[Mage] Rolling Ignites: are they back? - Page 6

drowsy · 01/12/08, 12:44 PM

Originally Posted by manly

Sometimes you get the fireblast ignite, sometimes you get fireball ignite. Something else to note; the order in the logs of fireball/fireblast is not consistent either. Maybe there is a link.

It looks like in all three of those pictures, the ignite damage was based off of the spell that appears second in the log, which would support your theory of a spell carrying an ignite payload that overwrites what is there when it lands. Is it possible that your fireblast might lag on occasion, causing it to get to the server after the fireball lands?

zurmagus · 01/12/08, 1:00 PM

Just for informational purposes here, you were at max Fire Blast range (26y) and tapping it just as the fireball finished?

Yeah, I was 25-26 yards away at the time.

manly · 01/12/08, 4:19 PM

Originally Posted by drowsy

It looks like in all three of those pictures, the ignite damage was based off of the spell that appears second in the log, which would support your theory of a spell carrying an ignite payload that overwrites what is there when it lands. Is it possible that your fireblast might lag on occasion, causing it to get to the server after the fireball lands?

You're correct, I actually had not noticed somehow. Might want to have more samples though, although so far it would fit the theory.

And no, I tried both 26 yards and ~15 yards many times over. It did not seem to affect the results.

BrTarolg · 01/12/08, 6:27 PM

Ok guys - from my personal experience in how games work, it would sound there is almost certainly a different way of detecting how a mob hits you and how a player hits you.

Firstly, there are two options - either the client or the server decides when the spell lands.

If it was the server, there is almost certainly no chance of it calculating the "crazy attack animations" which you are all experience - at best it will simply calculate the distance apart and just apply some base time to that (it probably doesn't even account for line of sight)

In fact, the fact i bet you could make hacks which both walk through walls and ignore line of sight issues suggests that these are calculated on the client.

However, as for outrunning shadowbolts, I bet the client calculates when NPC's hit you, however you also calculate when your own bolts hit (so if you target a player, you actually decide when your bolt hits). However, ignite procs (and timings - dot timings are almost certainly much more accurate than the timing of when the bolt hits) will always be decided by the server, and probably are properly synchronised across everyone.

Overall - i can't really do the calculations myself - but if you want to work out when your *actual* bolt hits, Look at when the first ignite tick happens, subtract 2 seconds from it + latency, and that is when the server decides your bolt hits.

Mode · 01/12/08, 11:55 PM

Has anyone considered doing some tests with warriors spell reflect?

As a warrior, I can tell you with 60% certainty that SOME spells can be reflected only if your spell reflect is up before the spell while others can be reflected if you hit it mid-flight. I can't give you examples off-hand as I'm terrible about using it in PvP, but I'll try to see if I can spot examples next time I'm in AV as prot.

To answer another question,

Spell animations continue after the target dies. Anyone that plays AV heavily has surely encountered the phenom of seeing/hearing spell animations fly into the graveyard.

Quantum · 01/13/08, 12:02 AM

If it helps the discussion somehow, casting Wrath can immediately proc Nature's Grace, even if the bolt just left your hands, and therefore the information whether the cast is a crit or not should be instantly determined. I assume that Wrath and Fireball are handled the same way.

Proudmoore · 01/13/08, 3:25 PM

Originally Posted by Quantum

If it helps the discussion somehow, casting Wrath can immediately proc Nature's Grace, even if the bolt just left your hands, and therefore the information whether the cast is a crit or not should be instantly determined. I assume that Wrath and Fireball are handled the same way.

Not necessarily. The behaviour of Combustion allowing Fireblast to take advantage of the bonus crit chance while a Fireball is still in flight (that will crit in the end) suggests that Fireball at least is not decided until it actually reaches the mob, and the damage it does is recorded. If whether the Fireball will crit or not is decided at the end of the cast, Combustion should dispel itself as you cast the fireball that will be the third crit.

Whether a spell crits/hits/misses or what not seems in every case I have examined, to be decided when the spell hits the target - when the animation of the ball actually reaches the target, or immediately in the case of spells like Scorch and Fireblast. Haste effects are applied as you begin casting, and bonus damage/mana cost effects are applied as the spell finishes casting - although I have onlylooked at Mage spells, I don't know about any other class.

