Hit rating - Page 20

alarge · 04/10/09, 7:00 PM

Originally Posted by Nooska

I am not saying to ignore the tool when the result doesn't suit you.
I am saying that you can only use a tool correctly if you understand its limitations.

The spreadsheet treats 99% hit as every shot doing 99% damage - this doesn't give a result you can use in a real situation.
This is not something that can be remedied as the spreadsheet cannot, under any circumstance, give you any estimation of which shots you will miss on or how that affects the rest of your damage.
The spreadhseet is designed to show you the average dps you would do under the set circumstances - average here being so averaged that you would never see this result.
Crit is handled the same way I believe, it is simply modified into the calculations of each shot as it doing an average extra damage.

As Kraxis shows by anecdotal evidence (which is fine here) being under the hit cap can easily leave you without any misses in a fight. You can just as easily have more misses than your miss percent indicates.

Regarding hit;

Beastmasters loose a lot of dps from the pet if not hit capped - this is consistent, so for beastmastery the spreadsheet will be more accurate - missing an arcane, missing a steady or missing an autoshot will not make a noticable difference.

Marksmen have CS they do not want to miss

Survival have Explosive shot.

MM and SV have 1 big shot that can really screw their dps if they miss - BM has a more averaged final result - this can only be used to infer that the spreadsheets method of averaging hit and crit gives a more true-to-life account of BM than of SV and MM, which will swing more because they have the big shots.

Also, budget wise, Hit never competes with agi or any other pure stat, it competes with crit on most pieces (either hit or crit) like haste and ArP are opposites, so the issue will mot often be; Hit or Crit - haste or ArP.

Hit is more valuable than crit, point for point, untill hit cap, because hit increases the amount of shots that can crit; 100% hit gives you full transferrance of crit - les than that and your crit % is diminished by the missed shots, as only hits can crit (you can't miss a crit).

---

Bottom line, my main point is that consistent damage is better than average damage in any real life situation.

Theorycrafting is good, by in a case where it concerns a game, we must continually keep in mind that the theoretical average doesn't mean much if the factual result is "either or" as long as we cannot predict the actual result in the game.

The tools we use are used not to achieve theoretical perferction, but to increase performance in an actual envronment - and the advice on the tools and on stats must always reflect that people have actual choices to make, and not just theoretical ones on paper.

As I've said before, I will say again; Its is a poor excuse that the spreadsheet shows higher dps if you miss on the crucial abilities to maintain the dps when it matters - and I doubt many raid leaders accept that excuse.
It doesn't matter to the raid or to anyone (including you) if your mega shot hits for 30 more damage if you miss it and so don't deal the damage needed from you. It will always be better if you deal 30 less damage, but can guarantee that you hit when needed.

You are vastly over-simplifying what the spreadsheet does. It doesn't simply apply hit% as an aggregate damage multiplier. As near as I can tell from the Calculations tab, it applies hit% to all the situations where hit applies -- so issues like missing ES and Chimera *are* covered.

Most of your arguments go back to the importance of hit. I'm claiming that the spreadsheet already adequately models the importance of hit. Assigning any *more* importance will lower your dps.

As other posters have indicated, there are no fights currently that require us to successfully hit with a number of shots in a sequence without a miss. Sustained average dps is what we bring to a raid. As a raid leader, I care about damage done and average dps sustained during the fight. I can't think of a single shot (or even a 20-sec segment) where every single person has to hit a "make or break" dps figure. And if there was one, crit and other random factors would cause just as much fluctuation as hit.

When you make the consistency argument, you continually just cover one side of the argument ("what if you miss?"). But the other side is actually more likely ("what if you don't miss?"). Let's say we have two gear choices:

* Choice 1: 100% hit
* Choice 2: 99% hit

[Over-simplified math, but it makes my point]

Let's say that your average dps for choice 1 was 5000. To make choice 2 more attractive than choice 1, your average dps needs to be >5000. For this argument, let's use 5001. Since you are only applying that dps 99% of the time, that means that your _base_ dps would need to be 5001/.99 = 5052 dps.

