whole issue with players talking about anecdotal evidence in all games all the time.

Humans seek patterns and if they don't find them to thier liking, they start making them up...it's what we do, it's a built-in mechanism to make sure we don't go insane with worry about the animals the might be right over that hill hunting us.

 

The outright denial of evidence and facts that we don't feel are correct is one of our greatest super-powers as humans.

 

Just look at the number of technologists that work with code and statistics on a daily basis who still wear the 'right shirt' on game day so 'thier' team will win...

 

The mind is a terrible thing. 🙂

 

[daHob]

This old canard. People have been twitch about game RNG since Ultima Online. 

 

First, real randomness is streaky. Streakiness is expected.

 

Is the RNG on a computer truly random? No.

Does that actually matter for the purposes of running a computer game? No.

Would that matter if you were making a scientific simulation where the statistical outcome is going to be used to predict things about the world? Probably.

 

Even if the RNG is on the game is bad, it's not going to be bad enough for an individual end user to be able to really notice. Certainly a string of 10 misses is /WAY/ to low of a sample size to indicate anything. Even if it were heavily biased there are hundreds of players pulling thousands of rolls from the same seed, more or less at random intervals, which churns the distribution you experience a lot. Also, it's pretty highly likely that you are pulling from several different RNGs, at least one per server. 

 

It's like people freaking out about how fair their d20 are. Ok, maybe 7 comes up 4% or the time instead of 5% over a VERY LARGE number of rolls. But you are only going to roll less than five thousand meaningful rolls in your gaming career, probably not all with that die. The probability skew from normal variance over that small a sample set is tiny. Certainly not enough to be noticeable.

 

Go ahead, but the bad one in the dice jail...

 

 

[Corruption]

17 hours ago, Jeuraud said:

Worse, many of the posters in this thread are saying that because Statistically there is no problem with the CoH Attack-RNG, my feelings of frustration are invalid... that's just fucked.

If there's no problem statistically, then what exactly is the problem? Sometimes you run into bad luck with the RNG, that's probability for you.

Edited July 31, 2019 by Corruption

[RikOz]

I was thinking about this just today, while once again whiffing my way through CoT ghosts/daemons/whichever they are.

 

I came to the conclusion that there were certain design decisions early on in development that were either lazy, or just poorly thought out, and this is most blatantly evident in the Accuracy/ToHit department, because this one area in particular directly leads to intense player frustration.

 

The "lazy" or "poorly thought out" decision in question here was making "hit or miss" a straight (to all appearances) yes/no question. You either hit and do damage, or you completely miss and do no damage. This is a "user experience" issue.

 

I'll compare to World of Warcraft, since that's the only other MMORPG I'm extensively familiar with.

 

In WoW, there are multiple mechanics that determine defense. There is your Armor, which is a number. There is Dodge, which is a number. There is Block, which is a number. And there is Parry, which is a number. All of these come into play when determining whether an attack hits or misses. And here is why I call this a "user experience" issue: When the target mob (or player, in PvP) causes you to "miss", there is a relevant sound effect that plays based on why you missed, and often a corresponding animation. You hear and see different things depending on the result of each "roll". You hear the sound of your weapon clanking against armor, clunking against a shield (for Block), or sliding off the other guy's weapon (for Parry). The point is that, while you may not be dealing damage, you are still making some kind of contact. The only genuine "swing and a miss" (assuming an opponent of roughly equal level) is when your opponent Dodges (and that still produces appropriate sound and/or animation). All of the other MMOs I've briefly tried (none for more than a month or so) did more or less the same thing.

 

In CoH, by contrast, you either Hit (POW!) or Miss (*whiff*). So if you end up with a series of unsuccessful attacks, the only audio-visual feedback you get is swinging at air, over and over and over and over, not to mention the word "MISS!" in big bold letters every single time.

 

This is the psychological difference between a basketball player missing shots that bounce off the rim or get blocked, and a player repeatedly failing to even hit the backboard. A baseball batter who grounds out is probably considerably less frustrated than another batter who strikes out on three pitches without even getting wood on the ball.

 

I mean, this type of discussion doesn't come up all that much on the WoW forums, because in that game the attention to user experience means the player never feels like they're just uselessly flailing at an enemy, when that enemy is at or near the same level as the player.

[Jeuraud]

Just a last little tidbit.

 

Lev 10 Wat/Ninj Sentinel

+0 Paper Mission, No Bosses, No AVs

3rd in a row with a +1 Named Lt.

