Here's a log of back-to-back flee stuns in NToV on Thursday: [Thu Nov 2 23:38:56 2017] You tell your raid, '3 more inc' [Thu Nov 2 23:38:57 2017] a glimmer drake begins to cast a spell. [Thu Nov 2 23:38:58 2017] An elder onyx drake hits YOU for 280 points of damage. [Thu Nov 2 23:38:58 2017] An elder onyx drake hits YOU for 100 points of damage. [Thu Nov 2 23:38:58 2017] You are stunned! [Thu Nov 2 23:38:58 2017] An elder onyx drake bashes YOU for 101 points of damage. [Thu Nov 2 23:38:59 2017] An elder onyx drake hits YOU for 110 points of damage. [Thu Nov 2 23:38:59 2017] An elder onyx drake hits YOU for 190 points of damage. [Thu Nov 2 23:39:00 2017] You are unstunned. [Thu Nov 2 23:39:01 2017] You are stunned! [Thu Nov 2 23:39:01 2017] An elder onyx drake hits YOU for 200 points of damage. [Thu Nov 2 23:39:01 2017] An elder onyx drake tries to hit YOU, but misses! [Thu Nov 2 23:39:02 2017] You feel the pain of a million stings. [Thu Nov 2 23:39:04 2017] You are unstunned. [Thu Nov 2 23:39:02 2017] An elder onyx drake tries to hit YOU, but misses! [Thu Nov 2 23:39:02 2017] An elder onyx drake hits YOU for 250 points of damage. [Thu Nov 2 23:39:04 2017] An elder onyx drake hits YOU for 190 points of damage. [Thu Nov 2 23:39:04 2017] You are unstunned. [Thu Nov 2 23:39:04 2017] You begin casting Harmshield. (This is from my ogre SK who was pulling at the time, so they're definitely flee stuns.) Two potentially interesting points: The 'stunned!' and 'unstunned.' messages in the first flee stun are 2 seconds apart. The second 'stunned!' is the very next line in the log file after the first 'unstunned.' That seems to suggest that flee stuns are chaining atm, ya?
Ergh, I messed up copypasta of the log. I trimmed a couple of extraneous lines out to make it more readable, but left the 23:39:04 unstun in two places for some reason. The first one should be deleted. Here is the full, raw log: [Thu Nov 2 23:38:56 2017] You tell your raid, '3 more inc' [Thu Nov 2 23:38:57 2017] a glimmer drake begins to cast a spell. [Thu Nov 2 23:38:58 2017] An elder onyx drake hits YOU for 280 points of damage. [Thu Nov 2 23:38:58 2017] an elder onyx drake was hit by non-melee for 15 points of damage. [Thu Nov 2 23:38:58 2017] An elder onyx drake hits YOU for 100 points of damage. [Thu Nov 2 23:38:58 2017] an elder onyx drake was hit by non-melee for 15 points of damage. [Thu Nov 2 23:38:58 2017] You writhe in the grip of agony. You have taken 1021 points of damage. [Thu Nov 2 23:38:58 2017] You are stunned! [Thu Nov 2 23:38:58 2017] An elder onyx drake bashes YOU for 101 points of damage. [Thu Nov 2 23:38:58 2017] an elder onyx drake was hit by non-melee for 15 points of damage. [Thu Nov 2 23:38:59 2017] Posh tells the raid, 'tells for Ancient Wurm Hide Robe - last call' [Thu Nov 2 23:38:59 2017] An elder onyx drake hits YOU for 110 points of damage. [Thu Nov 2 23:38:59 2017] an elder onyx drake was hit by non-melee for 15 points of damage. [Thu Nov 2 23:38:59 2017] An elder onyx drake hits YOU for 190 points of damage. [Thu Nov 2 23:38:59 2017] an elder onyx drake was hit by non-melee for 15 points of damage. [Thu Nov 2 23:39:00 2017] Your body is covered in a fine mist of slime. [Thu Nov 2 23:39:00 2017] You are unstunned. [Thu Nov 2 23:39:01 2017] You are stunned! [Thu Nov 2 23:39:01 2017] An elder onyx drake hits YOU for 200 points of damage. [Thu Nov 2 23:39:01 2017] an elder onyx drake was hit by non-melee for 15 points of damage. [Thu Nov 2 23:39:01 2017] An elder onyx drake tries to hit YOU, but misses! [Thu Nov 2 23:39:02 2017] You feel the pain of a million stings. [Thu Nov 2 23:39:02 2017] An elder onyx drake tries to hit YOU, but misses! [Thu Nov 2 23:39:02 2017] An elder onyx drake hits YOU for 250 points of damage. [Thu Nov 2 23:39:02 2017] an elder onyx drake was hit by non-melee for 15 points of damage. [Thu Nov 2 23:39:03 2017] You can't cast spells while stunned! [Thu Nov 2 23:39:03 2017] You can't cast spells while stunned! [Thu Nov 2 23:39:03 2017] You can't cast spells while stunned! [Thu Nov 2 23:39:03 2017] The nifiliks stop biting you. [Thu Nov 2 23:39:03 2017] You can't cast spells while stunned! [Thu Nov 2 23:39:04 2017] You can't cast spells while stunned! [Thu Nov 2 23:39:04 2017] An elder onyx drake hits YOU for 190 points of damage. [Thu Nov 2 23:39:04 2017] an elder onyx drake was hit by non-melee for 15 points of damage. [Thu Nov 2 23:39:04 2017] You can't cast spells while stunned! [Thu Nov 2 23:39:04 2017] You can't cast spells while stunned! [Thu Nov 2 23:39:04 2017] You are unstunned. [Thu Nov 2 23:39:04 2017] You begin casting Harmshield.