On average, for any segment of a fight, you doing 52 dps *more* than the guy with 100% hit. Every once in a while, you'll have a segment where you do less. The shorter the segment of time you look at, the less likely a miss is to occur. The longer the segment, the more misses are evened out by the higher base dps.

Now, you say that you are advocating "consistent dps" over "average higher dps". You can make that argument, but after being consistently out-dps'd in every fight of a reasonable length (but maybe not during segment 1:30-1:50 of a fight), you might change your tune. And in any event, that's a personal choice -- one not supported by the data.

As for hit not competing with red stats, I think you've missed a few cases:

* Red slot: 16 agi gem? Or 8 hit + 8 agi?

* +20 hit on hands? or +20 agi?

* Icewalker (+12 hit, +12 cri) on feet? Or +16 agi?

* Elixirs, food, etc.

Perhaps I wasn't fully clear on the 32 agi vs 30 hit case. I said "for me". For my character, in my existing gear, with Grim Toll switched out for an AP trinket (to bring me well below the hit cap). I didn't intend to imply that the same 1.06 ratio would hold true for other cases and certainly not for BM hunters where agi has considerably less value than it does for SV hunters.

Nooska · 04/11/09, 7:24 AM

From what I can tell juggling numbers in the spreadsheet, hit is applied as a damage multiplier.
Yes it covers missing the big shot stoo, but only part of them. To get the averaged out result you would have to fire an immense number of shots.

As a raid officer I also care about damage done. I also care that the damage is done when called on - hence consistently hitting every try is better than hitting slightly harder for some of the tries.

The reason we are stressing the point "what if you miss" is because thats the variable you are introducing when you undergear hit.

To respond to your examples:

Red Slot: Its 1:1 ratio, so hit would be the one to go for - and as you've said, the spreadsheet will agree with me for 99.99% of all cases where the hit gets utilized fully (and even for only missing 1 hit rating if you are BM specced because of the pet inheritance)

Hand enchant: again, hit wins if you are 20 or more under the cap (caveat: BM again)

Boot enchant: This one is more difficult because ts not an easy comparison, but, provided the hit is not wasted, icewalker will probably win; 12 crit rating provides roughly the same crit as ~24 agi, so you are gaining more crit from icewalker, and the comparison becomes 12 hit or 16 ap (with a slight crit reduction compared to the hit)

Elixirs and food are itemized on the same point basis as gems. If you are below the hit cap and the best hit food is not wasted, it will be the right choice - and the higher dps for 99.99% of cases. If you aren' low enough on hit for all the food to be utilized then another food would be the right choice, or regemming / chanting to have the food provide the full benefit.

When gemming, chanting and selfbuffing from food, elixirs and flasks, the first rule regarding hit is to not waste a lot of it, as it is then devalued, that is true.

Bottom line, I think we will have to agree to disagree - even thoug I do follow your point (and you seem to follow mine).
Ca we perhaps also agree that the discussion is moot, as the only likely case where it will come up is if you are hitcapped preraid, and dive directly into second tier of rating, thus skipping a tier of gear? From what I'm looking at the stat increases over 1 tier will never be enough to actually outgear the hit - unless you have a bad piece from t1 and a good piece from t2.

Har · 04/13/09, 2:55 PM

Originally Posted by Pijn

you either choose to go for higher dps by following the spreadsheet (which is still 'just' a mathematical approach to the real deal, you cant make programs to predict the weather on the long term either), or choose to go for better dps (read: consistent dps, knowing you wont miss any shot...ever)

If you really believe this, then you would be trying to trade as much of your crit as possible for AP. Crit just randomizes the amount of damage you do, so by your logic, it's bad DPS.

Honestly, our models properly weigh all stats, regardless of whether they're probability based or guaranteed. If you apply to a guild with a setup that has traded too much spreadsheet DPS to get hit capped, then you'll be in a situation where you have to explain how consistent slightly lower DPS is better than higher average DPS. If that's your decision, go ahead. Just don't tell them you heard someone here tell you to do it.

Shandara · 04/13/09, 4:57 PM

The spreadsheet assumes an infinite length fight when it comes to the effect of hit and crit.

This is necessary because otherwise a full simulation is needed. Fight length is only used to determine the effect of items/talents/abilities with (long) cooldowns.