Etherealz · 01/13/08, 6:45 PM

Originally Posted by Proudmoore

Not necessarily. The behaviour of Combustion allowing Fireblast to take advantage of the bonus crit chance while a Fireball is still in flight (that will crit in the end) suggests that Fireball at least is not decided until it actually reaches the mob, and the damage it does is recorded. If whether the Fireball will crit or not is decided at the end of the cast, Combustion should dispel itself as you cast the fireball that will be the third crit.

I don't believe that's a very solid proof. Combustion not being removed until the fireball reaches the target does not imply that the crit is not calculated until impact. There is no rule that says - once crit is calculated - combustion must be removed. More than likely combustion is just poorly coded to be "exploitable".

Enquillion · 01/14/08, 5:51 AM

Having read through this entire thread I found it a bit surprising that no one even considered the thought of casting Scorch with 2 sec intervals and perhaps possibly reaching a point where the Ignite would be so big that it would outweigh the drawback of casting a low DPS spell 75% the time (and 25% the time not even casting). Maybe this was blatantly obvious to be impossible, but I gave it a try.

In the name of science I respecced and equipped all the crit gear I could find, reaching 40% Scorch crit in the end, just to find out how it would behave. I hope you'll find some of this useful, it's my first post here and I pray you will consider it a contribution.

First of all I will say, that this is the kind of thing that would only work in a perfect world. I was of course using my G15 to get exactly 2 seconds between each Scorch, and while it many times did succeed in making the Ignite roll, it would generally fail far before the point where the Ignite got so big it could even be remotely worth doing this. I just had to be lucky with latency to have the Scorch land right on time.

Lowering the interval to 1.95, 1.98 and even 1.99 was bad, as the Ignite didn't get to tick between two Scorches, apparently it's that "sensitive" so to speak - in fact not even 2 was always dependable, and yet it would sometimes lose the Ignite if I crit with 4 sec between the first and the second. I didn't try using an interval larger than 2, as I saw no point; I already sometimes lost it with less.

Anyway, what you might find relevant, as I can't recall it being established yet: You do have two chances to make it roll.

10:06:06, Dr. Boom suffers 689 Fire damage from your Ignite.
-Scorch hit, I didn't copy it-
10:06:08, Dr. Boom suffers 691 Fire damage from your Ignite.
10:06:09, Your Scorch crits Dr. Boom for 1734 Fire damage.
10:06:10, Dr. Boom suffers 692 Fire damage from your Ignite.

The next Ignite tick was larger and all in all I got [edit: roughly] 60% bonus damage rather than 40% from my Ignite before it died in that particular case. However, it's not reliable at all to hope that the second crit will do the trick, except perhaps with the whole travel time mechanic.

Hate Monkey · 01/14/08, 6:53 AM

I do have one question to ask before I post anything about this: Under normal fireball crit situations, what is the highest Ignite tick you have seen, rolling, non rolling, or possible rolling.

I don't want to make a huge worthless post if my data turns out to be influenced by a different factor I haven't thought of yet.

kralizec · 01/14/08, 6:57 AM

As Enquillion said, the free ignite tick can be either the first (like all the examples posted before on this thread), but possibly also the second. The expected casting time would be around 4 seconds. Like, hasted pyros, or two scorch with a slight delay ?

Cardynal · 01/14/08, 11:00 AM

Originally Posted by Enquillion

Having read through this entire thread I found it a bit surprising that no one even considered the thought of casting Scorch with 2 sec intervals and perhaps possibly reaching a point where the Ignite would be so big that it would outweigh the drawback of casting a low DPS spell 75% the time (and 25% the time not even casting). Maybe this was blatantly obvious to be impossible, but I gave it a try.

In the name of science I respecced and equipped all the crit gear I could find, reaching 40% Scorch crit in the end, just to find out how it would behave. I hope you'll find some of this useful, it's my first post here and I pray you will consider it a contribution.

First of all I will say, that this is the kind of thing that would only work in a perfect world. I was of course using my G15 to get exactly 2 seconds between each Scorch, and while it many times did succeed in making the Ignite roll, it would generally fail far before the point where the Ignite got so big it could even be remotely worth doing this. I just had to be lucky with latency to have the Scorch land right on time.

