Jump to content


Photo

Sweet Informational Thread


  • This topic is locked This topic is locked
1206 replies to this topic

#41 Dustwhisper

Dustwhisper

    Don Flamenco

  • Members
  • 285 posts

Posted 21 November 2007 - 03:17 PM

Think I'm more interested in T6 helm/illidary cowl over getting T5 helm tbh as an upgrade/replacement of Grand engineer. I tend to prefer getting certain stats I need via gems since it's easier to do 1:1 replacements via gems than items where it's nearly impossible.

#42 Kyuki

Kyuki

    Piston Honda

  • Members
  • 114 posts

Posted 21 November 2007 - 03:26 PM

I dont want to sound rude, but if you are at that level of raiding, why would you want the T5 over T6 or Illidari cowl in the first place? :)

If you're not, then well, I hate to repeat myself, but capping hit elsewhere and getting the T5 helm would be a upgrade from the grand engineer.

It's about distributing your stats to gain the most from certain items to get a overall upgrade.

#43 Dustwhisper

Dustwhisper

    Don Flamenco

  • Members
  • 285 posts

Posted 21 November 2007 - 03:36 PM

I dont want to sound rude, but if you are at that level of raiding, why would you want the T5 over T6 or Illidari cowl in the first place? :)

If you're not, then well, I hate to repeat myself, but capping hit elsewhere and getting the T5 helm would be a upgrade from the grand engineer.

It's about distributing your stats to gain the most from certain items to get a overall upgrade.


I'm not too far away from it atleast, 3/5 hyjal, 5/9 BT with 7 bosses in last 8 days =D But still what my point was, I think the gain would be marginal and a waste of time to spend getting to atm. Not to mention it is my beliefe that T5 is better for warlocks to spend their points on than mages.

#44 Kyuki

Kyuki

    Piston Honda

  • Members
  • 114 posts

Posted 21 November 2007 - 03:58 PM

Certainly not worth having your guild going to Vashj just for that reason, that's a waste of time if nothing else, and you can very well wait till you get T6 then ofc :)

#45 Frenzi

Frenzi

    King Hippo

  • Members
  • 705 posts

Posted 21 November 2007 - 03:58 PM

For arcane mages, the cap is 76. For fire and frost mages, that number is 164.


This should be edited to reflect how Elemental Precision effects Frostbolt
Posted Image

#46 Pintofbrew

Pintofbrew

    Bald Bull

  • Members
  • 2,201 posts

Posted 21 November 2007 - 04:13 PM

And on the topic of bugged Ele Pre for frostbolt, I'd like to add that the Cowl loses more and more value if we take the hit out of the equation... 125 hit is cake to get without taking head into account with even the most modest of gemmings given T5 level gear. I appreciate that for a fire spec far from cap Cowl and the ZA head are better, but I maintain many of you are putting down the CSD merely because you prefer the itemization of the item with the meta socket.

#47 Kir

Kir

    Piston Honda

  • Members
  • 199 posts

Posted 21 November 2007 - 05:20 PM

(Not aimed at Ignus) Given that we're all capable of at least plugging some stats into a web-page can we all stop re-iterating the age-old "can someone please post how much +dmg is x haste/hit/crit/croissant?" please. It's getting tiring.


Fine, but this thread is not the TC thread, it's the FAQ for fresh mages coming here basically. Don't get an attitude about 'we've been over this a bajillion times, GOSH!', that's what this thread is meant to help solve.
At least put up the link to lhivera's page, Lhivera's Theorycraft Script, in the first post also, then. It's easier to use then the spreadsheet for most people.
If someone brings up things like this in the TC thread or elsewhere, you point them to this thread.

#48 Nabokov

Nabokov

    Glass Joe

  • Members
  • 3 posts

Posted 21 November 2007 - 05:58 PM

For most specs spell damage and spell haste in fact give almost identical benefit, with haste marginally ahead (1haste~=1.05dmg). This is perfectly reasonable given that it does not improve DPM. Any simulator will demonstrate equivallence including Vontre's Sheet and Lhivera's TC.


I'm working on modelling Spell Haste at the moment, so this caught my attention. I've seen the spell haste ratio quoted above (1 haste ~= 1.05 dmg) a few times on several topics here lately, and want to interject an important caveat that may seem obvious, but may not in fact be obvious to everyone: incremental amounts of spell haste will normally provide no DPS increase whatsoever.

Haste will always give you extra mobility, but haste only adds DPS once you've cast enough that you get an additional nuke. After that first extra nuke, the value of haste oscillates depending on the number of extra nukes you've cast, approaching some maximum limit defined by the average damage of your given nuke.

For example (warning, extremely simplified), if you have 20% haste (a passive 314 haste rating) and your fireballs do an average of 3000 damage, your fifth fireball is an 'extra,' and is effectively worth an additional 600 damage for each of those 5 fireballs (making the spell haste worth about 521 +dmg with empowered fireball). The next fireball you cast brings the effective bonus down to 500 damage per fireball, then 429, and so on until you've cast 10 fireballs, whereupon you're now back up to 600 additional damage per fireball.