This is a limitation of the spreadsheet format which has problems addressing specs that rely on procs (i.e. Survival) in specific and, as said, hit/crit in general. For Survival using the Rotation Test is probably more accurate than a pure mathematical frequency-based approach. But if you set the simulation length real short the variation between each iteration becomes bigger and bigger since RNG plays such a big role then.

This is more suited to a project like SimulationCraft.

deckstroy · 04/17/09, 2:59 AM

I can say that being hit capped is the most important thing. I say this from experience. I just joined a new guild where the other 3 or 4 hunters are all better geared than I (most of them have Journey's End and Envoy, some of them have Furyot5Flights,- I do not have any of those : / ) and I always out DPS them on boss fights. Always. I could say I am just a better player than they are, but last week on patchwerk I put out 5.3k DPS and the next hunter was only at 4.8k. They all ask me what I do differently and I tell them- GET HIT CAPPED!!!

Pijn · 04/17/09, 3:37 AM

Originally Posted by deckstroy

I can say that being hit capped is the most important thing. I say this from experience. I just joined a new guild where the other 3 or 4 hunters are all better geared than I (most of them have Journey's End and Envoy, some of them have Furyot5Flights,- I do not have any of those : / ) and I always out DPS them on boss fights. Always. I could say I am just a better player than they are, but last week on patchwerk I put out 5.3k DPS and the next hunter was only at 4.8k. They all ask me what I do differently and I tell them- GET HIT CAPPED!!!

Hehe, there has just been a big discussion about whether thats always true. Even though i realize it isnt always true (GET HIT CAPPED!), starting at a certain gear level, hit starts to overshadow other dps stats like agility. The spreadsheet will tell you where and when this is the case, but i believe at the end of Naxxramas, and now with Ulduar gear, the spreadsheet will always tell a hunter to be hit capped given the nature of hit rating and how it affects our damage. I hardly believe the spreadsheet will tell an Ulduar geared hunter to drop 1% below the hit cap in favor of a better item.

Like the Icecrown example i mentioned before: imagine hit rating being a rare stat in Icecrown, only to be found on a number of items (perhaps a polearm that gives 200 hit rating), where as hit is abundant in Ulduar. You have the choice of staying in Ulduar gear hit capped (which is impressive gear anyway), or get into your new Icecrown upgrades but 1-2% under the hitcap. I really believe the spreadsheet will tell you (any hunter) to stay in the Ulduar gear

(Until you get that polearm)

Har · 04/17/09, 12:34 PM

Originally Posted by Pijn

imagine hit rating being a rare stat in Icecrown, only to be found on a number of items (perhaps a polearm that gives 200 hit rating), where as hit is abundant in Ulduar. You have the choice of staying in Ulduar gear hit capped (which is impressive gear anyway), or get into your new Icecrown upgrades but 1-2% under the hitcap. I really believe the spreadsheet will tell you (any hunter) to stay in the Ulduar gear

(Until you get that polearm)

Nope, the spreadsheet averages out your performance. It will weigh hit rating appropriately heavily, but if the new gear is superior enough without the hit rating, it will suggest that to you.

alarge · 04/17/09, 1:32 PM

Originally Posted by deckstroy

I can say that being hit capped is the most important thing. I say this from experience. I just joined a new guild where the other 3 or 4 hunters are all better geared than I (most of them have Journey's End and Envoy, some of them have Furyot5Flights,- I do not have any of those : / ) and I always out DPS them on boss fights. Always. I could say I am just a better player than they are, but last week on patchwerk I put out 5.3k DPS and the next hunter was only at 4.8k. They all ask me what I do differently and I tell them- GET HIT CAPPED!!!

As several of us have stated, the answer isn't "GET HIT CAPPED!" it is "itemize properly!". Specifically, when you have choices between gear, gems, and enchants, pick the combination that will maximize your average dps. This cannot be accomplished by the application of a simple rule like "GET HIT CAPPED!". It requires some fairly complex calculations that take into account all the major facets of a stat. Many factors go into making the calculations unique for every player (talent choices, current gear, racials, buffs, etc.). Hence, the spreadsheet.