Lowering the interval to 1.95, 1.98 and even 1.99 was bad, as the Ignite didn't get to tick between two Scorches, apparently it's that "sensitive" so to speak - in fact not even 2 was always dependable, and yet it would sometimes lose the Ignite if I crit with 4 sec between the first and the second. I didn't try using an interval larger than 2, as I saw no point; I already sometimes lost it with less.

Anyway, what you might find relevant, as I can't recall it being established yet: You do have two chances to make it roll.

10:06:06, Dr. Boom suffers 689 Fire damage from your Ignite.
-Scorch hit, I didn't copy it-
10:06:08, Dr. Boom suffers 691 Fire damage from your Ignite.
10:06:09, Your Scorch crits Dr. Boom for 1734 Fire damage.
10:06:10, Dr. Boom suffers 692 Fire damage from your Ignite.

The next Ignite tick was larger and all in all I got [edit: roughly] 60% bonus damage rather than 40% from my Ignite before it died in that particular case. However, it's not reliable at all to hope that the second crit will do the trick, except perhaps with the whole travel time mechanic.

Post a full log please

From our experience, you cannot roll a scorch because there isn't a "flight time". Fireball having a flight time is a big part of this theory. If I had to guess...you're only experiencing the ignite push back...which is an intended ignite mechanic and only gives the 40% damage...not a roll.

manly · 01/14/08, 12:42 PM

First thing first, I did recommend testing if travel time was involved in the process, and the best way to test it would be with a g15 to do 2s scorches. However, if you did read what I posted, you would have seen that you need, as counter-intuitive as it might sound, cast times > 2s.

However, something that hasn't been said, but this just proves the point further, it is blatantly obvious that ignites are 'delayed' server-side before it gets applied. The proof is rather simple. It is in the screenshots I posted. If travel time was the only factor playing here, you would not see those cases happen. If the ignites were applied the time the spell was done casting on the server, then logically it is not possible to have the fireblast ignite override the fireball ignite.

There are a few explanations possible, namely:

1- travel time server side and travel time client side are unlinked (ie: the fireball 'lands' much earlier on the server, before your client-side fireblast, despise you seeing that you clearly land fireblast before fireball)
2- ignite is being applied sometime later after the crit lands.

What 2s scorch giving rolling ignites proves is that explanation #2 is correct. It does not proves or disproves #1. But it shows that #2 is a factor that does come into play.

-------
I often get 6-8k ignite ticks when stacking bloodlust/icy veins, but those are merely ignite stacking (ie: the accumulation of all ignites).

drowsy · 01/14/08, 2:23 PM

Wish I had timestamps on.

This was done with icy veins, heroism, and 42 haste rating. What is seen is the full value of the first ignite rolled into the second despite the second fireball crit landing before the first ignite tick. So the delay between a spell landing and the start of its ignite does affect how this all works. In this case I think that even though the fireball landed before the ignite tick, its payload was subject to the ignite delay people have noticed, pushing the new ignite "payload" just past the first tick.

What I think happened, in order:
Crit lands for 5399, ought to ignite 1080 x 2
Ignite starts
Crit lands for 5342, ought to ignite 1068 x 2
Ignite 1080 (free)
2nd ignite starts
Ignite 2149
Ignite 2148

Enquillion · 01/14/08, 2:44 PM

Originally Posted by Cardynal

Post a full log please

From our experience, you cannot roll a scorch because there isn't a "flight time". Fireball having a flight time is a big part of this theory. If I had to guess...you're only experiencing the ignite push back...which is an intended ignite mechanic and only gives the 40% damage...not a roll.

I'm sorry, but I kind of specced back and logged off because I was slightly disappointed that it didn't work out.

I'm well aware of how Ignite works, and it's clear to me that if you crit with exactly 2 second intervals it is possible to roll it, although it's unreliable. I'm sure some of you are fire specced normally and have a G15, so you could easily confirm it. It doesn't always happen, but it can happen, which I'm sure you will see.

Etherealz · 01/14/08, 5:21 PM

Originally Posted by Enquillion

I'm sorry, but I kind of specced back and logged off because I was slightly disappointed that it didn't work out.

