Learn something new every day. Though, like I said, I had little faith in my own testing since I had only tested for like an hour, so the whole three perfect casts I got were probably crits without me noticing because I had to focus on just casting at all, more than anything. The video is very insightful, too.
Okay, so I ran some numbers (again assuming the mage is using affliction to its best) to show kind of what I mean as far as depreciating effects go.
I took 1/s as a baseline; anything more than that received a reduced multiplier to the damage for that spell. For the first spell per second after 1, I reduced it by X (modifier) and for the second, I reduced it by X-.15. I also applied a multiplier for the first spell per second at 1.5x. e.g. for a 2.5/s mage, I assumed a 3/2/3/2 rotation on casts per second and the casts (and ticks, which always fell on a whole number) were multiplied by 1.5/0.75/0.6 in the X=0.75 case.
To achieve this, I had to break out the timelines into half and third seconds, so the data table is across 5 tabs now and would be difficult for me to put on. Some differences in damage appeared because I was now able to move the selfish boost to the right location (Trip2 doesnāt count towards it, AFAIK), amongst other small details.
The modifier of ā1ā was used to demonstrate the current situation, as I had been doing previously; in this case, the initial multiplier and subsequent mods were ignored.
As you can see, with this system, you can still reward skill but provide a slightly reduced amount. Some tweaking might be necessary to find the best initial damage increase and subsequent damage modifiers, but the point is to demonstrate.
Overall data table
| Class | Modifier | DPS | Average DPS |
|---|---|---|---|
| 0.5 | 0.25 | 1.659338182 | 1.509226415 |
| 1 | 0.25 | 2.644625455 | 2.419455381 |
| 1.5 | 0.25 | 2.854196667 | 2.762363042 |
| 2 | 0.25 | 3.241803333 | 3.059903757 |
| 2.5 | 0.25 | 3.372754667 | 3.286882835 |
| 0.5 | 0.5 | 1.659338182 | 1.509226415 |
| 1 | 0.5 | 2.644625455 | 2.419455381 |
| 1.5 | 0.5 | 2.994473333 | 2.89510552 |
| 2 | 0.5 | 3.538206667 | 3.334878387 |
| 2.5 | 0.5 | 3.870501333 | 3.743268967 |
| 0.5 | 0.75 | 1.659338182 | 1.509226415 |
| 1 | 0.75 | 2.644625455 | 2.419455381 |
| 1.5 | 0.75 | 3.13475 | 3.027847999 |
| 2 | 0.75 | 3.83461 | 3.609853016 |
| 2.5 | 0.75 | 4.368248 | 4.1996551 |
| 0.5 | 1 | 1.106225455 | 1.006150944 |
| 1 | 1 | 1.763083636 | 1.612970254 |
| 1.5 | 1 | 2.370386667 | 2.284050289 |
| 2 | 1 | 3.149213333 | 2.956517937 |
| 2.5 | 1 | 3.974226667 | 3.775383033 |
So i know you can get perfect casts from the shortcut. I was suggesting that the algorithm should be more strict on the upper end and less strict on the lower end.
If you make it so that you have a choice to cast slowly for major damage boosts or quickly for rapid fire, i think it would give some interesting changes to game play. You shouldnāt be able to get perfect casts from shortcutsā¦ Thats just silly. Iām definitely fine with shortcuts being able to successfully cast, but i think there should be some sort of penalty for not doing the āproperā rune (or at least a tradeoff)
They arenāt shortcuts. They are āalternative methodsā 
. They still take time to learn and master, and in some cases they take longer to learn than the āintendedā methods.
4 Likes
that doesnt change my point. there should be a hefty bonus to damage if you spend the time to draw out the runes as intended instead of cutting corners on the spells.
