I managed to squeeze in a few cycles in between other priorities, and spent some time today on the concept of rebuilding Fire At Will.
Bad News: (?)
Based on a series of tests I just performed, Fire at Will appears to completely inherit Accuracy stats from both the Captain, and any associated items.
I know a few folks here have run parses of combat logs, and performed tests of their own. Care to lend a hand in reproducing the reported Fire at Will issue? Because according to the steps I just took, there is no bug.
Here's how I was testing:
1) Give an NPC a +1000 Defense buff.
2) Fire at Will, and observe the miss-tastic light show.
3) Give myself a +1000 Accuracy buff (power-based, and equipment-based, in separate tests)
4) Fire at Will, and observe nearly every shot land as if neither of the above buffs were in place.
CONCLUSION: Fire at Will is using player and equipment Accuracy boosts when calculating its chance to hit.
So... where does that leave us?
Jeremy Randall
Cryptic - Lead Systems Designer
"Play smart!"
just yesterday i looked at a FAW users log, he was using very inaccurate AP beams and only had ~85 acc with them. the entry's for his FAW had an acc of ~65%. generally most beam users have at least 95% acc in all the logs i parsed, and then i see their FAW and its never above 75%
sooo thats why we make these claims. i'll try to get screens later
just yesterday i looked at a FAW users log, he was using very inaccurate AP beams and only had ~85 acc with them. the entry's for his FAW had an acc of ~65%. generally most beam users have at least 95% acc in all the logs i parsed, and then i see their FAW and its never above 75%
sooo thats why we make these claims. i'll try to get screens later
Perhaps there's some weird splitting of accuracy between the 2 targets per beam when using FAW. Bort, you should test that case just to be sure.
just yesterday i looked at a FAW users log, he was using very inaccurate AP beams and only had ~85 acc with them. the entry's for his FAW had an acc of ~65%. generally most beam users have at least 95% acc in all the logs i parsed, and then i see their FAW and its never above 75%
sooo thats why we make these claims. i'll try to get screens later
This may be a side effect of each pulse hitting two targets?
A regular beam fires 4 times in a cycle, hitting one target between 1 and 4 times.
A fire-at-will firest 5 times in a cycle, hitting up to two targets between 1 and 5 times each.
Is it possible that whatever parsing tool is being used is counting the misses on Fire at Will, but not counting the correct number of potential activations, therefore driving the Hit:Miss ratio off-center?
Jeremy Randall
Cryptic - Lead Systems Designer
"Play smart!"
This may be a side effect of each pulse hitting two targets?
A regular beam fires 4 times in a cycle, hitting one target between 1 and 4 times.
A fire-at-will firest 5 times in a cycle, hitting up to two targets between 1 and 5 times each.
Is it possible that whatever parsing tool is being used is counting the misses on Fire at Will, but not counting the correct number of potential activations, therefore driving the Hit:Miss ratio off-center?
possibly, but i recall seeing about as many 'miss' pop ups as there were damage dealt popups when i used FAW last.
BO3 also seems to have an issue. that issue being that it seems to have the same acc problem as FAW. i can have someone held and it doesn't mater, it will miss 25% of the time or more. wile BO2 has about the same acc of the beam that uses it
That was agianst a player flying along side them at about 14.18 impulse (both ships), target had elusive but the char I used is one of my old ones and dosen't have accurate all I used was 8 beam arrays and the two fire at will.
Phaser did not proc on engines as far as I could tell (was switching between the two clients on the same computer). Only used TSS, Emergency Power to Shields, Hazard Emitters and Tac Team on the target.
Assuming both tests were performed correctly their is only one explanation.
The error is not in equipment bonuses themselves, but in the specific equipment bonus from the weapon being applied to the weapon. The [Acc] mod itself.
Redo the test with a target having 1000 defense. Have two weapons of two energy types, one with a +1000 Acc mod and one without.
This does of course assume that your equipment accuracy bonus was from a console or something similar that applied to all your attacks in the test.
I was under the impression that acc and defense don't "cancel" each other out like that.
Maybe try a test here after you add the 1000 defense you then just remove the 1000 you added?
Why on earth did you even test it in that manner?
If faw ignores acc as some people seem to claim should adding 1000+ of it in fact done nothing if that was the case ?