I'm well aware of how Ignite works, and it's clear to me that if you crit with exactly 2 second intervals it is possible to roll it, although it's unreliable. I'm sure some of you are fire specced normally and have a G15, so you could easily confirm it. It doesn't always happen, but it can happen, which I'm sure you will see.

I ran a test on this just to be sure - Always fun beating on a warlock even if it's your own character.

1/14 15:51:34.311  Your Scorch crits Etherealx for 498 Fire damage.-----
1/14 15:51:34.764  Etherealx is afflicted by Fire Vulnerability.
1/14 15:51:34.874  You gain 15 Mana from Master of Elements.
1/14 15:51:35.108  Etherealx is afflicted by Ignite.
1/14 15:51:35.124  You gain Combustion (2).
1/14 15:51:36.374  Your Scorch hits Etherealx for 383 Fire damage.-----------
1/14 15:51:36.718  Etherealx is afflicted by Fire Vulnerability (2).
1/14 15:51:36.718  You gain Combustion (3).
1/14 15:51:36.827  Etherealx suffers 100 Fire damage from your Ignite.
1/14 15:51:38.374  Your Scorch crits Etherealx for 520 Fire damage.--------
1/14 15:51:38.733  Etherealx is afflicted by Fire Vulnerability (3).
1/14 15:51:39.124  You gain Combustion (4).
1/14 15:51:39.186  You gain 15 Mana from Master of Elements.
1/14 15:51:40.374  Your Scorch crits Etherealx for 550 Fire damage.--------
1/14 15:51:40.733  Combustion fades from you.
1/14 15:51:40.749  Etherealx is afflicted by Fire Vulnerability (4).
1/14 15:51:40.749  You gain 15 Mana from Master of Elements.
1/14 15:51:42.389  Your Scorch hits Etherealx for 417 Fire damage.---------
1/14 15:51:42.624  Etherealx suffers 264 Fire damage from your Ignite.
1/14 15:51:42.749  Etherealx is afflicted by Fire Vulnerability (5).
1/14 15:51:44.405  Your Scorch hits Etherealx for 415 Fire damage.----------
1/14 15:51:44.639  Etherealx suffers 264 Fire damage from your Ignite.
1/14 15:51:44.749  Ignite fades from Etherealx.

Crit damage = 1568
Expected ignite = 627.2
Actual ignite = 628
Actual ~= Expected
Yes its only a small sampling but I am quite sure that the ignite bug has to do with fireball's travel time. You can see that with the scorches timed at just over 2 seconds that the ignite is simply pushed back.

Edit: I ran a second test with scorch using 2060ms delays just to be sure.

1/14 16:48:02.858  Your Scorch crits Etherealx for 564 Fire damage.
1/14 16:48:03.061  Etherealx is afflicted by Fire Vulnerability (3).
1/14 16:48:03.468  You gain Combustion (4).
1/14 16:48:03.468  Etherealx is afflicted by Ignite.
1/14 16:48:03.514  You gain 15 Mana from Master of Elements.
1/14 16:48:05.093  Your Scorch crits Etherealx for 584 Fire damage.
1/14 16:48:05.093  Etherealx suffers 112 Fire damage from your Ignite.
1/14 16:48:05.468  Etherealx is afflicted by Fire Vulnerability (4).
1/14 16:48:05.874  You gain Combustion (5).
1/14 16:48:05.936  You gain 15 Mana from Master of Elements.
1/14 16:48:07.139  Your Scorch crits Etherealx for 594 Fire damage.
1/14 16:48:07.468  Etherealx is afflicted by Fire Vulnerability (5).
1/14 16:48:07.468  Combustion fades from you.
1/14 16:48:07.921  You gain 15 Mana from Master of Elements.
1/14 16:48:09.171  Your Scorch hits Etherealx for 448 Fire damage.
1/14 16:48:09.546  Etherealx suffers 292 Fire damage from your Ignite.
1/14 16:48:11.561  Your Scorch hits Etherealx for 448 Fire damage.
1/14 16:48:11.561  Etherealx suffers 292 Fire damage from your Ignite.
1/14 16:48:11.905  Ignite fades from Etherealx.

