Accuracy balancing

[msfmatmoo]

I feel like attacks are missing a lot more than they should no matter what type of character I use. I get that accuracy isn't perfect, but if five attacks in a row miss with an accuracy buff, I think that that's a bit much. I feel like accuracy may need to be balanced in the game.

[Psyonico]

what kind of accuracy buff do you have?

what level difference enemy are you fighting?

What type of enemy are you fighting?

 

These and other things make a big difference.  At level 22 you can slot Single Origin Enhancements, one Accuracy SO means you have a 95% chance for most* attacks to hit, which is the cap.

 

*I say most because there are powers like Controller AoE Immobilizes that have an accuracy penalty and require additional slotting to get to the 95% number.  It also is based on the type of enemy you are fighting.

[Greycat]

Yeah. There are a lot of factors that go into "I missed." Including just plain bad luck (thus, streakbreakers.) Kind of like "I popped a yellow and immediately missed" - you remember it because it was annoying, not because there was anything wrong.

 

Plus, if I'm recalling correctly, from ... 1-10?  you SHOULD  have an acc (or is it tohit?) bonus. But again - a lot of factors go into "I missed."

[ShardWarrior]

19 minutes ago, Psyonico said:

what kind of accuracy buff do you have?

what level difference enemy are you fighting?

What type of enemy are you fighting?

I would add what level are you and what enhancements are slotted?

 

Fighting CoT pre SO level can be painful.

[Vanden]

51 minutes ago, Psyonico said:

At level 22 you can slot Single Origin Enhancements, one Accuracy SO means you have a 95% chance for most* attacks to hit, which is the cap.

This is a worringly widespread belief. One even-level Acc SO is enough for a 95% chance to hit on even-level targets. To hit a +1 enemy 95% of the time, you need 46.2% enhancement. For +2, 69.6%. +3 needs 98%, and a +4 needs a whopping 143.6% enhancement to hit that 95% cap.

Edited May 8, 2020 by Vanden

[tidge]

If you want to entertain yourself, start a new tab just to monitor combat (specifically your ToHit rolls).

 

Read this page, especially the part about "The Streak Breaker": https://paragonwiki.com/wiki/Attack_Mechanics

 

This is partially user bias: It certainly feels like many more of my attacks are hitting thanks to STREAKBREAKER than should. I actually did some log captures to check, and I was humbled by how difficult it would be for me to prove (with any significance) that it was firing too often. Any drop in 'final to hit chance' can allow for more misses before streakbreaker helps you. I remain uncomfortable in my ignorance.

 

 

[Psyonico]

23 minutes ago, Vanden said:

This is a worringly widespread belief. One even-level Acc SO is enough for a 95% chance to hit on even-level targets. To hit a +1 enemy 95% of the time, you need 46.2% enhancement. For +2, 69.6%. +3 needs 98%, and a +4 needs a whopping 143.6% enhancement to hit that 95% cap.

hence why I said what is the level difference 😉

[SwitchFade]

1 hour ago, msfmatmoo said:

I feel like attacks are missing a lot more than they should no matter what type of character I use. I get that accuracy isn't perfect, but if five attacks in a row miss with an accuracy buff, I think that that's a bit much. I feel like accuracy may need to be balanced in the game.

The nature of event occurence under largely random variables dictates that statistical representation will be generally uniform about the median in a bell shape.

 

To validate this, please conduct an experiment where in all variables but one are controlled, in this case your attempts to attack and record a sample size of 30-50. Plot those into a single factor anova regression analysis.

 

You will discover that the r-square value will be high, the confidence high, the p value quite low and that accuracy is exactly where is should be based on the forumlae in the code.

 

Or, trust the multiple tests conducted and chalk your one occurence up to random variation on a small sample size and perceptual bias.

[Solvernia]

3 minutes ago, SwitchFade said:

The nature of event occurence under largely random variables dictates that statistical representation will be generally uniform about the median in a bell shape.

 

To validate this, please conduct an experiment where in all variables but one are controlled, in this case your attempts to attack and record a sample size of 30-50. Plot those into a single factor anova regression analysis.