The first stun is a bash stun, not a flee stun. That's why it's 2 seconds instead of 3. Either way, though, I don't think the idea is that you shouldn't be able to get stunned by the very next hit that lands after your first stun wears off. Just that you shouldn't be able to get chain stunned so as to go more than 3 seconds without seeing an "unstunned" message, however brief.
Yes, stuns aren't supposed to land while you're stunned-- there should be at least one hit between every stun. If this is ever not the case, please show me a log. I think the idea is exactly that you can't get stunned by the next hit. I kited a ton on AK - I wasn't perfect I got hit all the time I never recall getting stunned to stunned and thats what happening now. The new stun is overriding the old stun forcing an 'unstun' message is what it seems like to me.
ill look through my logs, but i know the mechanic is correct by sounds of it but something feels wrong about how it has been implemented currently, i also kited a lot on AK and i do recall being stunned running, but dont recall ever being chain stunned in place for 6-9 seconds.
We can see from Torven's logs that there isn't any kind of "cool down" - that is, you can definitely get stunned, get unstunned, then get re-stunned within the same second. I suppose it's possible that the percentage chance of a given hit to stun is too high? On the other hand, keep in mind that we've all now been playing for years on a server on which the flee stun percentage chance is 0%, which is going to color our perceptions.
It feels like the flee stun is landing while im stunned before I start moving and if i read the description correctly it should only apply as I am moving forward and the mob is behind me.
May need to refine the methods for detecting moving away. That was always the holdup in implementing this before.
This is Live: https://drive.google.com/open?id=1xh_9HEsWMLHuIiEpYj7sBynksAiF0EIu This is TAKP: https://drive.google.com/open?id=1JtoIK7PS3r8_-RLOGQo3JP4_zgSV6w6b For Live, I used a level 69 SK with 6 defense skill so that the green cons would hit almost every swing, which is useful for this test. Lower level NPCs attack slower and double attack less however. The TAKP example had roughly 13-14 mobs.
+1 to flee stun not working properly here. I do remember flee stuns happening on AK, but always assumed they were from a bash or something because of how infrequent they were. If I had to posit a guess, it's that the 10% chance implemented here is way too high. May I ask how this number was derived? Pulling more than 3 mobs is incredibly difficult you don't have a speed buff. It's also worth considering that NPCs have a very long "reach" here. If you are running directly away from a mob, they seem to be able to occasionally whack you from way out of melee range. If they're lucky, they stun you, catch up, and stick on you while continuing to stun. This is no good with a large pull, even if you have selos speed.
I assume the percentage comes from analysis of the log data - divide number of hits landed on an unstunned fleeing player by the number of stuns? On the other hand, I suppose there are a lot of ways this number could end up being distorted, since the presence of those 3 second stuns is likely the only indication in a log that a player is running away. I always figured this had something to do with a bit of lag rather than melee range, per se, but I do think it's true that you're more likely to get a "surprise hit" from a mob that seems like it should be out of melee range here on TAKP than you were on Al'Kabor. One relevant factor is of course the fact that mobs on TAKP (unlike on AK) can melee through walls, which can get really ridiculous in certain places. I suppose this probably makes pulling in KC and the Hole a bit harder than it strictly should be with flee stuns.