 

Sentinel

10 attacks @ 62.92% avg

High-  93.32%

Low-  18.34%

 

+1 Minion

6 attacks @ 50.96% avg

High-  91.19%

Low-  8.29%

 

+1 Named Lt

8 attacks @ 25.78% avg

High-  36.78%

Low-  0.27%

 

I ate a Yellow, Orange, and Purple and jumped in, and new almost immediately I was in trouble; this Named Lt was hitting way harder then it should, and it was not missing at all, despite my 25% defense. Now usually I'll cut and run, but after this thread I decided to see what happens.

I have 3 Knockdowns, and 1 Stun in my attack chain, and none of them hit the Lt, including the supposed 100% "when orange" AoE. I ate 2 Greens, and had my Dmg/Heal attack, so I was able to eventually take out the Lt, and then took a final Burst from the Minion, which I was able to kick in the nuts, and get a knockdown on the way to faceplant. All of the Lt's attacks except 1 were at 49.??%; that 1 attack was at 65%, so either it ignored my Def (25 vrs Melee/AoE/Range), or it had a base of 90%.

 

The Named Lt's Attack Sequence was not a quirk with the RNG-Attack sequence, something in the background was driving those numbers.

 

What might have been a quirk with the RNG-Attack sequence was that the Minion missed three in a row,  got a Streakbreaker, and then each subsequent attack went down in percentage; the Minion was able to hit me with it's last 4 attacks. I'm not sure how I feel about maybe being taken out because the MOB got a Streakbreaker.

Note, that the final attacks brought it's average down to the 50%.

 

 

The Named Lt is why people who use Def as protection, build to keep at the softcap, and everybody else is constantly talking about getting to softcap if possible. That Lt should have never happened in a Lev 10 +0 Paper Mission, but it did, and it's not the first time I've faced this situation. That's just one of the quirks of CoH.

 

Now you all go on and preach Statistics, rationalizing CoH's RNG-Attack sequence; I have almost 42 years of professional Training and Experience with Troubleshooting (Military and Civilian), that tells me otherwise.

Edited August 1, 2019 by Jeuraud

[oedipus_tex]

It's really only an issue at low levels and if they ever want to implement a solution, it would be mostly fixed by just giving low level enemies a penalty to their base Defense.

 

The mechanics are the way there are because of powersets like Force Field and the large level of abstraction required to make multiple powersets, each with wildly different effects, all filter into a single combat system. Powers either hit or miss with nothing in between because there are so many secondary effects to consider.

 

It might have been interesting if you had to roll for each tag on the power. (e.g. a power tagged ranged, fire rolled for both ranged and fire, and split the damage and secondary effects). But it would difficult to implement cleanly.

Edited August 1, 2019 by oedipus_tex

[Frostbiter]

There's already a sliding accuracy bonus from level 1-20 called Beginner's Luck. It starts large at +15% toHit and gradually decreases every level. I suspect what Jeuraud ran into was a named Boss scaled down to a Lt.

Edited August 1, 2019 by Frostbiter

[Razia]

On 7/30/2019 at 1:33 AM, Sylvar Panda said:

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. 😄

I had the same happen to me several time: Pop a yellow and miss. Or on my incarnate: Clarion boosted Farsight and my Snipe is missing. At first I thought this is server lag and I have to wait some secs until the server "notices" I popped a yellow.

[jubakumbi]

8 hours ago, Jeuraud said:

Now you all go on and preach Statistics, rationalizing CoH's RNG-Attack sequence; I have almost 42 years of professional Training and Experience with Troubleshooting (Military and Civilian), that tells me otherwise.

35 years of troubleshooting code (Law Enforcement and CIvilian) tells me all you have is some anacdotal evidence.

 

IMO, you have convinced yourself there is something wrong in the code and until you personally take the time to learn code or have a close trusted person do so (not influenced by Big Money because that is Evil it seems) your outlook wiil not change.

 

Your continued attacks on the very types statistical analysis that were likely used to create the very things you mainatained is just very odd, IMO.

 

Obviously, someone somewhere used numbers in a fashion that has made you think statstics are inherently Evil, something about a big boss making bas choices IIRC.

 

You are bascially saying all of the gathered evidence avialabe on how the code performs in this game is irrellevent, all the time gathering them is irrelevent, and that anyone who trie to use statstics has already failed, when, if fact, this is how code is tested.