If you're only dabbling in spell haste - say you have only 4% (62.8 rating), it's going to take you 25 nukes before you realize any benefit at all. After that, the benefit will dwindle to almost nothing until you cast your 50th nuke.

Hence, the more haste you get, the better, not just because of the raw boost it gives you, but because it's easier to actually obtain a benefit from it in the first place, and more likely that you'll maintain that benefit since the bonus nukes come more frequently.

EDIT: For "DPS" above, read instead "Damage." As others have pointed out, haste does in fact increase DPS in a linear fashion. By lazily referring to DPS, I ended up obfuscating my main point, which is that focusing on the DPS increase provided by spell haste obscures the variability of the benefits it provides at low levels.

#49 Vontre

Vontre

    Mr. Sandman

  • Allied Members
  • 5,458 posts

Posted 21 November 2007 - 06:30 PM

No, that's not an accurate way of modeling the scenario. You will notice that in any scenario with random interruptions, scorch will deal a greater percentage of it's potential damage than fireball (though fireball does more damage overall). Why is this? Because scorch is much faster. Random interruption means it can happen at any time. A longer cast spell means there is more potential time lost due to an interruption. Interruption can occur at any point within the spread of a 3 second interval; the average time lost due to an interruption would be 1.5 seconds, and the maximum possible time loss is 2.99~ seconds. For a scorch, interruption can occur at any point within a 1.5 second interval; the average is .75 seconds, and the maximum possible time loss is 1.49~ seconds. Haste has a very real value here, in that a faster spell is less likely to be interrupted at a bad time. If there are no interruptions, then the spell will almost certainly reach the number of interations required to gain that 'extra' cast.
Raiding is full of challenge. Sometimes there is fire. You have to not be in the fire.

#50 manly

manly

    Soda Popinski

  • Guild Members
  • 3,596 posts

Posted 21 November 2007 - 06:47 PM

Around 900 spell haste rating (unless I failed utterly when I decided to see if I could do it), fireballs become GCD limited. It's not possible to reach that much spell haste rating (passively), so it's always going to give a decent boost. The problem is that spell haste gear has crappy other stats, usually.

I could, of course, be utterly failing at "the math", though.

Only, you need 1570 haste rating to get 1.5s fireballs.

final_spell_cast_time = talented_spell_cast_time / (1 + spell_haste_in_percent)
1.5 = 3 / (1 + x)
x = 1 (ie: 100% spell haste -> 1570 haste)
<Eej> YOU"RE GONNA PULL
<Eej> IF YOU SQUEEZE OFF ANOTHER ARCANE BLAST
<Spectear> You've obviously never played with Manly.
<Spectear> That's hardly a reason to stop DPS.
Very Manly Staff

#51 Beaglemage

Beaglemage

    Glass Joe

  • Members
  • 22 posts

Posted 21 November 2007 - 07:01 PM

Is the rule of Damage > Crit universal? I was (probably mistakenly) under the impression that as a 10/48/3 mage equipping a crit proccing item like TLC, stacking as much crit as possible would be in my best interest (for the ignites and the TLC). Now I am not so sure.

#52 Nabokov

Nabokov

    Glass Joe

  • Members
  • 3 posts

Posted 21 November 2007 - 07:01 PM

Fair enough Vontre - I should have said that I'm primarily trying to model it for Patchwerk-style scenarios, just to get a grip on how much dps it can add in optimal conditions. Pursuing that, I totally overlooked the way it reduced interruption penalties and the real value that grants.

Your last sentence, "If there are no interruptions, then the spell will almost certainly reach the number of interations required to gain that 'extra' cast" more directly concerns what I'm modelling, but doesn't account for the fact that, once you get the extra cast off, its value diminishes until you get another extra cast. However, the longer the fight goes (and therefore the more extra casts you get), the better your average gain will be from haste. When you need to cast, say, 20 nukes (at an incidental 5% passive haste rating) for each extra cast, you're chances of ending the fight close to an optimal time is very low, where 'optimal time' means 'right after an extra cast.' The more haste you have, the better chance you have to end at an optimal time *and* the less it matters when the fight ends, because the more extra casts you get during a given encounter means your average gain rises.

#53 Snowy

Snowy

    Mitt Romney?

  • Moderators
  • 10,255 posts

Posted 21 November 2007 - 07:11 PM

Here's a generic question -- for a new mage, what badge rewards should they be aiming for first? This can dovetail nicely into my extremely casual Alliance mage, who I might play a few times a week. Which ones give the biggest bang for the buck? Is the Icon still the way to go? Which of the new rewards are most appealing?

This is somewhat more relevant for people who enter Karazhan now, since they'll be accumulating badges.