Seems to me a logical way to test it if your a dev... crank the number so high that it would be very easy to see the results.
Now what it wouldn't show is if FAW is multiplying the effect of ACC some how... say with some type of 0.8 multiplier to its effective values or something... which to be honesty would make logical sense anyway... Why should "FIRE everything" be as accurate as... "Take the shot" lol
here is an in practical log from a 40 minute match of a FAW user. FAW1 and FAW2 both have very much worse acc then just the beam firing normally. against pve targets, everything is almost always 100% acc or close to it, so faw isn't much of a problem there
it appears the point defense console is 100% accurate as well, hmm
i'll jump in my excelsior and stick FAW on it and take some readings
i think that not only does faw ignore your weapon acc mods, but your ships acc bonus as well. actually we were told that it does by design. so im kinda confused by 1 strange test showing him that it does actually work right, somehow
I think that this is an error in Cryptics logging software I used a single beam fire at will (5 hits) and visually saw 5 hits, the combatlog and the in game damage showing thingy showed 8 hits (double of 4 like a normal beam would) (edit: swings (the amount of times the weapon trys to hit) and it hit (amount of times it hit after accuracy applied) 6 times).
1 of those 5 hits missed on my visual confirmation.
In Game Chat Log:
Your Beam Array - Fire at Will I deals 0 Physical Damage to &#($. Your Beam Array - Fire at Will I dealt 545 (390) shield damage to &#($. Your Beam Array - Fire at Will I deals 43 (746) Phaser Damage to &#($. Your Beam Array - Fire at Will I deals 0 Physical Damage to &#($. Your Beam Array - Fire at Will I dealt 486 (348) shield damage to &#($. Your Beam Array - Fire at Will I deals 39 (665) Phaser Damage to &#($. Your Beam Array - Fire at Will I dealt 959 (687) shield damage to &#($. Your Beam Array - Fire at Will I deals 76 (1312) Phaser Damage(Critical) to &#($.
I notice that the reports of FAW being less accurate taken from combat logs taken in the field, not in laboratory conditions like Borticus's test.
I have an alternate theory: When you fire your guns manually, you're taking shots at targets that you think you can hit. But Fire At Will doesn't do that: It just takes shots at any old target, including obviously unhittable ones. So shots taken at targets you can't possibly hit result in automatic misses, which in turn depresses the accuracy stats for it. Take, for instance, an FAW activated when two potential targets are present: A Snoozer you always hit, and an Escort you never hit. Assuming the shots are distributed at random and thus are aimed at both targets, one of which you always hit, and one of which you always miss, FAW will then result in an accuracy of 50%. Shots taken manually result in 100% accuracy. This may explain the discrepancy in accuracy values.
I notice that the reports of FAW being less accurate taken from combat logs taken in the field, not in laboratory conditions like Borticus's test.
I have an alternate theory: When you fire your guns manually, you're taking shots at targets that you think you can hit. But Fire At Will doesn't do that: It just takes shots at any old target, including obviously unhittable ones. So shots taken at targets you can't possibly hit result in automatic misses, which in turn depresses the accuracy stats for it. Take, for instance, an FAW activated when two potential targets are present: A Snoozer you always hit, and an Escort you never hit. Assuming the shots are distributed at random and thus are aimed at both targets, one of which you always hit, and one of which you always miss, FAW will then result in an accuracy of 50%. Shots taken manually result in 100% accuracy. This may explain the discrepancy in accuracy values.
That's a nice theory, but several posters have showed depressed ACC values under near-lab conditions firing at a single target.
On the other hand, Bort's suspicion as to how the combat log and/or parser are handling the multiple attacks sounds very plausible.
Bort, do you have an internal parsing/data mining tool that could circumvent this?
Hilbert's combatlog script for ACT *does* for the record have a deficiency. It will only track the *first* instance of repeating damage, if the subsequent damage instances, within a given time are the same damage.
It's really bad at calculating plasma fire damage, for instance- if a plasma fire deals 114 plasma damage 8 times, it'll only read it dealing that damage once.
Now, obviously with phaser beams you have a different animal- but with this known behaviour suggests that at least such an issue is possible.
So I ask Bort: Can we has Cryptic combat parser please? (Assuming you have one)
I don't think Cryptic has a parser. I would rather hear from Mancom and see what he thinks. Cause there being an issue with the log or the way the parser is working with it does seem like one possible reason we are seeing odd numbers with FAW.