This way you get to choose between being an āorthodoxā spellcaster and a speedcaster
Unless you provide seriously diminishing returns and a massive increase on ācorrectā casts you will never achieve an equality within the realistic realm of cast speeds. As shown in the data, a 1/s caster is like 28% less effective than a 2.5/s caster even with huge modifiers in place. Currently itās like 58%
I completely understand your skepticism, given the damage most scoundrels pull. I took 2 approaches to estimating damage: an in depth damage model and in game testing. The model is the same model that I shared with the devs before the intitial scoundrel nerf, about which Riley said āYour numbers are pretty close to what I was tracking them as internallyā (also referring to mage and ranger numbers). I simply adjusted the model for a newer rotation and changed the basic damage number inputs to account for the nerf. Scoundrel damgage in particular is rather easy to calculate: bullets do a fixed amount of damage, bullets regenerate at a fixed speed, charges take a fixed amount of time, and curves add a fixed multiplier. The only difficult part is calculating the average card damage, given that there are so many possible combinations of card sequences, and finding the optimal rotation. However, cards only make up about 15% of a scoundrelās DPS (calculated with an allowance for choosing suboptimal cards with a given probability), so small errors in the card probabilities are negligible. Depending on which party buffs were applied, the model calculated a maximum scoundrel dps figure for a +4 iceheart of 65-80k (varies with buffs and how the buffs actually stack). You can easily make a similar model once you have the basic scoundrel numbers and theory down; if it becomes an issue you can just calculate card dps as 15% of overall as an estimate.
To ensure that the model was accurate and that buffs stacked properly, I tested actual damage numbers in game. Because I donāt actually play scoundrel, I canāt perfectly execute my rotation without missing any curves on the tiny target dummy. While a boss would be a different story, for testing purposes I simply didnāt use any curve shots or cards to record a base dps number. I calculated the curve buff separately and applied it retroactively. Because I wasnāt sure of how party buffs actually stack in practice (is the ranger buff + the bard buff a 1 + .05 + .05 multiplier, or a 1.05*1.05 multiplier?) I assembled a realistic party of a mage, scoundrel, and bard to apply all the buffs from those classes. When I executed my rotation mostly perfectly, with with no curves and no card usage, I got 39,500 dps over 1.75 minutes (I manually parsed the log and divided my total damage by the total time). Multiplying by the curve bonus and dividing by 0.85 to add the card estimated damage gives a total potential damage of 77,600 dps. I also didnāt use my super, which couldāve added even more dps if used correctly.
While executing everything perfectly in a real fight with curving is clearly a nigh impossible task, good scoundrels should be able to come close with practice. Fixing most of the errors scoundrels make that I listed previously will make a big difference in dps. Discipline and timing are almost as important as aiming for a scoundrel.
The data demonstrates that scoundrels have the potential to beat the current best rangers and even to reach the best mages in sustained damage. The top players in other classes are operating very close to their theoretical maxes (in terms of rotation/theory and speed for mages); scoundrels just havenāt caught up. By no means is the class so far off in damage as everyone seems to want to believe.
Your numbers are close enough that itās probably just a difference from what starting/ending point we chose. However, another key difference is that I took DPS from the middle section of the rotation, rather than just the beginning. In a long fight, itās realistic to assume that affliction will already be applied and about to wear off in order to phase out the initial damage reduction from having to apply the afflictions with none ticking in the background that is overstated by short term calculations.
I want to stress that skill and effort invested doesnāt linearly translate to casting speed improvements. As Archive demonstrated, itās much easier to go from 0.5 to 1 spell per second than it is to go from 1 to 1.5 or especially 1.5 to 2. Iāve only improved my cast speed by at most 0.1 spells per second in the last month or so of regular practicing, because even small improvements are very difficult at this point.
The ability to develop our own spell techniques and ways to play the class is a huge part of what gives mage itās depth. There is a huge variety of runes that are used by the community for the same spells: especially affliction. I know many mages who have made their own interesting techniques for spells that work for them, and sometimes them alone (the donut affliction, the tombstone affliction, and the spiral affliction are some unorthodox examples). The fact that they are all just as effective for people with different playstyles, controllers, ergonomics, and preferences makes the class that much better. If our techniques were punished simply because they werenāt close enough to āwhat they are supposed to beā, even though they work better, the class would be infinitely more boring. Rewarding people with extra DPS at slower cast speeds because theyāre drawing the ārightā rune only crushes the experimental core of learning to master mage.
Nobody is nearly this fast, lol.
1 Like
Yea please do, because as, like, the only high-end scoundrel left in the entire game there is sadly only theory left lol. Though everyone else I see playing scoundrel sucks so hard that it hardly matters if 1 or 2 players can rise up to ranger level. A new or average ranger will always be better than a new or average scoundrel, we tested this countless times because we played with a lot of fresh players, which are choosing these two classes.