Crit damage = 1742
Expected ignite = 696.8
Actual ignite = 696
Actual ~= Expected
Once again no bonus ignites using scorch

Running a similar test with rank1 fireball

1/14 16:10:37.202  Your Fireball crits Etherealx for 132 Fire damage.-- first crit
1/14 16:10:37.655  You gain Combustion (4).
1/14 16:10:37.655  Etherealx is afflicted by Ignite.
1/14 16:10:37.655  You gain 9 Mana from Master of Elements.
1/14 16:10:39.139  Etherealx suffers 1 Fire damage from your Fireball.
1/14 16:10:39.202  Your Fireball hits Etherealx for 88 Fire damage.-- hit
1/14 16:10:39.374  Etherealx suffers 27 Fire damage from your Ignite. - expected ignite
1/14 16:10:40.030  You gain Combustion (5).
1/14 16:10:41.186  Etherealx suffers 1 Fire damage from your Fireball.
1/14 16:10:41.233  Your Fireball crits Etherealx for 118 Fire damage. second crit
1/14 16:10:41.374  Etherealx suffers 26 Fire damage from your Ignite. ---free ignite
1/14 16:10:41.718  Ignite fades from Etherealx.
1/14 16:10:41.764  You gain 9 Mana from Master of Elements.
1/14 16:10:42.061  Etherealx is afflicted by Ignite.
1/14 16:10:42.061  You gain Combustion (6).
1/14 16:10:43.186  Etherealx suffers 1 Fire damage from your Fireball.
1/14 16:10:43.249  Your Fireball crits Etherealx for 128 Fire damage.--- third crit
1/14 16:10:43.655  Combustion fades from you.
1/14 16:10:43.780  You gain 9 Mana from Master of Elements.
1/14 16:10:45.202  Etherealx suffers 1 Fire damage from your Fireball.
1/14 16:10:45.264  Your Fireball hits Etherealx for 90 Fire damage.
1/14 16:10:45.686  Etherealx suffers 63 Fire damage from your Ignite.
1/14 16:10:47.218  Etherealx suffers 1 Fire damage from your Fireball.
1/14 16:10:47.358  Your Fireball hits Etherealx for 89 Fire damage.
1/14 16:10:47.686  Etherealx suffers 63 Fire damage from your Ignite.
1/14 16:10:48.077  Ignite fades from Etherealx.

Just for totals
Crit damage = 378
Expected ignite = 151.2
Actual ignite = 179

What I actually found interesting about the second test was in the middle of the sequence my combat log recorded a fireball crit, ignite faded after the crit landed, and then reapplied itself. Obviously it was some sort of ignite lag. The interesting part is that I still recieved the free ignite tick.

My thoughts on that would be that the ignite lag is only experienced by the client.

niduab · 01/15/08, 4:00 AM

Hi, I have a quick question of clarification.

If fireballs are being cast at speeds less then 2 seconds in duration (for example, 1.75s), does one lose damage from his/her ignites? Or is it more correct to say, 'One will not get any extra bonus ignite damage when fireballs have a cast time less then two seconds'.

If I'm wrong in both cases I would be grateful to be enlightened by someone else who has a firm grasp on the mechanics of these rolling ignites.

Enquillion · 01/15/08, 4:44 AM

I logged a friend's mage, he was conveniently fire specced. I'll have to tell you - I'm not imagining things. Apologies for not having a fancy combat log, merely copying from EavesDrop, but you should find the following interesting. The test was done, again, with having the G15 spam Scorch on a 2 sec interval.

YUID: Yet Uncaused Ignite Damage

09:17:51, Your Scorch crits Dr. Boom for 1697 Fire damage.
09:17:53, Your Scorch crits Dr. Boom for 1680 Fire damage.
09:17:55, Your Scorch hits Dr. Boom for 1102 Fire damage.
09:17:55, Dr. Boom suffers 675 Fire damage from your Ignite.

YUID: ~675

09:17:57, Your Scorch crits Dr. Boom for 1760 Fire damage.
09:17:59, Your Scorch hits Dr. Boom for 1152 Fire damage.
09:17:59, Dr. Boom suffers 689 Fire damage from your Ignite.