The STO combat log is far from perfect... perhaps something is up withe the way the parser is sorting the data from it.
Oh, and another thing: As I expressed in all my bug report threads, it could also be that FAW is not ignoring gear accuracy, but only using a wrong multiplier, so that a +10% [Acc] counts only for +1% or +0.1%. (We had wrong multipliers on the Plasma-Disruptor Hybrids, so it's not such a wild idea.) If this were true, it would explain why the +1000 Acc gear tests gave perfect accuracy. At this level, it doesn't matter if the multiplier was wrong.
(If it is a wrong multiplier, then it is impossible to determine with certainty for the players without access to dev tools, because a +1% change in accuracy disappears in the background noise of the random number generator and can't be reliably concluded from test data in the limited acc/def range we have at our disposal.)
Barely related but a thread would probably be overkill.
The devs should add phaser beam detonation like we see in balance of terror. It obviously couldn't be used like depth charges as shown (unless a cloaked ship was near your target) but it would be a cool weapon against carriers and difficult to hit escorts. It could do more focused damage than faw but less singular damage than beam overload or something.
well i parsed some random pug matches and got some data. no surprises lol, it had terrible accuracy and misses everything equally. it does not discriminate between spam and player
pug matchs with FAW3 at global thanks to tech doffs, activated as soon as it came off cooldown every time, on a fleet excelsior. it had 28.4 bonus accuracy thank to a vet boff, omega deflector, and accurate trait. test conducted with 8x [Tetryon Beam Array Mk XI [Acc]x2 [Dmg]]
this was not. this was a duel with 1 person, no spam. we were both moving at normal speed the whole time and he said his defense score was 73, quite high. FAW has its worst showing yet here.
so FAW is an ok garbage man skill. sure, it will hit spam most of the time and clear it, but it turns off any illusion of damage dealing wile its on. even in the duel with FAW on i could tell he was taking less damage then he was when my beams were just hitting him unbuffed.
beams already have a large damage deficiency vs cannons, and cannons have CRF that increases their rate of fire and damage dealing significantly on top of it. i can do a ton of damage in pvp with kdf cruisers, they are very effective tactical platforms. fed cruisers with their terrible turn rate and loathsome beam arrays cant even deal effective pressure damage anymore. beams need a rapid fire skill too, or they need to have a FAW like fire rate all the time, and faw can just be random target hit mode.
I don't think there actually is a Cryptic combat parser, but Bort's test seems to be rather dramatic: With enormous values like that, if the modifier were even slightly wrong, you'd get a massive gap between the defense and attack values, and you'd still thus have a misstastic lightshow.
It therefore seems likely that one of two things is occurring:
1. The combat parser in use is not correctly interpreting the data.
2. FAW fires upon bad targets in combat usage, thus resulting in higher levels of misses because those targets are not hittable to begin with.
Since I can't see any obvious flaw in Bort's testing methodology, and yet these two dramatically differing observations exist, there must be some missing link that explains this discrepancy, and one or both of those two seems to cover it.
1. The combat parser in use is not correctly interpreting the data.
2. FAW fires upon bad targets in combat usage, thus resulting in higher levels of misses because those targets are not hittable to begin with.
Empirical evidence from 1v1 private matches suggests that there is more to it though.
I would put forward two other possibilities:
FAW calculates hits in a strange way -- two examples that have been brought up include the possibility that it is using hit calculations for cannons instead of beams, or that it's faster firing rate (and/or other unique properties) is having unintended effects on the algorithm that determines whether an attack hits.
Borticus' test all but ruled the problem being a simple multiplier as such a modification would have been immediately noticeable after adding 1000 defense and ACC, but didn't rule out that the base accuracy isn't somehow lower. For example, if FAW suffers from an ACC penalty of 20, the results will look the same whether you're comparing 20 to 40 or 980 to 1000.
I would like to ask Borticus to try the test a different way. Adding +1,000 Accuracy in a magical dev way vs using FAW in the 'real world' could be causing the difference. Wasn't the main issue that the acc mod on the weapons themselves were not being recognized or that they were being misinterpreted somehow? An overall/general boost to acc that bypasses the weapons may be missing the whole problem, neh? And even if the test focuses on wep mods only, if the problem on the weapon mod is off by only .2 (just an example), boosting acc to +1,000 would kinda drown out and make differences difficult to discern.
