PSA: Reasonable probability that charm has been nerfed

[Splorf22]

New: Loraen (230 cha on average) vs L50 average mobs: 128 seconds +/- 22s
Old: Loraen (225 cha on average) vs L53 Ilis in Sebilis: 127 seconds +/- 28s

If I am doing my math correctly (never guaranteed!) the average charm on a L50 mob is now probably just over 2 minutes, unlikely (15%) to be over 2:30, and extremely unlikely (<1%) to be 3 minutes, which would be my estimate before the patch. It appears you can now charm L50 mobs with approximately the same duration as L53 mobs prepatch. I used a really simple procedure that somehow I went through graduate school without learning called bootstrapping to determine the variance.

caveat: These are just from my normal play so there are two that I broke early with the goblin ring. That probably skews the duration down slightly, but there is no good way to remove them from the set. On the other hand the fact that I only broke two out of almost 20 speaks for itself.

[Mon Jul 15 18:07:35 2013] You begin casting Allure.
[Mon Jul 15 18:08:04 2013] Your Allure spell has worn off. 29s - 6 = 23s
[Mon Jul 15 18:08:23 2013] You begin casting Allure.
[Mon Jul 15 18:15:17 2013] Your Allure spell has worn off. 6:54 - 6 = 406s
[Mon Jul 15 18:16:19 2013] You begin casting Allure.
[Mon Jul 15 18:22:12 2013] Your Allure spell has worn off. 5:53 - 6 = 347s
[Mon Jul 15 18:22:37 2013] You begin casting Allure.
[Mon Jul 15 18:25:08 2013] Your Allure spell has worn off. 2:31+ - 6 = 145s
[Mon Jul 15 18:37:42 2013] You begin casting Allure.
[Mon Jul 15 18:38:03 2013] Your Allure spell has worn off. :21 - 6 = 15s
[Mon Jul 15 18:38:20 2013] You begin casting Allure.
[Mon Jul 15 18:40:33 2013] Your Allure spell has worn off. 2:13 - 6 = 127s
[Mon Jul 15 18:54:41 2013] You begin casting Allure.
[Mon Jul 15 18:56:33 2013] Your Allure spell has worn off. 1:52 - 6 = 106s
[Mon Jul 15 18:56:57 2013] You begin casting Allure.
[Mon Jul 15 18:58:03 2013] Your Allure spell has worn off. 1:06 - 6 = 60s
[Mon Jul 15 18:58:14 2013] You begin casting Alluring Whispers.
[Mon Jul 15 18:59:57 2013] Your Alluring Whispers spell has worn off. 1:43 - 16 = 87s
[Mon Jul 15 19:00:20 2013] You begin casting Alluring Whispers.
[Mon Jul 15 19:01:03 2013] Your Alluring Whispers spell has worn off. :43 - 16 = 27s
[Mon Jul 15 19:01:13 2013] You begin casting Alluring Whispers.
[Mon Jul 15 19:02:15 2013] Your Alluring Whispers spell has worn off. 1:02 - 16 = 44s
[Mon Jul 15 19:03:05 2013] You begin casting Alluring Whispers.
[Mon Jul 15 19:07:28 2013] Your Alluring Whispers spell has worn off. 4:23 - 16 = 247s
[Tue Jul 16 10:38:38 2013] You begin casting Allure.
[Tue Jul 16 10:44:08 2013] Your Allure spell has worn off. 5:30+ - 6 = 324s
[Tue Jul 16 10:53:25 2013] You begin casting Allure.
[Tue Jul 16 10:55:38 2013] Your Allure spell has worn off. 2:13 - 6 = 127s
[Tue Jul 16 10:55:56 2013] You begin casting Allure.
[Tue Jul 16 10:56:45 2013] Your Allure spell has worn off. :49 - 6 = 43s
[Tue Jul 16 10:57:20 2013] You begin casting Allure.
[Tue Jul 16 10:58:33 2013] Your Allure spell has worn off. 1:13 - 6 = 67s
[Tue Jul 16 10:58:52 2013] You begin casting Allure.
[Tue Jul 16 10:59:51 2013] Your Allure spell has worn off. :59 - 6 = 53s
[Tue Jul 16 11:00:11 2013] You begin casting Allure.
[Tue Jul 16 11:00:57 2013] Your Allure spell has worn off. :46 - 6 = 40s

Old Data: http://www.project1999.org/forums/showthread.php?t=92423

[SamwiseRed]

most powerful class needs a nerf.