 

You will discover that the r-square value will be high, the confidence high, the p value quite low and that accuracy is exactly where is should be based on the forumlae in the code.

 

Or, trust the multiple tests conducted and chalk your one occurence up to random variation on a small sample size and perceptual bias.

i too know big words

[SwitchFade]

1 minute ago, Solvernia said:

i too know big words

Awesome!

 

If you do run the regression analysis, I'd love to see the data report and the plotline. If you don't mind, could you post it for us?

 

If there is indeed broken code, this would help is find it!

[Solvernia]

1 minute ago, SwitchFade said:

more big words

haha punch go whoosh

[SwitchFade]

1 hour ago, Solvernia said:

haha punch go whoosh

Interesting and questionable reaction when presented with feedback, advice and a willingness to consider your data.

 

Sans proof, conclusion: accuracy is fine and perceptual bias is confirmed.

[MTeague]

Also some mobs have extremely high defense beyond the normal

• Devouring Earth Boulder / Granite types - have very good smashing/lethal resistance defense.
• Carnies - have SUPER irritatingly high psionic resistance defense (speaking as a Mind/Kin, I took Nerve alpha JUST for them.)
• Paragon Protectors:   Love to pop Moment of Glory, and when they do, just stop attacking. Seriously. Save your endurance for the next minute, or go beat up some other critter.
• CoT:  love to hit you with Dark attacks that stack ToHit debuffs on you.  Even a group of 3 Spectral Daemon minions can savage your to-hit.
• Tsoo Sorcys:  all bets are off once they get close to you.  Hurricane is a Brutal ToHit debuff, AND they have a Darkness PBAE that does something too.

I'm sure there's plenty others that can be added to the list. 

Before i switch to sets, I use TWO accuracy enhancements in every power that can take them, even at SO / IO levels. 

Missing can be very paintful.  More damage means nothing if I whiff.

This is especially true if i'm on a Controller or on a support class that needs to HIT in order to heal anyone.  (did I mention I main a Mind/Kin?  coming and going....)

 

But I haven't noticed anything where i'm missing excessively unless a) I'm still using Training enhancers, OR, b) if one of the above foes or one like them is ruining my day.

 

And one can always team up with other characters who bring Defense Debuffs or Acc Buffs to the table.  Controls help a lot too.  A Bunch of CoT ghosts cant debuff your to hit if they're slept, held, confused, or stunned.

 

 

 

Edited May 8, 2020 by MTeague

[macskull]

3 hours ago, msfmatmoo said:

I feel like attacks are missing a lot more than they should no matter what type of character I use. I get that accuracy isn't perfect, but if five attacks in a row miss with an accuracy buff, I think that that's a bit much. I feel like accuracy may need to be balanced in the game.

When players bring up issues like this my reply is always "combat logs or it didn't happen" and one of four situations results:

1. Player is unable to produce said logs, leaving the question up to a matter of confirmation bias since there's no data to back up the claims
2. Player produces said logs and a review of them indicates the game mechanics are functioning as they should, though from a small data set the numbers may seem wrong
3. Player produces said logs and they appear to show a situation that should be impossible but I am unable to replicate
4. Player produces said logs and after review and testing there's something actually wrong

Numbers one and two are far and away the most common in these cases, number three happens sometimes, and I can count on one hand the number of times number four has happened. Since this game's combat is basically dice roll-based, there's always a bit of statistics and probability going on at work here. "I am missing all the time with a 95% chance to hit!" is one I hear all the time, but the more attacks you fire, the more likely you are to miss any one of them. As an example, if I fire ten attacks with the aforementioned 95% chance to hit the odds of all ten attacks landing is only about 60%. If I fire one hundred attacks with a 95% chance to hit those odds drop to 0.6%. Yes, you're going to notice the misses because missing with a 95% chance to hit is frustrating and you don't like it but you're forgetting about the 95% of attacks that do hit.

 

The best way to prove something is or isn't working is to test it yourself but you have to ensure you have a large enough sample size. A few months back I had a discussion with a user on the Homecoming Discord server who was pretty sure a proc was not firing as often as it should, so I set out to get some numbers and figure it out one way or the other. After a few hundred test attacks, I was surprised to see that the power was actually firing off substantially more than my math said it would, so... I kept attacking more, and wouldn't you know it, after roughly 3000 casts the number of times the proc fired was almost exactly the number I'd have expected. Another thing I noted over those 3000 attacks was that the total number of attacks that actually hit was 95.02%, almost exactly the expected 95%.