 

I am preaching nothing, rationalizing nothing - funny you use rationalize as an insult - I am simply presenting the data we have at hand.

Based upon my 35 years of code experience (every level) the data we have suggested there is nothing in the code other than a an RNG that simulates random about as well as other.

 

Now, if you are actually claiming that there is "math is done badly/wrong in the attack sequence code", that is a totally different issue, but is also backup up by the same arguments, because tests (of actual large meaningful datasets) have shown the results to be in the ranges intended.

 

If, like a poster above, you simply think the way the numbers are derived is broken somewhere, OK, but the same evidence we have from testing results shows that if this is the case, the end result still provides the percentages intended by the system.

 

You have stated you would not know how to look at the code, you won't trust anyone else, so you just enjoy your PoV I guess.

Edited August 1, 2019 by jubakumbi
speeling

[GM Capocollo]

8 hours ago, Jeuraud said:

The Named Lt is why people who use Def as protection, build to keep at the softcap, and everybody else is constantly talking about getting to softcap if possible. That Lt should have never happened in a Lev 10 +0 Paper Mission, but it did, and it's not the first time I've faced this situation. That's just one of the quirks of CoH.

 

Now you all go on and preach Statistics, rationalizing CoH's RNG-Attack sequence; I have almost 42 years of professional Training and Experience with Troubleshooting (Military and Civilian), that tells me otherwise.

 

 

You're doing a lot of extrapolating from a sample size of n = 1.  Probabilities aren't guarantees: just because something is rare doesn't mean it's impossible.

 

[daHob]

If anyone really thinks the RNG is skewed, you have all the data you need to demonstrate it. The combat tab shows what you rolled for every attack. Log scrape a sufficient number of data points (a thousand? you'd have to figure out how many is a good number for a random 1-100 range) and run a Chi-Square goodness of fit test. It will give a rating on how close the observed distribution is vs the theoretical distribution (an even chance of every number coming up). 

 

If it comes up really wonky, then you would have something to point at.

 

(link to Chi-Square test description http://www.stat.yale.edu/Courses/1997-98/101/chigf.htm )

[Honorbridge]

11 hours ago, Jeuraud said:

Just a last little tidbit.

 

Lev 10 Wat/Ninj Sentinel

+0 Paper Mission, No Bosses, No AVs

3rd in a row with a +1 Named Lt.

The emphasis on 3 in a row with +1 named LT seems odd.  Most paper missions have a named boss as the target.  With bosses off, it gets downgraded to a Lt.  Since missions can spawn enemies at mission difficulty level or one above, this seems perfectly normal.

11 hours ago, Jeuraud said:

 

Sentinel

10 attacks @ 62.92% avg

High-  93.32%

Low-  18.34%

 

+1 Minion

6 attacks @ 50.96% avg

High-  91.19%

Low-  8.29%

 

+1 Named Lt

8 attacks @ 25.78% avg

High-  36.78%

Low-  0.27%

I’m not sure what your numbers here represent.   It can’t be hit chance, because there is a minimum 5% chance to hit.  The LT couldn’t have had a .27 chance to hit.  Maybe these are the hit rolls?

11 hours ago, Jeuraud said:

I ate a Yellow, Orange, and Purple and jumped in, and new almost immediately I was in trouble; this Named Lt was hitting way harder then it should, and it was not missing at all, despite my 25% defense.

As you mentioned being level 10, were these small inspirations? Because a small purple is only worth 12.5% defense, a medium one would be 25%.  Or did you have enough defensive powers already to have 12.5% defense before the inspiration?  

 

Even in the event your defense was 25%, that still means you had a slightly better than 1 in 3 chance of being hit.  Chaining 4 hits in a row would happen 1 time in 5.

11 hours ago, Jeuraud said:

Now usually I'll cut and run, but after this thread I decided to see what happens.

I have 3 Knockdowns, and 1 Stun in my attack chain, and none of them hit the Lt, including the supposed 100% "when orange" AoE.

Did you mean the attack missed completely or that the status failed to mez?  As I said, I don’t know your sets, so I don’t know how that orange AoE is supposed to work.

11 hours ago, Jeuraud said:

I ate 2 Greens, and had my Dmg/Heal attack, so I was able to eventually take out the Lt, and then took a final Burst from the Minion, which I was able to kick in the nuts, and get a knockdown on the way to faceplant. All of the Lt's attacks except 1 were at 49.??%; that 1 attack was at 65%, so either it ignored my Def (25 vrs Melee/AoE/Range), or it had a base of 90%.