The reasons why ātheoryā or standing in front of dummies is not really telling anything I stated over and over, but I am slowly tiredā¦ scoundrels fail because a) no dungeon except perhaps airship is designed for what a scoundrel would even need to curve high ranks, consistently and b) the group is normally so much faster on trash; until your dream-combo popped, your curve is applied or your buff the mob is long gone, that seems to be the major issue, all this card-āpreparationā all the time for nothing.
On bosses itās a bit better, but thereās a hard cap which is actually quickly reached. You can only bring in the cards and combos you get, it is so much more luck involved than skill here; the stack is so very slow that thereās hardly any thinking or choosing involved, a boss is mostly also down before anything is unfolding here; and if you curve rank Vs on a boss, consistently, than this is it. Any good ranger or mage can outdps this.
The next thing is the buff, if you wanna buff the whole group - which is normally better for scoundrel-mains - you are loosing out on personal dmg. So the full basket-dmg (which afaik Scott plays) is only pseudo-dmg which scoundrel mains donāt normally choose, except if thereās more than one scoundrel in their reg groupā¦
Well sadly it is possible to reach the 70k+ like Cam said. I was at the test. The thing is now the curves difficulty, exact timing on shot releases, having all the boosts around you and having tiles that relly on a healerā¦ Its actually a very high skill class to squeeze that damage up. But I still havenāt seen this same test for other classes though. Would also be interesting. Shaman test would be fun xD. I doubt any scoundrel would get that 70k anytime soon though.
(Ps 3 bullet vs 4 bullet the 4 bullet is only really better evn solo if the fights are short like GRINDING and getting that extra delay tile. So go 3 bullet normally anyway)
You can curve up in all the dungeons; donāt limit yourself to curving sideways. The boss rooms are all especially open, with very large targets, which is where most scoundrels are pulling low numbers in the logs.
This is one of the downsides of scoundrel, that the buff sticks to one enemy instead of you. However, curving on the next trash opponent shouldnāt take more than 2 shots and should just give you time to recharge your chamber after emptying it on the last mob. Furthermore, if you coordinate with the group which enemies are attacked by which players, you can have more time to kill something that everyone else isnāt focusing on to make full use of your buff.
As I mentioned, my calculation involved the different probabilities of different card possibilities. The dps from cards is an average number. In general, once you determine which cards need to be prioritized and which cards should buff others, there are many options to almost always get the same important cards buffed. Furthermore, card damage only makes up about 15% of the total damage, so even without using any cards scoundrels can still pull 66,000 dps at maximum.
My damage calculations and tests used the break shot talent, and still got this level of dps. Itās intended to accurately simulate scoundrel mains.
Not at all. I saw atro placing his charts all over the thread, but otherwise saw people mentioning DPS numbers without referencing anything, so I was curious if there was a widely-known and accepted formula that I just wasnāt aware of or something. Thank you for sharing your formula.
Agreed. Even in hallways, I purposely curve around allies (mainly tanks) because a smaller curve bonus is better than no curve bonus. It fixes the body blocking problem, and maximizes my damage, so itās win-win.
Though I also have noticed that you can curve through some walls and ceilings, although youāre effectively doing so blind, you can still potentially get that damage increase.
The best idea I have seen for bringing rangers more in line with what mages and scoundrels are capable of is to reduce the remaining cool down of our special arrows and charged shot by a set amount of time whenever we successfully hit a weak point. This amount of time would need to be at least 1 second but more likely about 1.5 seconds. This would give more damage to the most skilled rangers who can successfully hit all weak points and make the class even more dynamic, fun, and challenging to play. I hope the devs are really looking at this as an option. It would be extremely fun and rewarding to play and would be the best change to the ranger class since weak points were introduced with Reborn.
6 Likes
Oh yea really, are you consistently curving vertically (and RANK V please a crappy I-III will do nothing!) or ever tried? Oh wait, you are not playing scoundrel main. Or scoundrel at all like 99% on forums who are giving tipps on what others all can do.
While there are a ton of mages ingame who put efforts in the class despite the huge skill-threshhold, nearly noone but noobs play scoundrels (and drop them again once they figure they suck), and nearly everyone else theorizes about the class, onlyā¦ which is a joke in itself.
The class misses the ranger buff, first of all, which happend after release to put rangers closer to mages and thus teared these 2 classes dmg apart. This pushed scoundrels necessarily to the bottom of logs, by dev choice.
Then it misses a deck revolving fast enough to do anything like ācard managementā, it misses a useful super completely, still, so why on earth should anyone care to develop fantastic skills (which are not even intended btw) to cope with all this.