[Chance Jackson]

One of the 1st things I do on all my toons is set up the 10 combat attributes I'm gonna monitor & I always include "last hit chance" I reccomend that you monitor it too

[agentx5]

Can we PLEASE get rid of the stupid 95% clamp function once and for all.

 

We have Defense%, Resistance%, Accuracy, and toHit% ; we didn't need to guarantee a 5% chance to miss for everybody.

Those layers work just fine, just fine-tune if needed.

 

It's quite frustrating as a well placed setup for an Assassinate on a Stalker and then your whole combo is ruined because the target moved.

Or a blaster preparing a sniper shot that goes wide unable to hit the broad side of the barn.

It's not "fun" it's just annoying and leaves the player feeling cheated.  And it can also ruin a team strategy that was planned.

Annoying things that aren't fun should be removed for Quality of Life reasons, IMHO.

 

You know you've screwed up as a designer (again, I'm talking Cryptic long ago, not current HomeComing Devs) when you have to have a "streak-breaker" code to force a hit to succeed!

 

I see this as an opportunity for HomeComing's team to shine and fix something Cryptic never should have created.  You can still leave a clamp function in at 100% to give a guarantee nobody will ever exceed that in error, but a 5% guaranteed chance to miss no matter what is just stupid.

Edited May 12, 2020 by agentx5

[macskull]

48 minutes ago, agentx5 said:

Can we PLEASE get rid of the stupid 95% clamp function once and for all.

 

We have Defense%, Resistance%, Accuracy, and toHit% ; we didn't need to guarantee a 5% chance to miss for everybody.

Those layers work just fine, just fine-tune if needed.

 

It's quite frustrating as a well placed setup for an Assassinate on a Stalker and then your whole combo is ruined because the target moved.

Or a blaster preparing a sniper shot that goes wide unable to hit the broad side of the barn.

It's not "fun" it's just annoying and leaves the player feeling cheated.  And it can also ruin a team strategy that was planned.

Annoying things that aren't fun should be removed for Quality of Life reasons, IMHO.

 

You know you've screwed up as a designer (again, I'm talking Cryptic long ago, not current HomeComing Devs) when you have to have a "streak-breaker" code to force a hit to succeed!

 

I see this as an opportunity for HomeComing's team to shine and fix something Cryptic never should have created.  You can still leave a clamp function in at 100% to give a guarantee nobody will ever exceed that in error, but a 5% guaranteed chance to miss no matter what is just stupid.

No thanks

[Ukase]

In all candor, as surprising as it may seem to those who are familiar with me - I often imagine myself as the character using the attack. 

Just anecdotal - but I find it interesting that:
Nemesis Staff misses more often than not when in melee compared to being at range. 
Missing a stationary target with a melee attack ...well, it happens...but I think it shouldn't. Hitting a box with a sword with no npcs around to distract you should be a given. And yet, I whiff from time to time. 
Different attacks have different base accuracy...and a lot of them just seem like they were plucked out of thin air. I'd love to know the "why" behind these chosen values. 
Incarnate Judgement - Pyronic has a greater chance to miss than Ionic.  Huh? Why? Cuz fire's hotter than electricity? Is it hotter? Probably not. These are the kinds of thoughts that kept me out of the better schools. 

What really baffles me is why the folks at Cryptic didn't use endurance as the accuracy check. If you're rested - you're far more likely to hit what you aim for - but when you're tired, you're more likely to miss. And yet, the game mechanics don't care about this at all. Probably just as well. I wouldn't hit a thing until I got Agility to t-3, lol. 

 

[Sovera]

On 5/8/2020 at 7:18 PM, Vanden said:

This is a worringly widespread belief. One even-level Acc SO is enough for a 95% chance to hit on even-level targets. To hit a +1 enemy 95% of the time, you need 46.2% enhancement. For +2, 69.6%. +3 needs 98%, and a +4 needs a whopping 143.6% enhancement to hit that 95% cap.