[koros]

Last night in Mistmoore, on my 28 enc, I soloed the last half of the level using the same mobs I had used to previously solo the first part. Charm seemed no different (better if anything)

[koros]

Also your math looks way off in a few areas - 4 of your samples were well over 3 minutes, with 1 (the 2:31 where you broke charm) having the potential to be. Where did the <1% chance for 3 min charm duration come from? Assuming a normal distribution of charm durations (which I doubt from this, although the sample size is small) the standard deviation would have to be huge.

[Swish]

Kudos to the OP for gathering the info... I guess time will tell on how it feels for everyone. All the enchanters will adjust regardless, just get to 60 and its all easier :p

[Splorf22]

Last night in Mistmoore, on my 28 enc, I soloed the last half of the level using the same mobs I had used to previously solo the first part. Charm seemed no different (better if anything)

I think its extremely hard for the human brain to measure small differences in the average value of highly random processes. How about you check your logs and post? This goes for you too Writ3r.

Also your math looks way off in a few areas - 4 of your samples were well over 3 minutes, with 1 (the 2:31 where you broke charm) having the potential to be. Where did the <1% chance for 3 min charm duration come from? Assuming a normal distribution of charm durations (which I doubt from this, although the sample size is small) the standard deviation would have to be huge.

I said the average has a 1% chance to be over 3 minutes, not an individual charm. I resampled the data many times to create a distribution for the mean, which should be more or less gaussian even though the distribution for an individual charm is not.

[koros]

Misread that part.

However, I still think that 1. The sample size is likely too small, especially for a bootstrapping method, and 2. It's entirely possible (and looks likely) that the distribution is bimodal, or something else entirely. Perhaps MR in seb even got jacked up.

Regarding my experience charming in MM, due to my limited mana regen and the fact that I charm soloed for 3 hours, significant differences in mean charm duration should have made me unable to keep the entire area clear as consistently as I did, mana was tight last time I did this.

I'll do some analysis when I get home from work.

[Barkingturtle]

I suspect J-Boots are about 3% slower, too.

[Swish]

WTS Tinfoil hat with "tashan" effect built in.

[Splorf22]

Misread that part.

However, I still think that 1. The sample size is likely too small, especially for a bootstrapping method, and 2. It's entirely possible (and looks likely) that the distribution is bimodal, or something else entirely. Perhaps MR in seb even got jacked up.

Regarding my experience charming in MM, due to my limited mana regen and the fact that I charm soloed for 3 hours, significant differences in mean charm duration should have made me unable to keep the entire area clear as consistently as I did, mana was tight last time I did this.

I'll do some analysis when I get home from work.

If my analysis is correct, there is a ~10% probability that I was just unlucky with my tests. It's certainly not impossible, so I titled the thread 'reasonable probability'.

Also the biggest issue with MM is probably that L20 mobs will kill each other pretty fast, so you probably never need long charms when soloing. It's at 50+ when mobs HP starts going crazy that you have more breaks.

I suspect J-Boots are about 3% slower, too.

That's probably because you don't know how to use estimators and confidence intervals to actually measure things. Of course, if you did your ego might not be so starved for attention that you have to troll the boards all day.

[Barkingturtle]

That's probably because you don't know how to use estimators and confidence intervals to actually measure things. Of course, if you did your ego might not be so starved for attention that you have to troll the boards all day.

Oh calm down, Champ. The J-Boots thing is a classic EQ joke and meant to inspire lulz amongst Classic EQ enthusiasts, not angst. Man, I don't remember what I did to you but it must have been hurtful. I'm sorry.

[Splorf22]

umadbro

Will you ever make a post that is not a variant of umadbro?

Anyway, as much as I enjoy watching you embarrass yourself with your usual vapid comments, I'd prefer that you stay off what has the potential to be an interesting mathematical thread.

[Widan]

onyxia deep breaths more too

[Spitty]

This thread isn't as...charming...as I expected it to be.

[Barkingturtle]

Will you ever make a post that is not a variant of umadbro?


Well, I made a pretty funny J-Boots joke a few posts up but it seems to have only made you mad. Guess I just can't win. You do seem kinda mad, though. Am I the only one seeing this?