YUID: ~691

09:18:01, Your Scorch crits Dr. Boom for 1678 Fire damage.
09:18:01, Dr. Boom suffers 690 Fire damage from your Ignite. <-- Key line, check timestamps

YUID: ~671 (the ~691 is done at the same time as the Scorch crits)

09:18:03, Your Scorch crits Dr. Boom for 1731 Fire damage.
09:18:05, Your Scorch crits Dr. Boom for 1726 Fire damage.
09:18:07, Dr. Boom suffers 1371 Fire damage from your Ignite.
09:18:09, Dr. Boom suffers 1371 Fire damage from your Ignite.

Now, realistically these last two ticks should have done: (1731+1726)/5+(671/2)=1027
However, they did 344 extra. 344*2 = 688, so that means I got that much bonus damage from somewhere. Logically, I got it from critting at the same time as it ticked, thus adding another ~690 tick which split across the last two.

Total Scorch crit damage: ~10272
Expected Ignite damage: ~4109
Actual Ignite damage: ~4796
Bonus Ignite damage: ~687 - I repeat that I believe this comes from the Ignite tick coinciding with the Scorch crit at 09:18:01

I hope this has some value to the discussion.

Acetyle · 01/15/08, 5:53 AM

Took me a couple of days, but I've just finished reading through this entire thread. This post is basically taking everything I gathered and understood from this thread and additionally detail my own theory as to how things are handled server-side and how they apply to this topic.

First of all though, a (very) technical view on how I think things are handled server-side (basically reversed from the observations ppl have made here, and my knowledge of software and the game itself):
1. Server-side code is multi threaded. There's at least one thread handling all the background mechanics, and separate threads are handing requests as they arrive from clients.
2. Spells are handled by the request handling thread as they arrive, any future consequences are scheduled to be handled by a background processing thread.
3. Instants are handled and resolved by the thread that handles that request, and not by the background processing thread.
4. Effects that are the consequence of another effect are applied by scheduling as well. This means the the handling of a fireball's cast completion actually schedules an ignite to be applied, rather than apply it itself. Same goes for the scorch debuff, as seen by previous combat logs. This causes the delays seen between fireballs applying damage, and the time consequent debuffs are applied. From the previous combat logs is, it seems that the time of the ticks is decided at this stage as well, not when the ignite is removed from queue and applied.

The difference between the handling of a fireblast (solely by the request handling thread) and the handling of a fireball (both, by scheduling) explains how the "rolling" ignites and free ignite ticks take place. It's a basic race condition where a fireball has been received and it's resolution (finished cast) is scheduled to be handled by a background thread.

Example:

Assumptions: both the target and the caster are stationary - flight time is constant, and both flight time and latency are ignored for now.

T0 - start time.
C - fireball cast time. Assumed to be >2s for this example.
R - your response time, meaning how close to the previous fireball cast you manage to queue another one. Assumed to be <1s.
Q - time it takes the server to remove an action from the queue and process it.

T0:          You hit your fireball key. Request is sent to the server. Request received, and it's completion is scheduled to be resolved at (T0 + C).
T0 + C:      Fireball cast is resolved, hit/crit is decided, ignite gets calculated if needed. Let's say it is a crit, ignite gets calculated now, and is queued to be applied.
T0 + C + R:  You start casting another fireball. Request is sent to the server. Request received, and it's completion is scheduled to be resolved at (T0 + 2C + R).
T0 + C + Q:  ignite is applied. (By observation of combat logs posted earlier in this thread)
T0 + C + 2:  First ignite ticks.
T0 + 2C + R: Second fireball cast is resolved, hit/crit is decided, ignite gets calculated if needed. Assuming it is another crit, ignite gets calculated now, with 1 tick still left on the target, and is queued to be applied. 
T0 + 2C + 4: Second ignite tick, also referred to as a "free" ignite. While it was calculated when the last fireball crit, it was not removed from the schedule, and is processed.
T0 + 2C + Q: Second fireball ignite is processed, and is applied to the target. Carrying the payload as calculated on (T0 + 2C + R).

R might be higher than Q, but for simplicity of this example I listed R first, although combat logs suggest it may be much lower.

This theory explains the 100-200ms window you have in order to take advantage of this behavior. To have your ignite "roll" you need consecutive fireballs to finish their cast while an ignite payload is still on the target, but have the server process the actual ignite action after an ignite tick has been processed. This is due to the fact the 2nd ignite ticks are not getting rescheduled, only first ignite ticks are. A 100-200ms time frame is reasonable for such a scenario.

