Arcanaville

When scrappers argue about performance, they are almost never talking about performance related to anything remotely near where the game's PvE content is balanced around. Most long time Scrapper forum posters understand that.

Its true that because SR focuses almost exclusively on defense, if other sets can soft cap they will then have an advantage over SR which now has less orthogonal protections to beef up. But at that point, we're talking about high end min/maxing, not powerset balancing. Whether every set should present the same ultra-high end opportunties to the min/max player is a separate debate, and one completely disconnected from normal game balance.

If anything, this *might* become an interesting topic in I19 when the Incarnate system launches. Otherwise its mostly just the way Scrappers roll. The Scrapper forum has been talking about min/max performance issues since almost the beginning. Its historically been the focus of many min/max discussions, even occasionally ones that aren't specifically related to Scrappers.

Some of the critters that stand on the streets seem to be AI coded to run if their allies are shot at, before they themselves are even fired upon. Its most common when the group is engaged in some crime, like trying to steal a purse or break into a bus stop for some reason, but it also seems to happen even when its just two or three critters standing around sometimes.

I don't know that. In I11, did you know that the average blaster was half-way to being perpetually debt-capped?

Meaning: half of all blasters were worse off.

And even for my maximally enhanced scrappers, this is not absolutely true. There is still some endurance and health management going on even with both Health and Stamina, when I have both.

Point still being, though, that if players decide to lower downtime with the fitness pool, even if a majority of players choose to do so that doesn't mean the opportunity cost is incorrect. It simply means its an option many players choose to take.

Whether this is something that requires separate action below level 20 when stamina first comes available is at least in part a philosophical question on whether downtime is ever acceptable. If its not, the entire premise of the resource management elements of the game are wrong, and tinkering with Rest would be a trivial gesture in the face of such an incredibly serious design flaw. It also begs the question of the point to sets with downtime mitigators available at lower levels (such as quick recovery). The regeneration set itself was designed to be a low downtime powerset as one of its advantages. That's a meaningless, almost ludicrous statement if the design intent of the game is to minimize player downtime in the first place.

Yes, the devs choose not to balance the game around a build most players don't have.


They don't, won't, and shouldn't. They balance around what players *use*. The devs datamine player performance in the game, and balance around the averages. Those averages do include what the average players slots with, and undoubtedly includes some level of invention enhancements. I guarantee you the average SR player is not running around with 45% defense.

As to:

Oh, I'm very sure the average player doesn't fully understand how defense works, much less what softcapping involves or how it works. Right now I think the 45% number is being thrown around much like the 2:1 rule was in the past, with equally deleterious consequences.

I still get questions from people who think the +56% in their slotting means they have 56% defense.

Yes. Take a soft-capped SR or Shields character, find a Malta mission, and go stand in front of an engineer. Watch how often the spawned turret hits you. This still surprises people when they actually perform this test.

You're talking about how the game is implemented. I'm talking about how the game should be designed. Those are two completely different things.

You have to start somewhere. Let's say you start with the minion health table. You make a spreadsheet and you put numbers for minion health from level 1 to level 55. You might even use an equation to generate those numbers.

But from there, the damage table for a standard player archetype (say, what we define as "1.0 damage modifier") should be computed relative to that table. It should be normalized to that table. You might decide that the damage of a scale 1.0 attack for a 1.0 modifier table should be 25% of minion health at level one and 10% of minion health at level 50, and then compute a formula to generate those numbers on a linear scale.

Then from there you decide that Bosses will start off with three times minion health at level one and rapidly increase to five times minion health at level 20 and then level off to six times minion health at level 55, and then generate equations to generate those tables.

From an initial generated table, the other tables get generated relative to that one, based on how you want the difficulty and other factors to scale in the game. This is the only real way to make reasonable targets and then actually try to hit them. If its just all random numbers, its virtually impossible to enforce any sort of relationship between them, and you get weird jumpy changes in difficulty, odd corner cases in scaling, and even exploitable glitches (like, say, being able to spawn level 1 AVs).