[lecompte]

Please ignore all off topic posts -- including Splorf's responses to Barkingturtle.

Splorf/Loraen. I don't have prepatch data for 60 enchanter charms but I bet you do for mobs outside of Seb. Name a couple, I'll put charisma at 230, like yours, and post my logs for you to review.

[Splorf22]

umadbro

The saga continues.

Anyway back on topic, I have logs of me doing quite a bit of charming in many zones. The problems are a) some of the time I broke charm with the ring b) its not easy to tell the level of the mob I was charming (although I suppose something could be done there with the parsing). Your best bet Lecompte is the data on the Ilis that I linked in the first post.

[lecompte]

That'd be fine, but with that we can't exclude MR buff to mobs in Seb or a nerf to tash. I was hoping to exclude some other potential factors while checking the hypothesis.

[Splorf22]

I was very careful to dispell them first. It should be good data.

[lecompte]

I was very careful to dispell them first. It should be good data.

I'm sorry, I didn't mean an actual MR buff, I mean an increase to their racial/class magic resistance.

[Nirgon]

If indeed charm is less effective, this is a fix

[-Catherin-]

spent most of the day in soloing in seb today. charm did not really feel any different, or if it was it was negligible. trying to lull on the other hand was atrocious. resists and crit resists all over the place.

[Splorf22]

I'm sorry, I didn't mean an actual MR buff, I mean an increase to their racial/class magic resistance.

The mobs in this sample are all from howling stones south and east (2 devourers and 1 sepulcher iirc).

If indeed charm is less effective, this is a fix

Evidence? And please don't say 'Well back in the day I talked to this guy" like you usually do.

Look I believe charm is somewhat overpowered although not insanely so like it was before Kunark here. From a pure balance standpoint a 20% nerf to the average duration isn't unreasonable. But based on what I have read in Xornn's guide, charm here is already weaker than it was on live. I even found quotes of people on live who were charming in the planes before Kunark. Unfortunately it took Verant a while to really understand their game, and the first couple expansions are full of serious balance issues. That's just classic.

spent most of the day in soloing in seb today. charm did not really feel any different, or if it was it was negligible. trying to lull on the other hand was atrocious. resists and crit resists all over the place.

I'd be happy to be wrong :D I just don't trust people's feelings when it comes to this is all.

[pasi]

My sample pool was small enough that I simply chalked things up to bad luck yesterday, but I noticed more resists (and charm breaks) than usual. Again, small sample size and potential for variance.

With that said, people need to also acknowledge that not everything in the patch is listed. One of the largest changes that seems to be getting no mention is that NPC mana appears to be functioning properly since the patch. That, or Theft of Thought is actually taking mana from NPCs. I soloed Ghost of Kindle for someone yesterday which normally requires me to chain mez her from 39% down, but she was actually oom from spamming nukes. This was a new feeling for P99 for me - NPCs before felt like their mana pool was in the 5 digits. I'll have to play around some more to see if this is a global thing.

[Hawala]

Last night in Mistmoore, on my 28 enc, I soloed the last half of the level using the same mobs I had used to previously solo the first part. Charm seemed no different (better if anything)

Same at level 35, better.

[koros]

Splorf,

This data makes me quite certain either 1. Charm wasn't nerfed, or 2. Your sample size is WAY too low, additionally it's likely some unknown instrumental variable was in effect here due to the strange distribution.

Population mean of 127, with a standard deviation of 121 and a standard error of estimation of 28. This doesn't add up. I think bootstrapping was definitely the wrong method to use.

[Splorf22]

Splorf,

This data makes me quite certain either 1. Charm wasn't nerfed, or 2. Your sample size is WAY too low, additionally it's likely some unknown instrumental variable was in effect here due to the strange distribution.

Population mean of 127, with a standard deviation of 121 and a standard error of estimation of 28. This doesn't add up. I think bootstrapping was definitely the wrong method to use.

I guess I'm a bit confused? I thought Central Limit Theorem implied that the standard error of the distribution of the mean should be 1/sqrt(n) that of the original distribution. 121 / sqrt(18) = 28.5 which is pretty close.

