Derra wrote: »

Vapirko wrote: »

Okay good to know. In terms of ultis, I'm just going off what other people have told me in game. The health desync happens constantly. I didn't know lag would cause say a dawnbreaker animation to not show up at all.

Weaponswap/dodgeroll animationcancel allows you to skip most offensive animations entirely.

Lmao this thread is so old. Back when I was still floundering around looking for something to blame my losses on.

Why would I get a macro when I can do everything a macro can by hand? Don’t make excuses for dying, instead learn to be better

People need to realize that everyone is playing in lag, it is not just them. There are times when you might see an animation for a skill and it hits you 2 seconds later. There are times where you get CC'd and for whatever reason you can't break free, or when you do you can't use your skills for a few seconds because the game hasn't internally registered that you've broken free despite your character moving.

So many people are under the belief that they encounter players cheating or using macros in this game when in reality it is literally the game, and those people you think are cheating are going through the same thing. I look at my death recap sometimes and see I get hit with 5-6 things I didn't even see, and other players whisper me after I kill them and say they experience the same thing.

To understand one aspect of this issue you have to know why light attack weaving is important. It is a critical ability because it maximizes damage per unit time. The way to do this, is to exactly time hitting the ability key after the LA key. If you hit the abilities or LA too soon, one or both won't fire = dps loss. If you hit the ability too long after the LA, you will get damage from both, but you lose time when you are doing no damage.

The optimum time for Ability after LA is about 0.6 seconds (I forget the exact number). For the following, you need combat metrics. Go to your target dummy and use a single ability and do ~ 10 LA/ abilities. Then, open CM and use "combat log". CM reports time of every event to the thousandth of a second. Deduct the time of the LA from the time of the ability. This is the "excess time" you have wasted. Ideally you want (LA - Ability) = 0.002 seconds. (in other words the LA and ability hit at almost the same time).This is the most efficient LA weaving.

The problem is humans , unlike computers, are inconsistent. The best dps testers on dummies are super consistent and minimize excess time. But human reaction time is 0.25 seconds.

Here are three logs. In the first one, I tried to simply ensure that I got equal numbers of LA and abilities. Then I did the above calculation to find the "excess time" for each attack.

Excess time (seconds): 0.3, .14, .2, .25, .3, .3, .2, .34, .11, .15, .25, .18, .04, .04, .23, .15. Mean: 0.198 Standard Deviation: .09. Not a very good parse.

Here is some of a parse on a non-trial dummy which was 35k dps.: 0.11, 0.46,. 0.01, .13, .1, .13, .16, .09, .04, .002. Mean .12 SD ,13. Much better. But you can see a few terrible time loss rounds (.46) and one perfect one (0.002). That last one times the light attack and ability to within 2 one thousandths of a second-- way faster than human reaction time.

Now what if you see something like this in cyrodil: 0.002, 0.002, 0.004, 0.007, 0.08, 0.002, 0.003. Mean 0.014, SD .03. In this example the player hit 5/7 combos with an "excess loss" of 4 one thousandths of a second or less. Note the mean is ten fold lower than the first 2 parses and the SD is also very low.

The difference in these three parses is analyzable statistically and , as you might expect, the P value of data set 1 vs 3 is <0.0001. But, the p value of set 2 vs 3 is also significant: 0.05. The p value between sets 1 and 2 does not reach statistical significance. (Unpaired T test )

The point of all this is the simplest macro is to program a mouse button so that when you hit it the following happens: (LA, Pause 0.6 seconds, Ability). You create a button that does this for your damage abilities. This will guarantee, within the noise created by lag, that your "excess time" is minimized so that you get the maximum dps. This would be the only practical macro in Cyrodil, is my guess. "chains" of abilities are cancelled out by player movement etc.

What about lag? Could it mask simple macro use? The point is that statistics basically exist to discover the population means underlying noisy samples.

One thing I don't know is whether ESO uses a timestamp mechanism, ie allows the local machine to time stamp its abilities which are then decoded by the server. This would affect whether lag could "mask" macro use. If the mousestrokes are timestamped, then lag could cause them to hit you all at once, but it would have less of an effect on the time gap between LA and abilities. Statistical testing as shown above, however, suggests one could (especially with larger sample sizes) easily distinguish between users who are weaving their own abilities and users who are using simple mouse macros.

What about the people parsing 100k on trial dummies? I can guarantee there are players a a lot better than me, with higher dps and more skill in every way. They must, by the above logic, have better "excess loss" scores. (or some of them use simple mouse macros!). Surely you can improve your timing, with thousands and thousands of practice strokes and hundreds and hundreds of hours of game time. I'm curious what the mean "excess loss" of these players would be (not sure if this data exists). But, I would argue that the standard deviation of a human will "always" be higher than a macro. This (low standard deviation) is the fundamental issue IMO.

You opinion on Magcro is misinformed, regardless of how good blastbones is.