The models have to be normalized to make the relationships between the various numbers obvious and easy to enforce. Those models should then be able to generate all the actual numbers of the game automatically. What you're talking about is how the numbers are presented to the players, which has virtually nothing to do with how the numbers are modeled and engineered in the actual game design.

More power to you if you figure out how to do that.

Not really. Elude is stronger than OTWS. Overload is stronger than Elude except for positional Psi. Hibernate isn't in the same class as either. If we are specifically talking about scrappers, Elude is also stronger than Unstoppable, Power Surge, and Strength of Will (its a little hazier for Unstop and PS for Tankers and Brutes due to their 90% resistance caps).

Elude is actually one of the strongest tier 9s out there. Its main failing these days is for players already building very high performance SR builds with the invention system, its power can become redundant. Keep in mind that Elude supplants toggles and passives, so for a long time (before I6 and the passive resistances were added to SR) Elude was going up against "tier 9s" that were really multiple powers. In the old days, Elude was going up against perma-unstoppable plus invincibility and Instant Healing plus Dull Pain, and competing favorably. Today, Elude doesn't compete against Strength of Will, it competes against the entire Willpower set, and while it probably underperforms the entire Willpower set, that's not because it underperforms Strength of Will itself. Its underperforming SoW plus RTTC + Fast Healing + Mind Over Body plus everything else in the set.

Capped defenses alone is not the best thing you can get anymore in this game. But its one of the strongest single things you can get in this game, before you stack other things in and around it. Sure, if I could trade Elude for Strength of Will on my soft capped SR scrapper, I probably would because I am, in fact softcapped which limits the benefit of Elude. But most SR scrappers are not soft capped, and I wouldn't trade out Elude from the SR set for another tier 9 which would be weaker for most of the other SR players out there.

Its a better emergency power for players that haven't spent time and money soft-capping their SR builds. If all you have is 30%-35% defense, pushing all the way to the soft cap is still pretty sweet for three minutes.

Besides Rularuu watchers, DE quartz eminators have a +100% tohit buff. Turrets and pets have essentially a +25% tohit buff. On the red side tohit buffs are a little more common: some villains have Build Up. Some have tactics (the guards in the bank missions in particular).

The most common place you'll see tohit buffs in the higher end game blue side are the vengeance buffs from Nemesis LTs. You don't want to be running around the docks in PI and kill the two Nemesis LTs that flank the Warhulk before taking him on, because your softcapped defense just went bye-bye (+30% tohit per buff if I remember correctly).

Actually, the health normalized model is the better one from a design perspective. It tells you what's actually happening. Take damage and health. The most important number is the ratio of damage to health, not either number separately. If its 0.25 at level one and smoothly decreases to 0.125 at level fifty, that tells me something about the game design. If all I have are the damage tables and health tables, I have to compute this myself, because those two tables separately tell me absolutely nothing.

Which, by the way, I have spreadsheets that do just that. Its how I actually know that in fact we get offensively weaker as we level, and by exactly how much.

The math is always easier to get correct when designed around normalized standard scales. When it isn't done that way, weird mistakes tend to happen.

The minimum you can reduce tohit to, before accuracy buffs, is 5%. So if you have 90% defense and your attacker has base 50% chance to hit you, you are good up to 45% tohit buffs (which would make the attacker have a net tohit of 95% before your defense, and then 5% counting defense).

Pets and turrets have base 75% chance tohit still, not 50%. So one area Elude could help even a soft-capped character besides PvP is if they find themselves facing off against a lot of those things. If you're in a hyperdense Malta mission, for example, and a bunch of engineers start dropping turrets everywhere, they will seem to pass right through your "softcapped" 45% defense with their 75% base tohit.

However, if you're facing something with a *lot* of tohit buffage, like those DE enimators, then not even Elude will save you unless its heavily slotted. Those generate a +100% tohit buff, which is going to cut right through softcapped defenses, softcap+Elude, and even leak a little through softcap+slotted Elude. They used to make a mess out of perma-elude back in the day.

Most people mean "45% defense" when they talk about soft capping, but technically speaking the defense soft cap is only 45% for things with base 50% chance to hit you. To soft cap a pet or turret, which has base 75% chance to hit, you technically need 70% defense. But most people ignore that situation and refer primarily to the act of building or reaching 45% defense, which corresponds to the base 50% chance to hit that almost everything except pets and turrets (and technically players in PvE) have.