A Lt class mob at +1 has about a 63% chance to hit (50% base critter chance times 1.1 for being +1 times 1.15 for being a Lt.).  Those hit chances are in line with expected results for having used 1 small purple inspiration. (63-12.5=49.5). The inspiration probably wore off just before that last attack.

11 hours ago, Jeuraud said:

The Named Lt's Attack Sequence was not a quirk with the RNG-Attack sequence, something in the background was driving those numbers.

 

What might have been a quirk with the RNG-Attack sequence was that the Minion missed three in a row,  got a Streakbreaker, and then each subsequent attack went down in percentage; the Minion was able to hit me with it's last 4 attacks. I'm not sure how I feel about maybe being taken out because the MOB got a Streakbreaker.

Note, that the final attacks brought it's average down to the 50%.

I’m not sure what you are describing here.  Are you sure you were looking at the hit chance and not the hit roll?  Hit rolls are not percentages, but just numbers.  It doesn’t matter exactly what your roll is, just that it be less than the hit chance to be a hit.

11 hours ago, Jeuraud said:

The Named Lt is why people who use Def as protection, build to keep at the softcap, and everybody else is constantly talking about getting to softcap if possible. That Lt should have never happened in a Lev 10 +0 Paper Mission, but it did, and it's not the first time I've faced this situation. That's just one of the quirks of CoH.

What part of the LT should have not been happening?  Being +1 is perfectly normal, and the fact that it was a downgraded boss means it kept all its powers, just had them weakened.

11 hours ago, Jeuraud said:

Now you all go on and preach Statistics, rationalizing CoH's RNG-Attack sequence; I have almost 42 years of professional Training and Experience with Troubleshooting (Military and Civilian), that tells me otherwise.

The combat system is basically a math problem.  You use math (which statistics is one kind of) to solve math problems.

[Blackbird71]

On 7/30/2019 at 9:56 PM, Jeuraud said:

many of the posters in this thread are saying that because Statistically there is no problem with the CoH Attack-RNG, my feelings of frustration are invalid...

Does double-take

 

If statistically there is no problem with the outcome of the RNG, what other sort of response do you actually expect?!?  Seriously, If the collective data shows anything resembling a true random distribution, what other possible conclusion should we draw other than that there is no problem with the RNG?  Are you really expecting someone to say, "The results all show that everything is working as it should, so we assume there's a serious problem"?  As has been explained before, code is not verified or validated by reading the code, but by testing the outcomes and matching the results to what is expected.  This is exactly what a statistical analysis is to an RNG; it's verifying the accuracy of the outcome.

 

You've got your teeth sunk so far into this that at this point I expect that even if someone did as you asked, and did a full deep-dive of the code and explained it to you line by line how the RNG worked, and it was the purest and most true RNG possible for a computer to simulate, you would still doubt the conclusion.  I don't see any way that you are going to be satisfied with any answer other than hearing that the RNG is hopelessly broken.

 

13 hours ago, Jeuraud said:

Now you all go on and preach Statistics, rationalizing CoH's RNG-Attack sequence; I have almost 42 years of professional Training and Experience with Troubleshooting (Military and Civilian), that tells me otherwise.

All due respect to your experience, it clearly has taught you very little about statistics or computer code.  I have military experience under my belt as well, and one of the things I learned was that one mark of a good officer was knowing the limitations of your own knowledge and recognizing when you needed to yield to the experts.  I'm also an engineer, and I know enough to tell you that in this case, what your training and experience is telling you is wrong.  You've heard the experts speak; are you willing to listen?

 

Let me try one analogy that might explain the disconnect here.  If you were presented with a new type of scope that was supposed to provide accurate aim at extended range, how would you verify that it worked as intended?  Would you disassemble the scope and analyze the lens size, shape, composition, etc.?  Or would you put the scope on a weapon, take it down to the range and test fire it, and compare it to other equipment for accuracy?

 

In this case, what you are asking for (an examination of the code) is the equivalent of disassembling the scope and inspecting its components.  Statistical analysis of RNG results is the equivalent of field testing the scope.  Which do you think gives a more accurate assessment of performance?

Edited August 1, 2019 by Blackbird71

[daHob]

The weirdest part about this whole thing to me is that this game is the most transparent about to hit stuff of any MMO I've ever seen. It shows you the chance to hit and the roll on every attack. Are people assuming they'd show that and then lie about it?

