Info about attack speed and question about aggdef

[______________Rustot]

I have found such formulas:

attackspeed = max( 1, weapatkspeed + 0.75 - aggdef - initiative / 600)
rechspeed = max( 1, weaprechspeed + 0.75 - aggdef - initiative / 300)
here aggdef is -1 .. +1

for example, MA speeds is 1.25s (I've not mistake here?). Then with max aggdef we achieve the maximal speed at ZERO initiative (1.25 + 0.75 - 1 - 0 == 1.0), with middle aggdef at initiative 600 (1.25 + 0.75 - 0 - 600/600). With middle aggdef and init=300 we have maximal recharge speed, but 1.5s attack speed.

In other words, declared attack and recharge speeds are given for zero initiative and 3/4 aggdef position from middle to max. Full aggdef diapason regulate speeds in 2 seconds diapason. 300 initiative subtracts 0.5s from attack speed and 1s from recharge speed.

I already saw the similar formulas, but without 0.75. but I have some reasons to consider it is correct. Can somebody confirm or deny these formulas? It is desirable by results of experiments.

It was info. Now question.

Whether aggdef influences something except attack speed and probability to evade? Ie if i decrease aggef, can I lose something except speed? For example, maybe aggdef influences probability of hit, or probability of making criticals, or something else?

[_______bunderscratch]

what I heared about evaide : decrease overall dmg and critical hit probability.
about aggr/def : all inits, all evaides and concealment.

all _very_ arguable.

Also I heared it's now not a init/600 init/300 for attack-recharge but:
first opinion:
init/1000 init/1000
second opinion
init/1000 for attack only, recharge not decreased.

Accoding with both opinions this was introduced in some patches.

[_________________Zuh]

I am about to perform precise testing on speed and slider when 14.2 comes out. As you can have a total IPR, then you can raise progressively RI and test all the way...


Beside that, where do your formulas comes from ? Because indeed they are quite accurate.

[______________Rustot]

Originally posted by Zuh
Beside that, where do your formulas comes from ? Because indeed they are quite accurate.

One people said that this formulas are used for animation of attack/recharge timebar in AO client.
There is no guarantee, that on a server side the same formulas are used, but also there is no reason to use different.

[Myga]

Well I dont believe server and client use same timer. For instance I normally fire speedcapped, 1s atk, 1s recharge. Then comes a doc mob and gives me a init debuff, now on client my recharge stays 1s but the atk speed is suddenly 3secs or so. And the hit messages and atk/recharge bar appearing on client are usually totally out of synch.

Another funny thing, a 1s atk, 2.5sec recharge weapon like gluegun appear to fire speedcapped for me with +15 crit scope for about 100 ranged init and agg-def on full agg.

[______________Rustot]

PS. For nanoinit attack time (hands movements) is

time = attacktime - nanoinit/200 - aggdef

recharge time (nano lock) imho fixed

[______________Rustot]

Originally posted by MSag
And the hit messages and atk/recharge bar appearing on client are usually totally out of synch.

Hence only experiments will help

[______________Krneki]

Timings are a bit different (or the chat isn't displaying correctly). In theory with maxed physical init i should be hitting 30 hits in 60 sec with my fists, but I do only about 26-27, but if I time the attack/recharge bars when i'm out of range to hit and no animation is taking place they both charge and recharge 30 times.

[______________Rustot]

maybe was 3-4 evadings per minute?

[______________Krneki]

If it did, there would be a time gap betwin hits...and I don't think a lvl 170 can evade 1150 MA Skill

[______________Rustot]

Also it can be result of unsynchronized lags on a server.

For example

make hit;
sleep until current time + 1 sec
make hit;
sleep until current time + 1 sec
...
is wrong

var = current time + 1 sec;
make hit
sleep until var
make hit
var = var + 1 sec
sleep until var
make hit
var = var + 1 sec
sleep until var
...
is right, because sleep on nonrealtime os cannot be strictly given time

In such case the experiments are useless, too large error

[_________________Zuh]

Thanx for the info about the weapons speed. Looks very accurate.


I haven't thought of the +0.75/+0.75 to weapon speed when I did my first tests (and found wierd results). We are doing more tests, and I'll keep you aware of them.

While your forumlas seems tip-top. The good ones .