I don't claim to be experts like many of you, but it doesn't seem like the test was an apples to apples one.
But I do appreciate a test being done. TY Borticus. but please don't stop here.
This may be a side effect of each pulse hitting two targets?
A regular beam fires 4 times in a cycle, hitting one target between 1 and 4 times.
A fire-at-will firest 5 times in a cycle, hitting up to two targets between 1 and 5 times each.
Is it possible that whatever parsing tool is being used is counting the misses on Fire at Will, but not counting the correct number of potential activations, therefore driving the Hit:Miss ratio off-center?
Given that information you aren't forced to only account for hits and misses. But sure the guys could try. You could also just check to see if FAW is adding as much damage as it should be.
We know what normal phaser fire does. We know how many activations FAW will trigger given a known amount of targets. Check to see if the incremental damage dealt by FAW matches those expectations. If it does, it could just be a reporting issue. If it doesn't, it would be a FAW issue.
If you're having difficulty with a metric, and you think the metric itself may be the issue, try a different one.
The hit/miss is based off of the difference between the two values. So the first test would have had a difference of 1000. Presumably pushing the hit rate to its lowest. The second test would have had a difference of 0, putting the hit rate at its highest. So. Say acc mods of any type were not being accounted for. That was the premise. The second test would have had virtually the same results as the first and Bort wouldn't have needed to do any heavy lifting on the percentages. Issue resolved. If there WAS some deviation from expectations (never missing) but it wasn't catastrophic then we would have a measurement of a partial change in the acc or def mods or even base values.
Good News:
CONCLUSION: Fire at Will is using player and equipment Accuracy boosts when calculating its chance to hit.
So... where does that leave us?
Great test, simple and informative. However, just to cover all the bases, I would repeat the same test vs. 2 NPC's, then 3, then 4, etc. just to look for consistency (just in case BFAW behaves differently vs. 1 or multiple targets). Assuming that checks out, then I would investigate the following:
An error in the reporting of hits/misses
1. parse manually and see if the "accuracy bug" shows up.
- If it does, then the error is in the delivery of hit/miss information by the game.
- If it does not, then the error is in the delivery or interpretation of information by the parsing programs.
I would not proceed to speculate until I knew the above outcomes.
Comments
sooo thats why we make these claims. i'll try to get screens later
R.I.P
This may be a side effect of each pulse hitting two targets?
A regular beam fires 4 times in a cycle, hitting one target between 1 and 4 times.
A fire-at-will firest 5 times in a cycle, hitting up to two targets between 1 and 5 times each.
Is it possible that whatever parsing tool is being used is counting the misses on Fire at Will, but not counting the correct number of potential activations, therefore driving the Hit:Miss ratio off-center?
Cryptic - Lead Systems Designer
"Play smart!"
Like my fanpage!
https://www.facebook.com/CaptainBMoney913
Join Date: August 29th 2010
possibly, but i recall seeing about as many 'miss' pop ups as there were damage dealt popups when i used FAW last.
BO3 also seems to have an issue. that issue being that it seems to have the same acc problem as FAW. i can have someone held and it doesn't mater, it will miss 25% of the time or more. wile BO2 has about the same acc of the beam that uses it
Nooooooo!!!1@
http://i.imgur.com/hmuMPSY.png
Parser I use:
Info:
http://www.reddit.com/r/sto/comments/yoc4e/how_to_act_a_semicomprehensive_guide/
Script:https://raw.github.com/Abydos1/act-sto-plugin/master/sto.cs
That was agianst a player flying along side them at about 14.18 impulse (both ships), target had elusive but the char I used is one of my old ones and dosen't have accurate all I used was 8 beam arrays and the two fire at will.
Phaser did not proc on engines as far as I could tell (was switching between the two clients on the same computer). Only used TSS, Emergency Power to Shields, Hazard Emitters and Tac Team on the target.
I sorta gottta giggle at this response :P
PaxOttomana: gawd mirror event is like fighting a tsi premade, they keep comin and comin!
Maybe try a test here after you add the 1000 defense you then just remove the 1000 you added?
Why on earth did you even test it in that manner?
Do you even Science Bro?