As for all these ātestsā as I stated over and over the mobs are down once you got your buffs in, normally once your first curve is placed even. Or you got 4 backs of players around and a wall behind you well gl with vertical curving (which I still need to see btwā¦ not 1 player in the game showed me a vid with consistently vertical ranks Vs, from close-upā¦ and even if so, if heās the only one the class is designed for 1 person huh?).
There are no dummies in dungeons and in live play scoundrels will not ever go up high enough due to all the actions of other classes. We got logs over logs confirming that.
And donāt tell me what cards are doing, I got several hundred hours into that class since the beta and I hate it by now, because thereās just no way to make any sense outta it. Knowing perfect combos is easy, but cards are too slow and too many to match even bossfights lengths. I loved the scoundrel at first, because the movement is better for my joints than ranger and the card-thing sounded interesting, but Iām a healer main and the no-skill shaman is 1000x better as 2nd (dmg) class. So with time I decided to only play scoundrel when I must or to farm. But yea way too many hours and struggles wasted with it to at least say something about it.
Plus we had an influx of ever-new scoundrels, players who started in reborn, in the guild. All of them quit by now, who stayed and successfully developed the class up to do level 10s was rangers, solely.
This alone should be alarming and trigger a need to care for that class already.
And no wonder, it is so much easier than dealing with a broke class to go ranger which is at least a working class or switch to shaman which requires almost no skill and easily tops a rangerās dmg 3 times.
Theory is fine, but if nobody in the game employs it (for whatever reason), then it is proven useless. If nobody plays it or gets good at it, after all the time the class is out by now, there definitly is a problem and itās not fixable by players. If it was weād already see alot of scoundrels performing well, using workarounds like the ominous vertical curve (again, try thatā¦) etc.
PS: @Reyterra feel free to come to a group and test your scoundrel dmg against our rangers (which all got a +4 weapon by now so max gear helps to compare). I have yet to see a good scoundrel, sorry, and people tell all sort of things until they see the logs. Edit: Ok nvm you got no 30 scoundrel and apparently did not one single shard dungeon yetā¦ then see for yourself once you reached the endgame please, this whole thread is about that.
One could say the same of mages, if nobody had ever put the time into it. Then the meta would be solely ranger.
You are correct, my scoundrel isnāt 30 yet and I havenāt run any shard dungeons. Though, Iām far from blitzing the class to 30, since like my 30 runemage (which I honestly only played until about level 5), itāll be pointless if I cant use the class correctly. I do dungeons with it, I do overworld content with it, I grind with it, but I dump the exp elsewhere to level classes that I may play later, but am not concerned with practicing with atm.
Iām significantly more focused on practicing technique than leveling, else - like my runemage - Iāll have a 30 with a skill equivalent of getting off one in three casts of fire 2.
As far as parsing, I believe you, although I havenāt been able to get a good read of anyone (aside from obvious differences) since Iām unsure of how to safely put a parsing program on my quest, since Iām fairly certain the thing doesnāt even have a currently compatible antivirus program on the market yet.
When I make it to that gear level with scoundrel, Iāll drop by. By then, I should (hopefully) have at least 100 hours of practice.
I had a very brief period where I mained scoundrel it is very easy to curve vertically if you have to.
I know that your just defending yourself but donāt get so mad when someone says something that you donāt agree with
Carnage at least has one pretty good scoundrel by the name Truth or veritasium. I recommend asking him for tips if you want to get better at the class
Dummy shots are all good.
But if you have more then one mage in the party in a shard or raid, (even with mage muted) the biggest problem for the scoundrel, at least on my side, is that I can no longer consistently curve shots.
Correct me if Iām wrong, but I believe curv power and release point depend on frame rate. I get easier curves on ultra grafics in my house, but in a raid the same shot gets a different angle, my guess is because of the lower frame count.
Iām basing this of my experience as a fisherman where frame drop counts towards your swing. Iām betting scoundrel has the same system.