So how does flight time fit into this?
My theory is that when a spell that has a flight time is resolved, the server calculates the time it will take for it to hit the target, and schedules the effect to take place on that time. The flight time actually helps us as it delays the application of the ignite debuff.
(My personal experience shows that shadow bolts hit me before I see the actual bolt reach me, happened to me several times, but I never thought to SS it. I have a normal flying mount. Using an epic flying mount which has a speed very close or even higher than that of the bolt will skew testing IMO)

Conclusions:
Cast time + flight time need both to be adjusted in order to have an ignite "roll". You want the cast to be completed while an ignite is still up on the target, but the flight time + server side queue processing needs to be longer than the time left untill the next ignite ticks.
So cast time needs to be Q + 2 < C < Q +4: It needs to be longer than it takes the first ignite to tick (so as not to reschedule it), but also be shorter than the time it'll take the last ignite to tick (which is not a problem).
Further more, it also needs to be close enough to Q + 4: It needs to "hit" the target Qs before an ignite tick in order for that tick to still take place unaffected by the new hit. Since many of the combat logs we saw here show Q to be ~1, assuming an additional flight time of ~1 second (standing at or close to max range) explains why we need fireball's cast time to be close to 2s.

On the other hand, Q was sometimes shown to be far less than 1s, meaning we'll probably need some addon to calculate it for us (a-la Quartz's latenzy colored cast-bar, maybe even on that cast-bar) so we can reposition ourselves to adjust flight time.
Assuming that server-side load is not subject to spikes and is pretty constant/stable, and that it directly affects Q, an addon may not be needed, since it can probably be calculated once for an entire raid.

Make sense?

Etherealz · 01/15/08, 11:14 AM

Originally Posted by Enquillion

I logged a friend's mage, he was conveniently fire specced. I'll have to tell you - I'm not imagining things. Apologies for not having a fancy combat log, merely copying from EavesDrop, but you should find the following interesting. The test was done, again, with having the G15 spam Scorch on a 2 sec interval.

Text

When realms come back up I'll try and reproduce scorch rolling ignites again - It would be nice if your log went down to milliseconds as it could have something to do with the ignite delay time for you vs me. Possibly I need to look at something higher / lower to achieve the rolling scorch ignite. I suspect it's has a much smaller window.

manly · 01/15/08, 11:27 AM

To be honest, everything points towards needing your fireball to be mid-air while the ignite application is scheduled to be processed, which so far we know nothing more than a vague 'takes 100-200ms time'. It is still hard to tell whether travel time is involved in reproducing the rolling ignite behavior, and is certainly a mess to test.

If anything, the next step is probably writing a mod. Ideally you want it to record the timings involved for when you do get a rolling ignite, and optionally also record the free/lost ignite damage (lost ignite to fireball/fireblast, but I guess you might as well count mobs dying early ?).

Once we know more precisely the scheduling time for ignite to be applied we can better modify the mod so that it would suggest you a proper time to cast your fireballs. Merely keeping track of:

1- your lag in ms
2- time left for the next ignite tick
3- your fireball cast time (incl. cooldowns currently active)

would be awesome to start with.

[DRF]Solmyr · 01/17/08, 3:33 PM

Here's some data that... well, my guild isn't happy about... I pulled and wiped us on our first attempt... =P