MT_Head:

I think while you're researching this subject, you should also research defense stacking. The defense caps don't apply to any individual power, but to total defense numbers. If you have a bunch of defenses that are all stacking together, the caps apply to the totals, not the individual powers. And they apply to each type of defense separately. "Type" in this case refers not to powers, but to attack (and defense) types: i.e. Melee_attack, Ranged_attack, AoE_attack, Smashing_attack, Lethal_attack, Fire_attack, Cold_attack, Energy_attack, Negative_Energy_attack, Psionic_attack. These are attack types, and (almost) every attack is tagged to be one or more of these. Defense powers buff the corresponding defense numbers that defend against these kinds of attacks. You always get to use your best type against any attack: if you're attacked by an attack that is both Melee_attack and Fire_attack, you get to use the higher of your Melee_attack defenses and your Fire_attack defenses. The higher of the two, not both.

Theoretically speaking, you could be soft-capped to Ranged_attack, but not Melee_attack because your ranged_attack defense was 45% but your Melee_attack defense was only 20%, say.

Not really: the purple patch would see to that.

What I'm talking about is that in a very real sense, health and damage go up mainly for vanity sake, not for any numerical game design reason. Suppose I have an attack that does 10 points of damage, and it does 10 points of damage from level 1 to level 50. And suppose a minion has 40 health at level 1 and that same 40 health at level 50. It takes the same four hits to kill him at level 1 and level 50.

Now suppose that instead of that I make that attack increase in damage from 10 points to 100 points at level 50, and the minion increases to 400 health. Its still taking four swings to kill him: my attacks are still doing 25% of his health per swing. All I've done is moved the decimal place on the numbers: in fact I could lie about it and just *print* the numbers in the combat chat that way, while internally the numbers were still exactly the same.

That doesn't mean a level 1 can go attack and kill a level 50. You could still have level scalers that said when a level 1 attacks a level 50 their attacks are attenuated to only 1% of their normal strength or whatever, so it would still take 400 hits to kill that target.

Its just that this would be psychologically unacceptable to most players. My damage has to go up. And what I attack must get stronger than what I was attacking yesterday. So the game designers have to apply scaling tables to make everything get bigger, in the correct proportions. And as I mentioned, they actually scale health up faster than damage, so while we hit harder, the things we're hitting gain even more health than we gain damage, so in relative terms we actually get slightly weaker as we level.

This has nothing really to do with leveling and fighting different levels, because that can be addressed with combat modifiers. This is more of a case where the numbers the game *shows* to the players have to get bigger to signal increasing strength. They can't just show "hit for 25% of the minion's health" all the time. And if they actually said that and the number got smaller as you leveled, so that by level 50 you were hitting "for 12% of the minion's health" the players would be completely baffled.

The problem is that content isn't balanced that way. Its balanced on the premise of managing health and endurance, not blowing them off in one fight and having nothing left for the next fight so you have to wait around to recover them. The game simply isn't designed to allow people to recover to full easily after every single fight.

Out of combat recovery and rest are touchy subjects for that reason. Health and Endurance are supposed to be managed resources, but often players treat them like dogs would act if you dumped out all their food for the month into their dishes.


The game just isn't balanced around "that which does not kill you doesn't matter." It would take significant amount of work to rebalance around that principle. And regeneration and healing would become significantly less interesting and practically impossible to balance. Those are high costs just to change the entire foundation of the game from its resource management roots to a more per-kill shooter foundation.

If I could work on just one piece of existing tech, it would be the critter AI. I think they could be made a lot smarter, and a lot less frustrating, simultaneously. But my understanding of the critter AI is limited by what I can test for, which limits what sort of suggestions I can make.

My only real contribution there is that critters now make up their minds to shoot you in the face a lot more intelligently than in the past, ironically by making them a whole lot stupider.

I do think that the flee code should add a line of sight check. Every two seconds, the fleeing critter should attempt to target the player or players. If the fleeing critter cannot target any member of the player team with line of sight targeting, the critter should stop running. That's the limit of how far I think a fleeing critter should go: out of sight, so you have to nominally pursue. Not completely out of the zip code.