ā¦that and I more or less main scoundrel with so terrible damage numbers that Iām swapping to Ranger again
Ya I had the same problem in the raid yesterday with frame dropping making it very hard to curveā¦ But I sadly have to blame it on my pc specs. But if I hear someone with higher pc specs have that problem too then I feel like getting a better pc wonāt change that problemā¦
I think it has to do with the tracking being unable to properly, well, track due to latency. Quest can have a similar problem if you move your head too fast, and you can even lose tracking altogether and get booted back to oculus home; had that happen a few times when I was trying to spin around a lot in dungeons and use my body movements to efficiently curve while moving through the dungeon. No wasted movement kinda thing, ya know? It can be especially bad with mageās tracking, and Iām sure the threshhold for casting is probably a little lower, but still.
Though all of my practice (aside from the very beginning) is on live monsters and dungeon runs (because they scale, so my level doesnāt matter), because practice. And as far as I can tell, the needed curve angle for rank V is ~70Ā° or more, but like Burnator said, lag can lower the rank a bit.
Edit: Also, I pretty much only curve vertically, if that helps too.
Donāt worry, I didnāt exactly do a lot of curving during the dev boss event, mostly crashing.
I donāt really defend myself, that is the pointā¦ I donāt really care about OG players and their e-penis comparisons, but I see one new guy after the next quit over this one class, so I do think something else needs to change but giving them smart tipps.
It is basically two classes now which are easily below tank-dmg when new players drop in with them, mage and scoundrel, now while the mage has a clear perspective to be imba if you are willing to put in the time, the scoundrel does not, so the only advise is stay away from it.
And once again, show me a video with vertical high ranks, IV, V (not I-III ranks that I can do from whatever angle, sure thingā¦ but I-III help zero with the issue, not even the on the line talent kicks in there).
If you state it is all-so-easy show it for once what you mean, or anyone claiming the vertical curves are a thing; I think it is not, because the range & space must normally be so much wider to proc the higher ranks. I really think all that is just airy talk over stuff nobody in the game seriously looked into, yet, and first of all, a mechanics which was never intended for the class in the first place.
(And for all that curving advise I curve IVs and Vs consistently on bosses, most of the time - but even in perfect fights, nice cards, all-Vs, no movement required etc. the dmg is by far not topping anyone who also put hundreds hours into whatever class; and that is really un-rewarding af, because at this point you can change nothing but change the whole class).
Nope, Iām just someone who spent hours figuring out exactly how the class works, optimizing the class, and developing tilesets for the class.
Maybe that explains why the damage numbers are so low? If the only people playing the class are pulling low damage because theyāre noobs, then why does the class need to be buffed? So those noobs can do better than people who play other classes? Balancing isnāt done from the low end up. As Reyterra mentioned, low skill mages do worse than low skill rangers and I think thatās perfectly fine.
In a normal boss fight, you can go through your deck about 5 times (around 90-120 second fight). Fights are often much longer as groups are progressing through harder content. Thatās plenty of time to make decisions about each card and create powerful combos, with probability normalizing itself. If RNG was really a huge factor, then there would also be dps logs where scoundrels did extraordinarily well because they got all the right cards and got lucky.
My testing and math didnāt even include the extra damage boost you can get from pre-setting up cards before the fight (i.e. putting a buffed poison card up and a flame card on your belt before you even begin the fight, to guarantee a good start).
So if everyone fails to put in the same study and work that mages and rangers do, they should get buffed to match those practiced rangers/mages for free? Rangers and mages have had a year and a half of time to hone their class and approach maximum performance.
By not using my curves or cards at all, I pulled 40k dps in game. What tank pulls 40k again?
If a scoundrel approaches the theoretical cap like other classes do they would be substantially rewarded
Whatās so hard to grasp about vertical curving? Itās the same curve mechanic, just not in the horizontal plane. Iām sure not all your curves are perfectly flat; some of your shots probably are angled a little up. Just adjust that so that they are mostly up and youāre curving vertically. From my limited experience testing scoundrel, I find vertical curves to be easier than horizontal ones. They go over things in the way, I donāt have to worry about monitoring my peripherals, and enemies are tall. Hitboxes arenāt square, and most enemies in Orbus are around twice as tall as they are wide. Therefore, thereās a significantly larger cross section of the enemy to aim at and a larger margin of error.
Iām pretty out of practice for scoundrel, as you mentioned:
However, I pulled out my gun and gave some vertical curves a try. I want to point out that the dummy is much smaller and harder to hit than a boss, and that trash usually comes in packs (if you miss you still might hit something else next to it). With practice, Iām sure a scoundrel main could do much better than I did on a variety of enemies.
Regardless: https://youtu.be/CuMTnQmX9As
2 Likes