The error is not in equipment bonuses themselves, but in the specific equipment bonus from the weapon being applied to the weapon. The [Acc] mod itself.
Redo the test with a target having 1000 defense. Have two weapons of two energy types, one with a +1000 Acc mod and one without.
This does of course assume that your equipment accuracy bonus was from a console or something similar that applied to all your attacks in the test.
If faw ignores acc as some people seem to claim should adding 1000+ of it in fact done nothing if that was the case ?
Seems to me a logical way to test it if your a dev... crank the number so high that it would be very easy to see the results.
Now what it wouldn't show is if FAW is multiplying the effect of ACC some how... say with some type of 0.8 multiplier to its effective values or something... which to be honesty would make logical sense anyway... Why should "FIRE everything" be as accurate as... "Take the shot" lol
https://i2.minus.com/ihYMMTz19ZjZ6.JPG
it appears the point defense console is 100% accurate as well, hmm
i'll jump in my excelsior and stick FAW on it and take some readings
i think that not only does faw ignore your weapon acc mods, but your ships acc bonus as well. actually we were told that it does by design. so im kinda confused by 1 strange test showing him that it does actually work right, somehow
1 of those 5 hits missed on my visual confirmation.
In Game Chat Log:
Your Beam Array - Fire at Will I deals 0 Physical Damage to &#($. Your Beam Array - Fire at Will I dealt 545 (390) shield damage to &#($. Your Beam Array - Fire at Will I deals 43 (746) Phaser Damage to &#($. Your Beam Array - Fire at Will I deals 0 Physical Damage to &#($. Your Beam Array - Fire at Will I dealt 486 (348) shield damage to &#($. Your Beam Array - Fire at Will I deals 39 (665) Phaser Damage to &#($. Your Beam Array - Fire at Will I dealt 959 (687) shield damage to &#($. Your Beam Array - Fire at Will I deals 76 (1312) Phaser Damage(Critical) to &#($.
Parser:
http://i.imgur.com/2XwY3AR.png
http://files1.guildlaunch.net/guild/library/205090/Untitled.png
[SIGPIC]http://file3.guildlaunch.net/205090/DVhexishensig.jpg[/SIGPIC]
Cruisers with mk x common in infected elite http://www.youtube.com/watch?v=Q82PqoFFxjc
Cruisers with good gear in infected elite http://www.youtube.com/watch?v=WMnFljZD9m8
Soloing Infected Elite http://www.youtube.com/watch?v=XaEFICFx4E8&feature=youtu.be
I have an alternate theory: When you fire your guns manually, you're taking shots at targets that you think you can hit. But Fire At Will doesn't do that: It just takes shots at any old target, including obviously unhittable ones. So shots taken at targets you can't possibly hit result in automatic misses, which in turn depresses the accuracy stats for it. Take, for instance, an FAW activated when two potential targets are present: A Snoozer you always hit, and an Escort you never hit. Assuming the shots are distributed at random and thus are aimed at both targets, one of which you always hit, and one of which you always miss, FAW will then result in an accuracy of 50%. Shots taken manually result in 100% accuracy. This may explain the discrepancy in accuracy values.
That's a nice theory, but several posters have showed depressed ACC values under near-lab conditions firing at a single target.
On the other hand, Bort's suspicion as to how the combat log and/or parser are handling the multiple attacks sounds very plausible.
Bort, do you have an internal parsing/data mining tool that could circumvent this?
It's really bad at calculating plasma fire damage, for instance- if a plasma fire deals 114 plasma damage 8 times, it'll only read it dealing that damage once.
Now, obviously with phaser beams you have a different animal- but with this known behaviour suggests that at least such an issue is possible.
So I ask Bort: Can we has Cryptic combat parser please? (Assuming you have one)
The STO combat log is far from perfect... perhaps something is up withe the way the parser is sorting the data from it.
Food for thought for Borticus: Is FAW currently using the beam or the cannon accuracy formula? (http://hilbertguide.com/blog/#2012-08-18)
Oh, and another thing: As I expressed in all my bug report threads, it could also be that FAW is not ignoring gear accuracy, but only using a wrong multiplier, so that a +10% [Acc] counts only for +1% or +0.1%. (We had wrong multipliers on the Plasma-Disruptor Hybrids, so it's not such a wild idea.) If this were true, it would explain why the +1000 Acc gear tests gave perfect accuracy. At this level, it doesn't matter if the multiplier was wrong.