Mostly AAO, although to be honest I haven't done any calculations based on the new Burn. Between that and the new Fiery Embrace, its possible Fire/Fire is up there especially on AoE damage with Burn.

Simply put, endurance costs are normalized. Damage isn't because of the big numbers principle: players simply want to see bigger numbers as a sign of progress. So level 30s have to do more damage and take more damage, even if the actual number of shots it takes to kill them is similar to that of level 28s. The numbers are a visual cue that you're getting more powerful, even though the enemies are also getting more powerful.

That visual cue isn't necessary for things like endurance. If it were, what would happen is your endurance bar would get bigger, but the endurance costs of your powers would also get bigger as they became more powerful, and we'd be back where we started from.

Its really not that endurance and recharge don't scale, its really that health and damage both are specifically designed to scale upwards, both along similar curves (but with different slopes) to give the illusion of getting stronger. In actual fact, relative to the health increase of enemies, we actually get weaker as we level upward: it takes more shots for a level 50 player to kill a level 50 boss than it does for a level 5 player to kill a level 5 boss.

Its all "pay no attention to the man behind the curtain" aspects of game design that it is better for most players not to investigate too deeply, lest it hamper your ability to enjoy any game, much less this one.

This is often stated as the reason to increase recovery: because most players take stamina (assuming this is true) there's obviously a problem that needs to be rectified by making stamina less "necessary."

That's false. Suppose tomorrow I were to add a passive power that could be slotted to add +50% damage to all your attacks constantly. I'm sure that would be pretty popular also. Question: if as many players took that power as took Stamina, would that prove that damage was too low and needed to be buffed?

No, actually it would prove that there are some things that most players will always want more of, no matter how much you give. And in the example above, the obvious game design statement is that I was an idiot for adding such a power in the first place, and clearly it shouldn't even exist.


Exclusive options are, for the most part, required to have roughly equal value propositions in a well designed game. Non-exclusive options are not required to have that property, and stamina is not strictly speaking a component of an exclusive option (and engineering a specific example to make it appear to be such doesn't actually make it such). Non-exclusive options obey the laws of synergy which allow for certain options out of a group of non-exclusive options to be far more popular than others simply because their benefit is far more general or generally synergistic.

Haha, brain glitch. Big enough one that I'll keep it there to remind myself.


(For those less familiar, Shields cannot be paired with Claws, Katana, Dual Blades, or Spines.)

Of the three sets he mentioned (Dark, Elec, Energy) only Dark has generally faster cast times and rooting times. The other two should be similar or worse than KM in this regard.

I'm afraid you're not going to get a singular consensus answer. Debating DPS is one of the things the Scrapper forums has been doing for years, and for a long enough period of time for most of the old hands to realize that there's no one absolute answer to that question: a lot depends on different variables. Those variables include:

1. At what level of recharge? The answer is somewhat different if you assume no recharge, or assume a small amount of slotted recharge, or assume you will build with Hasten, or finally if you're talking not about a build you intend to level up with but a level 50 build that can use the best inventions. And then the amount you want to spend on those can also change the equation by opening doors to different optimized builds.

Recharge is important because most offensive sets have some really great attacks in terms of doing a lot of damage in a short period of time (DPA), and others not so good. The higher your recharge, the more you can use the better powers and the less you need to use the less concentrated powers. Some powersets start off with a bunch of powers that are decent overall, and do well with any level of recharge. Some start off worse because they have some stinkers, but at higher levels of recharge the stinkers disappear and they are left with nothing but more potent attacks. Dark Melee, for example, can't even build a complete attack chain with zero recharge. But some of its attacks do very high DPA, and a high recharge build can compensate for its lack of attacks by allowing the player to cycle the ones it does have very fast.


2. Single target vs AoE. If you are talking about doing the most damage to kill the most targets as fast as possible when there are lots of targets around, AoE will generally do more damage than single target attacks. The break even is roughly three targets or so. If you can hit more than three or four targets consistently with all your AoEs, your AoEs will generally outperform all single target attacks out there.