Edited August 2, 2019 by daHob

[jubakumbi]

28 minutes ago, daHob said:

The weirdest part about this whole thing to me is that this game is the most transparent about to hit stuff of any MMO I've ever seen. It shows you the chance to hit and the roll on every attack. Are people assuming they'd show that and then lie about it?

IMO, the assuption is more along the lines of 'If this code does not do what I expect, then it must be broken'.

I personally deal with this on a daily basis from people that do not understand technology...it is a very, very common outlook.

In fact, as I type this, I am listening to my wife on the phone dealing with a client that is determined the code is broken, when they made a mistake.

 

Even though I have dealt with this for so many years, it still gets me everytime.

When the poeple pushing the idea there must be something broken start telling us that we are all part of some kind of conspiracy to hide it, either through big money, lazy, or out of some level of spite/hate to make consumers upset, then it gets really interesting.

[Twisted Toon]

I'm sure someone had mentioned Arcanaville earlier. I'm also 99.9% certain that she has tested the combat code 6 ways from Sunday. She also never worked for, in any way, Cyptic, Paragon Studios, or NCSoft. If she didn't find anything wrong with the code, then I doubt there is anything wrong. If you doubt Arcanaville, when it comes to how things work in the game, then there's no hope for you.

 

I think Acranaville knows more about the game than any other individual, Dev or not.

[Myrmidon]

6 hours ago, Twisted Toon said:

I'm sure someone had mentioned Arcanaville earlier. I'm also 99.9% certain that she has tested the combat code 6 ways from Sunday. She also never worked for, in any way, Cyptic, Paragon Studios, or NCSoft. If she didn't find anything wrong with the code, then I doubt there is anything wrong. If you doubt Arcanaville, when it comes to how things work in the game, then there's no hope for you.

 

I think Acranaville knows more about the game than any other individual, Dev or not.

 

As I recall, she said that they couldn’t afford to pay her.😂

[Twisted Toon]

21 minutes ago, Myrmidon said:

 

As I recall, she said that they couldn’t afford to pay her.😂

I just realized that I said her name 3 times in that last post, and there's no hiding form the Arcanaville... 😢

[arcanaville]

On 8/2/2019 at 5:02 AM, Twisted Toon said:

I'm sure someone had mentioned Arcanaville earlier. I'm also 99.9% certain that she has tested the combat code 6 ways from Sunday. She also never worked for, in any way, Cyptic, Paragon Studios, or NCSoft. If she didn't find anything wrong with the code, then I doubt there is anything wrong. If you doubt Arcanaville, when it comes to how things work in the game, then there's no hope for you.

 

I think Acranaville knows more about the game than any other individual, Dev or not.

Well, I did test both the combat random number generator and the streakbreaker.  I was eventually able to test the combat random number generator to several million swings using combat logs.  In fact, I was able to test it well enough to actually find an actual problem with it, just one no one would ever observe.  The system rounded random rolls to the nearest 0.01%, at least in the combat logs.  If it also did that in actual combat, and I think it did something like that, then there was a teeny tiny problem, the random rolls ranged from 0.00% to 100.00%.  That looks like 10,000 possible rolls, but it is actually 10,001 possible rolls: it includes both zero and 100.00.  Moreover, both zero and 100.00 showed up half as often as any other number, demonstrating this was a fencepost problem not a random generation problem.  So it is kind of possible that a defense champ would be getting hit maybe one in a thousand times more often.  Which human brains really cannot notice.  I was also able to reverse engineer precisely where the streakbreaker cutoffs were from the data, accurate to about 0.2%.  So I'm pretty sure I was testing correctly.

 

I never found a problem in the combat generator that would *noticeably* affect combat.  And I'm pretty sure if there was one, it would have shown up in the data.

 

Incidentally, when the game was still running I had an open policy that if anyone thought there was a bug in the combat mechanics or the combat random rolls they could send me a demorecord to analyze.  I think I analyzed about two hundred of them.  No actual problems in those either.

 

One more correction: I did do a project for Paragon under contract once, near the end.  I helped them restore standard rewards for custom critters used in Architect missions by creating a system that allowed the game to offer those rewards if the custom critters weren't "crippled" by the designer to make them easy to kill.  Basically, you had to be at least as strong as the average normal critter at every level range to generate normal rewards.