I can say that I know the distribution for charm for these types of mobs. You can think of it as being two uniform distrubtions: a "good" charm with say 10% chance, U(6-15 min) and a "normal" charm with say 90% chance, U(0-6 minutes). As your level gets higher or lower I'm not sure whether the midpoint changes or the chance of good/normal changes or both.

[koros]

You're right on central limit theorem, that's where I got the numbers I mentioned from.

Given the very high standard deviation, (only 68% of charms last 1 tick to 4 minutes? ) I don't think we can infer that we've reached an accurate mean.

Additionally, due to what seems to be negative kurtosis and positive skewness from the distribution, more samples are in order. A lot more.

[falkun]

If you have two data points that are different than the others (read: 2 intentional breaks vs. "natural" breaks) and you know which two data points those are, why can't you just highlight+delete them? You've got a relatively small sample size, shouldn't be that hard to find them and delete them from the set.

[newsmurf]

i spend more time at ng and other varius areas of seb than catherine and i can agree with all of you to some degree, the mr of mobs has been buffed.

[SirAlvarex]

On the inverse, when I was pulling in Seb on Sunday I was resisting about 70% of the spells the lower level frogs were throwing at me. I eventually stopped charm pulling casters because "whats the point?"

To point you at your own thread, I think they may have implemented the 255 cap without telling us. (http://www.project1999.org/forums/showthread.php?t=98648)

[ripwind]

To point you at your own thread, I think they may have implemented the 255 cap without telling us. (http://www.project1999.org/forums/showthread.php?t=98648)

Sometimes code changes go up to production by accident, too.

[Splorf22]

If you have two data points that are different than the others (read: 2 intentional breaks vs. "natural" breaks) and you know which two data points those are, why can't you just highlight+delete them? You've got a relatively small sample size, shouldn't be that hard to find them and delete them from the set.

Because those are not random. The ones I break are likely to be the better ones.

Actually, probably the best thing to do would be resample from the set of unintentionally broken charms that lasted at least as long as this one.

Now that I think about that, that's a pretty good idea. Hmm.

[falkun]

No, that would artificially alter the data.

Take all data with the same controls (CHA, player_level, mob_level, debuffs (tash & malo), "natural" charm break, zone, mob), and only discard data that does not match the controls. If you start discarding data because it doesn't fit what you expect, then you are fitting the data to your conclusion(s), not the conclusion(s) to your data.

[Splorf22]

Imagine I am charm soloing for XP. I charm something and send it at thing two. If charm breaks, I recharm it. Finally after it kills the enemy I break charm and kill it. The charms I break are disproportionately the longer ones. Example:

Fight #1: Charm (100s), break, Charm (100s), self break, kill
Fight #2: Charm (200s), self break, kill

If we simply average those we get 133s (and the actual duration is higher!). If we throw out the two that were broken by the caster, we get 100s, which is clearly wrong. And this bias will persist as you add more samples.

I am suggesting "pretend the charm was allowed to continue until it broke on its own by replacing it with one that broke naturally that lasted at least that long"

[falkun]

So you are artificially lowering the mean by instigating the charm break. Wouldn't these be offset by the super short natural charm breaks to get your true distribution? Either way, let the data speak for itself. Unbiased data cannot be right or wrong, it just is, only analysis can be correct or incorrect.

[-Catherin-]

i spend more time at ng and other varius areas of seb than catherine and i can agree with all of you to some degree, the mr of mobs has been buffed.

Im not sure where you get that you are at seb camps more than me :p but after spending more time there I'm still convinced there is no change, or if there is, it is too negligible to prove or disprove

[Splorf22]

Yeah I'm starting to believe I was just pretty unlucky :( I had fun doing the math though!

[lecompte]

Yeah I'm starting to believe I was just pretty unlucky :( I had fun doing the math though!



And that is what is important. /me pats Splorf22 on the head.

[wwoneo]

most powerful class needs a nerf.

See, this kind of logic is the reason why games start to fail. It's not that the "most powerful" classes need nerfs. If there's REALLY an imbalance, then the correct course of action is to buff the crappy classes! If you want to complain that course of action makes the game easier, then buff the mobs too, but the psychological aspect of nerfing your character is quite frustrating. P.S. I don't play a chanter.

[Sithel1988]

charm lower level mobs???????????????????

[Bohab]

what does suck is charm break mid DDoS when you are left defenseless, that is more unlucky than anything

Especially when it's an essence harvester...