10% was derived from logs I made of a character loaded with heavy items to make them walk slowly so I could heal them, not run into stuff, prevent from outdistancing and keep the NPC behind me after choosing NPCs that didn't kick or bash. I used regex to remove text in between 'you are stunned! you are no longer stunned' to eliminate hits on me while I was stunned, then did simple line counts to determine hit and stun counts. If anybody wants to independently verify the results, use this regex in notepad++: (?-s)^.*You are stunned(?s).*?(?-s)you are no longer stunned.*(\R|\z) I tested a jopal hunter in PoFire which is a level 63 rogue NPC on my level 64 ranger. The results were 74 stuns in 704 hits. I also tested a level 5 puma, which is low enough level to not bash/kick. Result was 6 stuns from 60 hits on the same level 64 ranger. I suppose I could do a test with SoWed NPCs while at normal run speed to be certain that run speed doesn't affect stun rate. I would say that NPCs on TAKP do probably run (or at least walk) a little faster than they should, or it sure looks that way to me. If that is true, then it could be contributing to issues with stuns. Also certainly the server does timer checks more frequently than Sony's servers.
Mob speed is very close. I have worked hard on getting it right. Typical mobs will run you down without sow. Strafe and you can just outrun them. The fact remains, that on AK people pulling inside had run3. And outside always had sow. You made sure you didn't get hit. Should mobs that bash, stun with hits too? Or are they mutually exclusive? Is it really a mechanic to make up for non-bashing mobs?
My other thought, is the stun could be an offhand attack, has a chance to stun from behind? Maybe charm a mob, give it two weapons and see if there is a correlation to which hand causes the stun.
Would that affect encumbrance and therefor hits/stuns? I don't know -- just wondering. If another parse gets done, maybe an ensnare instead of being overweight?
Just did another log with SoWed and hasted Dervish Cutthroats with an unecumbered running character. 136 stuns in 1089 hits, which is 12.5%. I just spoke with Demonstar who had also come up with 12% when he implemented it into EQEmu, so if anything I had it too low.
Another thought. Perhaps the chance on AK was 10% per round, and here it's coded as 10% per hit, so when mobs double, there's a 20% chance of the stun landing per round.
We just did some Vyemm fighting, and this Stun runing back wards prevents the warrior to return to the position after he gets summoned after a knockback i think the way we are determine wether you are running from a mob or not
Mm, interesting. I agree that shouldn't be happening. On Al'Kabor, we experimented with summon-kiting against the Rathe Council before Hobart patched in the fix that reduced their damage as their health went down. We found that non-ogre tanks eventually got stunned and died, but that ogres never did. If 3-second flee stuns were landing, this wouldn't have worked. So flee stuns should likely only be triggered by blows that land from behind the moving toon.
... and three seconds after posting the above, I realized that there's no reason an ogre's frontal stun immunity wouldn't apply to a flee stun just as much as a bash stun. Nevermind!
Doesnt matter if you are Ogre or not, when you get summoned, you start by default with your back turned and get rekt.
12 years on live, 5 years on AK, pulling regularly for AOEs, raids and low level AoEs for tradeskilling. I've pulled countless hours of KC, Seb and the hole AoE on live and AK. I have never in my life be stunned as much as i have in the past week. I got stunned 8 times running from the first 3 jugs back to the Shroom king. Happened on 2 pulls. The logs are the logs, but something doesn't feel right here.
The Vyemm scenario seems to be flee stuns when the tank isn't moving is the issue. It's instant summon/stun. For flee stun, you have to be moving away from the mob but the way it's currently working is a 10% chance to stun stationary PC's, seemingly much higher than 10%, but that's a different debate.
I see the problem with Vyemm/summoning. I'm going to reduce the stun chance to 5% temporarily until we have a better solution for checking forward movement. 12% stun chance when moving however is what Live is currently like. Either Sony increased this chance at some point (not an impossibility), or you all have bad memories; and Sony was making the game easier over the years, not harder. I'm not sure how I'm going to resolve this in a way that satisfies everybody and likely won't be able to without an online Al'Kabor to run tests on. Contrary to popular belief however, my goal is NOT to make a harder than AK server.
The problem Torven is that currently however the flee stun is implemented feels too simple. Like that 12% demon parse is lacking some variable that doesn't parse well into a log. Currently its like flee stun from behind moving 10%. Something is missing. Play the server, then go play live - do the same pulls (I did this during foragees) - you will see that on TAKP the stuns are landing more frequently. I dont know why. Something is missing. Some data isn't in play. Play takp and tell me if this is YOUR experience when you pulled even 1-2 mobs.
Flee stuns were a bit jarring at first, but I'm getting used to 'em. Run3 will make them way less common, I think. I'm all for the change, both for AKuracy and balance. Out of curiosity, Torven, did you calculate confidence intervals for the Live and AK stun chance? If not, I could probably whip 'em up if you point me toward the data. (Or is it all upthread?) Out of other curiosity, Haynar, how does the server currently compute a player's velocity? Does it continually maintain speed and heading as states in a client object, and query 'em when needed? Or just keep track of recent position history and estimate velocity (when needed) from that? Detecting strafing and spell- or melee-based push seem like hard problems!