07:50'16.312	Solmyr's Scorch hits Teron Gorefiend for 1097 Fire damage
07:50'17.999	Solmyr's Scorch crits Teron Gorefiend for 1703 Fire damage
07:50'19.555	Solmyr's Scorch hits Teron Gorefiend for 1298 Fire damage
07:50'20.817	Solmyr's Ignite dots Teron Gorefiend for 341 Fire damage
07:50'22.196	Solmyr's Fireball crits Teron Gorefiend for 6173 Fire damage
07:50'24.132	Solmyr's Fireball crits Teron Gorefiend for 6133 Fire damage
07:50'25.849	Solmyr's Fireball crits Teron Gorefiend for 6283 Fire damage
07:50'27.680	Solmyr's Fireball crits Teron Gorefiend for 6158 Fire damage
07:50'29.678	Solmyr's Fireball dots Teron Gorefiend for 32 Fire damage
07:50'29.710	Solmyr's Fireball hits Teron Gorefiend for 4023 Fire damage
07:50'30.432	Solmyr's Scorch hits Teron Gorefiend for 1605 Fire damage
07:50'30.443	Solmyr's Ignite dots Teron Gorefiend for 5119 Fire damage
07:50'31.744	Solmyr's Fireball dots Teron Gorefiend for 31 Fire damage
07:50'32.424	Solmyr's Ignite dots Teron Gorefiend for 5120 Fire damage
07:50'32.424	Solmyr's Arcane Missiles hits Teron Gorefiend for 878 Arcane damage
07:50'32.880	Solmyr's Arcane Missiles hits Teron Gorefiend for 659 Arcane damage (190 resisted)
07:50'33.424	Solmyr's Arcane Missiles hits Teron Gorefiend for 878 Arcane damage
07:50'33.753	Solmyr's Fireball dots Teron Gorefiend for 31 Fire damage
07:50'33.753	Solmyr's Arcane Missiles hits Teron Gorefiend for 878 Arcane damage
07:50'34.273	Solmyr's Arcane Missiles crits Teron Gorefiend for 1317 Arcane damage
07:50'34.435	Solmyr's Molten Armor hits Teron Gorefiend for 112 Fire damage

Total Fireball Crit Damage = 24747
Total Expected Fireball Ignite Ticks = 4949.4 (x2)
BUT, I had one tick left on a scorch from applying the debuff = 341

Grand Total Expected Ignite Ticks = 5119.9
Actual Ticks = 5119 and 5120
NO GAIN =(

The fireballs crit ~2 secondsish apart, I stood just out of melee range of Teron. I switched to AM to try and not pull, but it was already too late...

Kaimani · 01/17/08, 5:45 PM

As Manly and others have repeated over and over again, it seems that in order to get the rolling behavior you need to have the ignite tick while your next fireball crit is in the air. The log you've posted shows 4 crits right next to each other with no ignites in between, therefore no rolling would be possible. This is the exact same behavior you would witness while rapidly critting scorches - the ignites are merely added up.

For 'free' damage to actually occur, there would at least have to be ignite ticks between those crits, all other timing issues aside. You were casting too fast.

nephalos666 · 01/17/08, 5:51 PM

That sounds to be about right (in response to the flight speed of direct damage, thrown spells. in this case fireball)

I have simultaneously charged a target as a mage let loose a fireball at the same target and the travel speed of the spell as well as the movement speed of the charge skill are the same. Having flown an epic flyer with a trinket and as a paladin on the PTRs i could probably safely say that the flight speed of thrown spells are somewhere around the range of 280% to 300% normal movement speed, the same as an epic flyer boosted by a trinket or a paladin aura. And while the flight speed in relation to the distance from the target does nothing for your DPS, it would however have an effect on an attempt at a rolling ignite burn. You couldn't try it any closer than probably 20 yards or you wouldn't have enough time to get off a second fireball in the chain to hopefully crit at the same time the ignite ticked.

My guess is (from seeing fire in action on my own mage and having chain critted with fireballs on bosses in old world instances, seeing something akin to an additive or rolling ignite) the optimal range seems to be 35-36 in-game yards. testing the cast time/travel time of my frostbolts against targets at max range, i'll notice it takes about 1-1.5 full seconds for the first bolt to hit the target while during the casting of a second one. this is at 36 yards (using Arctic Reach talents). Problem is, is there any way to come up with a type of RangeMod for your UI that tells you exactly how many yards you're standing from your target? because if the rolling ignite is dependent not only on the cast time but also the flight time, then having a mod that would show you your exact range from the target would help immensely in the attempt at getting a rolling ignite to go for any period of time.

Enquillion · 01/18/08, 4:02 AM

As far as I know, no mod can track the exact range to your target, but you can get a basic idea depending on which spells are in range to be cast. However, I can't see anything bad about being far away, there's only a "too close" scenario here, isn't there? So 36 yards to be able to renew the Scorch would be ideal.

I wouldn't really say it's dependent on a certain flight time - just a minimum flight time, that's what I've gathered at least.