#54 manly

manly

    Soda Popinski

  • Guild Members
  • 3,596 posts

Posted 21 November 2007 - 07:27 PM

Well, it's not exactly a dps gain only in the case of 'once every 25 casts' (illustrative number here). If the fight lasts 2 min, then you might end up with a clipped fireball at the end of the fight without haste, whereas haste gear would have had you won that one.
<Eej> YOU"RE GONNA PULL
<Eej> IF YOU SQUEEZE OFF ANOTHER ARCANE BLAST
<Spectear> You've obviously never played with Manly.
<Spectear> That's hardly a reason to stop DPS.
Very Manly Staff

#55 Doroteasenjk

Doroteasenjk

    Piston Honda

  • Members
  • 117 posts

Posted 21 November 2007 - 07:48 PM

I'm working on modelling Spell Haste at the moment, so this caught my attention. I've seen the spell haste ratio quoted above (1 haste ~= 1.05 dmg) a few times on several topics here lately, and want to interject an important caveat that may seem obvious, but may not in fact be obvious to everyone: incremental amounts of spell haste will normally provide no DPS increase whatsoever.

Haste will always give you extra mobility, but haste only adds DPS once you've cast enough that you get an additional nuke. After that first extra nuke, the value of haste oscillates depending on the number of extra nukes you've cast, approaching some maximum limit defined by the average damage of your given nuke.


You are confusing DAMAGE with DPS. Obviously, the damage increase is a step function, as you correctly point out, but the damage per second scales linearly with the increased haste rating. The caveat, of course, for benefiting from increased haste rating is that lag (brain, communication, computer or pushback) will always be there to interfere.

Assume for ease in my example, that you cast 5 Fireballs over 15 seconds, unhasted, each delivering exactly 2000 damage. The total damage is 10,000 and the DPS in the 15 second interval is 666.66. With 16 point increase in haste (1.01%), you will cast 5 Fireballs in 14.849 seconds. The total damage in the 14.849 second interval is 10,000 (the same), and the DPS in the 14.849 second interval is 673.43.

For that matter, because the instruments themselves don't measure fractions of seconds very reliably, you may not see a damage meter reading that shows that increased DPS unless you get that extra Fireball off in the measured interval. However, this is more a question of a sampling error; the underlying DPS increase is present, but it may take a while for it to register properly.

#56 Nork

Nork

    Bald Bull

  • Members
  • 1,067 posts

Posted 21 November 2007 - 08:16 PM

The followup to that is that on a 14.9 second fight, you'd see a 25% damage boost from that 1% haste increase (you hit with five fireballs vs. four if you were hasteless). Since fight durations are rarely exact multiples of three seconds, you have to count by averages, and a 1% haste increase is 1% more damage on average.

It gets a little more interesting with farming, since it only takes a few seconds to kill something while farming. For the purposes of raiding, though, X% haste is an X% dps boost.

#57 Prod

Prod

    Von Kaiser

  • Members
  • 61 posts

Posted 21 November 2007 - 08:51 PM

There are more questions I see concerning talent specs, such as deep arcane/frost if a 2nd mage puts WC up for you. I've always gone by the rule deep frost is best but you may want to include the math in the OP.

#58 Nabokov

Nabokov

    Glass Joe

  • Members
  • 3 posts

Posted 21 November 2007 - 09:02 PM

Ack, I should have known better than to be so lazy with my terminology - indeed, the main thing I wanted to highlight was the stepwise increase in damage (not DPS). In fact, in retrospect my main point should have been to say that DPS isn't the best metric for thinking about haste when choosing gear. I humbly submit to all the related corrections, including the prospect of an finishing an otherwise-clipped fireball, and the potentially outsized returns from wearing small amounts of haste.

#59 spiderella

spiderella

    My internal monologue has Tourette's Syndrome

  • Members
  • 693 posts

Posted 22 November 2007 - 04:58 PM

Here's a generic question -- for a new mage, what badge rewards should they be aiming for first? This can dovetail nicely into my extremely casual Alliance mage, who I might play a few times a week. Which ones give the biggest bang for the buck? Is the Icon still the way to go? Which of the new rewards are most appealing?

This is somewhat more relevant for people who enter Karazhan now, since they'll be accumulating badges.


Best practices / twinking a mage:

* Icon, off-hand, maybe cloak from badges
* PvP honor gear
* Tailoring gear

can provide 12ish pieces of gear that could last into t6 content.

EDIT: Forgot to mention that Scryer's Bloodgem can be a very quick and easy fix to low hit rating, it doesn't really complement the Icon but that plus the new badge trinket could be a nice combo.

#60 Caryna

Caryna

    is awesome

  • Members
  • 941 posts

Posted 22 November 2007 - 05:01 PM

Here's a generic question -- for a new mage, what badge rewards should they be aiming for first? This can dovetail nicely into my extremely casual Alliance mage, who I might play a few times a week. Which ones give the biggest bang for the buck? Is the Icon still the way to go? Which of the new rewards are most appealing?

This is somewhat more relevant for people who enter Karazhan now, since they'll be accumulating badges.

I'd go for bracers, offhand, wand and the icon of course (which is exactly what I got as soon as 2.3 went live). And start with whatever item is the best upgrade for you.

And if your mage is a tailor then Spellstrike and Spellfire sets are so good you'll have a hard time picking up any item that would break a set bonus on them.
[10:05:49] <Nat> how do u know if a unicorn is a virgin?




0 user(s) are reading this topic

0 members, 0 guests, 0 anonymous users