I'll give you an example. That neckpiece you have has a blue slot. You've elected to put a yellow hit gem in that slot, bringing you 6 points over the hit cap but abandoning the +8 AP socket bonus. If you were SV-spec'd, you'd probably get better results by putting a +8hit, +12 stam gem in that socket. You'd be 2 under the hit cap, but you'd have gained 8 AP and 12 stamina. On my sheet at least, 8AP+12 stamina is worth more than 2 hit.

Pijn · 04/17/09, 1:53 PM

Originally Posted by Har

Nope, the spreadsheet averages out your performance. It will weigh hit rating appropriately heavily, but if the new gear is superior enough without the hit rating, it will suggest that to you.

Hah, it was just a prediction, im saying the spreadsheet will say the new gear isn't 'superior enough' to outweigh a loss of 2% hit (Again this is about high end-game gear, the relative nature of hit rating makes it a strong itemrating). Anyway, lets drop this hypothetical case, was just an example of how important hit rating can be.

Nooska · 04/17/09, 1:54 PM

Originally Posted by Har

Nope, the spreadsheet averages out your performance. It will weigh hit rating appropriately heavily, but if the new gear is superior enough without the hit rating, it will suggest that to you.

Actually.. it wont.

At least not when looking at the gear planner - from the experiments I've done with switching gear and looking at how its worked out, the gearplanner can only possibly give you the upgrades if you switch 1 piece of gear, and a single piece will have to outweigh all the loss from losing hit before it becomes a positive delta even if switching all the gear would result in a dps gain (you will never get below 8% miss).

The sentiment is correct but the loss of the first hit rating hurts the most - but again, as has been said, the spreadsheet will likely never show a dps increase by losing hit - except maybe in the case of Beast-tamers Shoulders form Kazrogal, as that adds a flat 3% increase - which can definately outweigh even 1% hit loss (for BM).

(I actually can't remember if there is hit on the BTS)

alarge · 04/17/09, 1:57 PM

Originally Posted by Nooska

Actually.. it wont.

At least not when looking at the gear planner - from the experiments I've done with switching gear and looking at how its worked out, the gearplanner can only possibly give you the upgrades if you switch 1 piece of gear, and a single piece will have to outweigh all the loss from losing hit before it becomes a positive delta even if switching all the gear would result in a dps gain (you will never get below 8% miss).

The sentiment is correct but the loss of the first hit rating hurts the most - but again, as has been said, the spreadsheet will likely never show a dps increase by losing hit - except maybe in the case of Beast-tamers Shoulders form Kazrogal, as that adds a flat 3% increase - which can definately outweigh even 1% hit loss (for BM).

(I actually can't remember if there is hit on the BTS)

I just gave you an example above of a situation where losing hit actually gives you a higher dps.

Are you arguing against the principle? Or are you iterating through all the possible pieces of gear that might force this situation and are unable to find a case where a piece results in a loss of hit, but extra dps?

Nooska · 04/17/09, 2:20 PM

I was responding to Har's statement about t he spreadsheet suggesting gear. Even though it is correct in principle the difference has to be enourmous even though the average would be true for more pieces, because it only weighs each piece individually against what you currently have.

Also while you are correct in your example I have 2 problems with it for the argument you are making:
1) you are not exchanging equal (item)values of stats, and any increase in ap would be worth 2 hit if its replacing an overcpaped number (in your case for a sv spec as BM would have the opposite because of the inheritance) - so you are exchanging 2 hit rating for 12 ap (8 ap + 12/3 stam - or is it 12/4?) - in essence you are exchanging 2 rating points for 12 -the 12 will always be better - and thats without counting the passive stam (health) benefit.
2) Most importantly, the spreadsheet will not suggest any such thing - the spreadsheet doesn't suggest what gems you use, you tell the spreadsheet and it suggests gear - minor technical distinction, but since the subject I was replying to was what the spreadsheeet will suggest / tell you its a point.

Har · 04/17/09, 3:42 PM

I chose my words poorly. If you measure your DPS with two different setups, one of which uses gear that has no hit, but has enough other stats that the overall DPS is higher, I consider that the spreadsheet "suggesting" that setup to you.