It's a 'worryingly' widespread thing because people tend to not budge their difficulty slider. It's super common even at level 40+ to see people running TFs at +0

 

11 hours ago, agentx5 said:

Can we PLEASE get rid of the stupid 95% clamp function once and for all.

 

We have Defense%, Resistance%, Accuracy, and toHit% ; we didn't need to guarantee a 5% chance to miss for everybody.

Those layers work just fine, just fine-tune if needed.

 

It's quite frustrating as a well placed setup for an Assassinate on a Stalker and then your whole combo is ruined because the target moved.

Or a blaster preparing a sniper shot that goes wide unable to hit the broad side of the barn.

It's not "fun" it's just annoying and leaves the player feeling cheated.  And it can also ruin a team strategy that was planned.

Annoying things that aren't fun should be removed for Quality of Life reasons, IMHO.

 

You know you've screwed up as a designer (again, I'm talking Cryptic long ago, not current HomeComing Devs) when you have to have a "streak-breaker" code to force a hit to succeed!

 

I see this as an opportunity for HomeComing's team to shine and fix something Cryptic never should have created.  You can still leave a clamp function in at 100% to give a guarantee nobody will ever exceed that in error, but a 5% guaranteed chance to miss no matter what is just stupid.

+1. Get rid of the 5%. If it's a worry for NPCs then let them keep it. If it's a worry for PvP then PvP can keep it.

 

In the 'quite frustrating scenario is lenghty animations that miss, with a special place in hell when the animation SHOWS that it will miss AKA me using something like Flaming Arrow and the character is nose to the target (sometimes a box I need to break), AND THEY DO THE SLOW ANIMATION TO THE SKY BEFORE SHOOTING THE CEILING!

[Blastit]

I think the 5% minimum 95% maximum hit chance is good for keeping people humble. Some variance in gameplay is good because you then have to possibly readjust your tactics.

[AerialAssault]

You notice the misses on 95% chance more because they're rarer. How many times have you used Assassin's Strike or Total Focus and it has hit? I guarantee it is more times than you've missed!

 

The 5% miss chance is the equivalent of rolling a 1 in D&D, which is always a miss, regardless of whether your attack bonus is +9 or +99. Without that element of failure, the challenge of the game is reduced significantly. Is it annoying? Yes. Is it bad design? No.

[Greycat]

Oh, if I didn't mention it... I disagree with removing that 5% - with very few exceptions. Why am I missing fires with extinguishers in Steel Canyon? Why am I missing boxes? I don't *think* we tend to miss objects in Mayhem missions. Those.. should really probably be just hittable.

[Doomguide2005]

On 5/8/2020 at 1:36 PM, Psyonico said:

what kind of accuracy buff do you have?

what level difference enemy are you fighting?

What type of enemy are you fighting?

 

<more stuff>

I'd add 

 

What is your To Hit buff?

 

And then I'd do as @tidge did and point them to the Attack Mechanics page of the wiki.  There is lot of not understanding Accuracy vs To Hit (as well as 'English' usage vs game terminology) involved.

[Naraka]

On 5/8/2020 at 4:43 PM, SwitchFade said:

Interesting and questionable reaction when presented with feedback, advice and a willingness to consider your data.

 

Sans proof, conclusion: accuracy is fine and perceptual bias is confirmed.

You do realize you aren't replying to the OP, right?

[SwitchFade]

6 minutes ago, Naraka said:

You do realize you aren't replying to the OP, right?

Obviously. I didn't mention the original poster, original post or any names.

 

Fairly clearly, I quoted someone and directed the comment to their response to feedback and advice.

 

Further, I asked that specific person to post data we could analyze. Notice, by looking carefully at the conversation progression, that that specific person posted snarky comments FIRST, to which I requested they run a regression analysis and post, since they decided to chime in. THEN, I replied after they posted another snide comment. I concluded with it, as it was evident they were uninterested in feedback. The OP never weighed in on it.

 

I'm sure you realize that though, as you were pointed in making sure you knew that I knew. Which is quite clear I do, if one were to follow the chain of conversation. 😌