But if you're trying to bring down one single tough target like an AV or a pylon, AoEs are usually (not always) rubbish, because they do less damage to a single target in the same amount of time that a single target attack would. If you care about AoE, the rough rule of thumb is pick the set with more of them. In general, three AoEs beat two AoEs no matter how good the two are and how bad the three are. Not always true, but usually true simply because if you hit enough things with them, no AoE is actually bad.


3. Variable damage buffs. Different sets have different ways to self-buff damage, and how you use these powers can have a big impact on your damage output. If you can saturate Soul Drain every time you use it - which requires just happening to have the right number of targets all standing around you at the moment it recharges (not easy to arrange if you're also trying to kill them continuously) Soul Drain is the best self damage buff Scrappers can get. On the other hand, if you cannot consistently do that, its just average. Follow Up is fast enough that it can actually be stacked: with enough recharge you can self buff yourself multiple times. But there are practical limits to that because FU itself has low DPA: there is a slight diminishing return aspect to that. And Power Siphon is even more complex.

Different number crunchers will factor these in different ways, and it can be sufficiently debatable that many people will compare DPS without them, and then add them back in afterwards so that at least the non-buffed DPS numbers can be debated without getting into an unresolvable argument over the best way to account for these powers.


4. Then there's secondaries. It can get doubly complex when you throw in certain buffing secondaries. Shields in particular.


So here's the broad strokes. If all you care about is damage output, and you want that damage over a wide range of situations, Fiery Melee is your first stop. Its bonus DoT makes it a legitimate competitor for among the highest levels of damage across the board, period.

Claws is also a contender at low to medium recharge, because it has intrinsic recharge bonuses and frankly a broken adjustment to its AoE attacks, making it a decent single target performer and a really good AoE performer. It doesn't have quite enough single target DPA oomph to put it at the top with ultrahigh recharge, though.

At high and really high recharge levels, Dark Melee starts to get really good at single target damage, because it has two really good single target performers. Once you get high enough recharge, they start to dominate your attack chain.


Two wild cards from Going Rogue:

Somewhere in the mix is the new MA, but I think its not quite high enough to challenge Fire and Dark Melee at the moment. Its good, and much better than it was, but probably not the best.

Kinetic Melee is a bigger wild card. With an ultrahigh recharge build, and saturated Power Siphon, I don't know if we've seen the full potential of KM yet. But I wouldn't place bets on it until someone has done a top end analysis of the maximum capability of KM. Its also decent at lower recharge, but probably not a top performer. At the highest levels of recharge, though, when Power Siphon can be basically perma? I think its in the mix.

It'll probably take more time and analysis from more people before we converge on a more accepted opinion on these two. But my guess: MA ends up near the top at low to medium recharge, better than average at high recharge but not quite a top contender at high recharge. KM ends up near the top at very high recharge, probably among the top contenders.


Bottom line: if you're looking for the absolute best on-paper level 50 money is no object scrapper damage dealer, you're probably looking at either Dark Melee/Shields or Fiery Melee/Shields. Maybe Claws/Shields for AoE more than single target. Or, if you're going all out on recharge and willing to do the homework yourself, KM/Shields for single target may be a strong play.

I remember when a lot of people made this claim on the CO forums with regard to energy building. Turns out that mechanic is a lot more fun to some players and really boring to others. For me personally it tended to make a mockery of the power activation constraint system, which eventually led to quite honestly much more boring play for me.

So yeah, its easy to say that changing it in this way or that way can only be better, but that's easy to say when you don't like it now and assume no one else could possibly like it either, so there's no downside.

Even I think there's room for improvement in the endurance management system, but I would never put it up to a vote. I'd sooner set it by random number generator.

Then I still don't know what you're referring to. Kinetic Melee's roots seem to be appropriate to their cast times, and KM's cast times are not excessive compared to the average cast times for comparable melee sets. It doesn't root a lot more overall. Certainly not excessively so compared to energy melee or electric melee.

Supposed to, no I don't know. That's a question for Castle. I do know it doesn't: only the initial target is affected by -recovery. All the arcing jumps seem to have energy damage and knockdown, but no -recovery.