An example- eventually I'll want to replace my Grim Toll (with lots of hit on it). I'll only be able to regem a portion of the hit I lose, but if the final loss to my hit rating is outweighed by (for sake of argument) a metric ass-load of agility, I'll take the upgrade. In fact, I'll take any upgrade that promises me more DPS when modeled in the spreadsheet.

Hit is extremely valuable, but not to the point of stupidity. Take whatever gear/talent setup the DPS spreadsheet tells you has the most DPS.

alarge · 04/17/09, 3:43 PM

Originally Posted by Nooska

I was responding to Har's statement about t he spreadsheet suggesting gear. Even though it is correct in principle the difference has to be enourmous even though the average would be true for more pieces, because it only weighs each piece individually against what you currently have.

Also while you are correct in your example I have 2 problems with it for the argument you are making:
1) you are not exchanging equal (item)values of stats, and any increase in ap would be worth 2 hit if its replacing an overcpaped number (in your case for a sv spec as BM would have the opposite because of the inheritance) - so you are exchanging 2 hit rating for 12 ap (8 ap + 12/3 stam - or is it 12/4?) - in essence you are exchanging 2 rating points for 12 -the 12 will always be better - and thats without counting the passive stam (health) benefit.
2) Most importantly, the spreadsheet will not suggest any such thing - the spreadsheet doesn't suggest what gems you use, you tell the spreadsheet and it suggests gear - minor technical distinction, but since the subject I was replying to was what the spreadsheeet will suggest / tell you its a point.

1) Ahhh. But in fact, many gear choices *do* involve unequal number of itemization points. This is particularly true in the case where you are raiding new content. The case where you are comparing equal levels of useful item points is a trivial case. The more interesting version involves unequal cases (e.g., 32 agi or 30 hit?). I have argued that the rule of "always choose hit" only applies well to the trivial case and is inadequate (and even incorrect) for the more general case.

2) The spreadsheet "suggests" in three ways that I use:

[a] The relative stats values found on the overview page
[b] The gear planner
[c] Actually making the changes to your gear/gems/enchants and comparing the resulting dps

Both [a] and [c] can be used to analyze gem choices. Only the gear planner will not give you gem choices (although even then, you can choose different default gem choices to see different effects).

[c] is the most powerful tool in the sheet. This is only way you can actually fully evaluate gear changes, including secondary impacts (e.g., I replace one piece and then also re-enchant another to accommodate the hit change that comes from the first piece).

Nooska · 04/17/09, 4:22 PM

1: yes, they do often involve unequal gear numbers, but noone has ever said that overcapping hit was anything but a watse and lacking 2 hit rating was better than 6 over (Caveat: exception BM due to pet inheritance, better to be 7 over than 1 under).
I am aware of what you have argued, and I still oppose that point, you say that the "always choose hit" only applies to trivial cases and is incorrect in general. This is not true. Hit will always be the better stat IF you are not capped and IF the values are relatively equal all told. Your 32 agi v 30 hit isn't worth anything in item comparison, as hit never replaces agi, it replaces crit (or haste or ArP) - when gemming or enchanting it is a different matter, but then the values aren't equal either and it is a matter of course that the higher value will most often yield the better result.And btw, gear of the same ilvl have the same itemization points spent - it follows the ilvl, so you will always be advised to take the hit gear from the same ilvl until you are capped (or slightly over for BM)

2: neither a or c is anything the spreadsheet suggests - a is based on your current stats and you can't use the calcattributes to replace stats. If you leave the gemslto open and do so, you will get a result, but this is not a matter of the spreadsheet suggesting anything, thats a matter of you using the spreadsheet. C is especially a matter of using the spreadsheet, and not of it suggesting anything.
The only place the spreadsheet suggests anything is in the gear planner. This is the place the calculations are used to make suggestions, and this is what my previous response was directed at- that the way it does it will rarely make a suggestion to replace hit with anything else, unless the level of items is so great as to nor be factual for anyone that needs the spreadsheet to optimize. There is siply not enough difference in Naxx and Uld items for that case to ever be relevant (barring naxx10 -> Uld 25 HD or somesuch).