(If it is a wrong multiplier, then it is impossible to determine with certainty for the players without access to dev tools, because a +1% change in accuracy disappears in the background noise of the random number generator and can't be reliably concluded from test data in the limited acc/def range we have at our disposal.)
The devs should add phaser beam detonation like we see in balance of terror. It obviously couldn't be used like depth charges as shown (unless a cloaked ship was near your target) but it would be a cool weapon against carriers and difficult to hit escorts. It could do more focused damage than faw but less singular damage than beam overload or something.
Good luck with faw and thanks for the updates.
https://youtube.com/channel/UCJZ5FBJ9bFaZ6yAFiNpZiRQ/featured?view_as=subscriber
Twitter:
https://twitter.com/CaptainCidStorm
pug matchs with FAW3 at global thanks to tech doffs, activated as soon as it came off cooldown every time, on a fleet excelsior. it had 28.4 bonus accuracy thank to a vet boff, omega deflector, and accurate trait. test conducted with 8x [Tetryon Beam Array Mk XI [Acc]x2 [Dmg]]
https://i.minus.com/i1odg7MuTJzmX.JPG
https://i.minus.com/ibvaBSw5HTDemE.JPG
https://i.minus.com/i0mIRLaxlvNbo.JPG
these were all hectic matches full of targets at variable speed and plenty of spam with high evasion.
https://i2.minus.com/idS4Pa8eLATd1.JPG
this was not. this was a duel with 1 person, no spam. we were both moving at normal speed the whole time and he said his defense score was 73, quite high. FAW has its worst showing yet here.
so FAW is an ok garbage man skill. sure, it will hit spam most of the time and clear it, but it turns off any illusion of damage dealing wile its on. even in the duel with FAW on i could tell he was taking less damage then he was when my beams were just hitting him unbuffed.
beams already have a large damage deficiency vs cannons, and cannons have CRF that increases their rate of fire and damage dealing significantly on top of it. i can do a ton of damage in pvp with kdf cruisers, they are very effective tactical platforms. fed cruisers with their terrible turn rate and loathsome beam arrays cant even deal effective pressure damage anymore. beams need a rapid fire skill too, or they need to have a FAW like fire rate all the time, and faw can just be random target hit mode.
It therefore seems likely that one of two things is occurring:
1. The combat parser in use is not correctly interpreting the data.
2. FAW fires upon bad targets in combat usage, thus resulting in higher levels of misses because those targets are not hittable to begin with.
Since I can't see any obvious flaw in Bort's testing methodology, and yet these two dramatically differing observations exist, there must be some missing link that explains this discrepancy, and one or both of those two seems to cover it.
Empirical evidence from 1v1 private matches suggests that there is more to it though.
I would put forward two other possibilities:
I don't claim to be experts like many of you, but it doesn't seem like the test was an apples to apples one.
But I do appreciate a test being done. TY Borticus. but please don't stop here.
Given that information you aren't forced to only account for hits and misses. But sure the guys could try. You could also just check to see if FAW is adding as much damage as it should be.
We know what normal phaser fire does. We know how many activations FAW will trigger given a known amount of targets. Check to see if the incremental damage dealt by FAW matches those expectations. If it does, it could just be a reporting issue. If it doesn't, it would be a FAW issue.
If you're having difficulty with a metric, and you think the metric itself may be the issue, try a different one.
Maybe there's something being introduced when there's more than one. Like it starts using the wrong formula for later targets.
CommanderDonatra@Capt.Sisko: ahhh is it supposed to do that?
Norvo Tigan@dontdrunkimshoot: hell ya, maybe
Great test, simple and informative. However, just to cover all the bases, I would repeat the same test vs. 2 NPC's, then 3, then 4, etc. just to look for consistency (just in case BFAW behaves differently vs. 1 or multiple targets). Assuming that checks out, then I would investigate the following:
An error in the reporting of hits/misses
1. parse manually and see if the "accuracy bug" shows up.
- If it does, then the error is in the delivery of hit/miss information by the game.
- If it does not, then the error is in the delivery or interpretation of information by the parsing programs.
I would not proceed to speculate until I knew the above outcomes.