Arcanaville

Your current system uses only a single polynomial expansion for damage within a bounded space (eight maximum team members). By definition it will have a calculable global maxima and minima within that space, and so long as the relationships are all linear, the equations will either have a solution within the space or point to a solution along a boundary condition. For there to be no closed form calculation that solves for this, you would need to change your methodology entirely to something with either nonlinear or non-polynomially expandable expressions.

I'm not sure why you think it would take millions of equations though since all archetypes fundamentally follow the same game mechanical rules, just with different constants. Its worth noting that in spite of the fact that your code contains many different components representing the different archetypes, the entire program expresses just a single fundamental equation for calculating damage: essentially the equation I used above. In fact, when I last looked at damage mitigation discretely there were really only five fundamental equations in total expressed within the entire system (the tohit algorithm, the damage and resistance equation which includes heals, the +maxhealth mechanic, the mechanics of regeneration, and the mechanics of recharge).

For those that aren't aware of the history of Pocket D, Pocket D is the successor to a previous place called Paragon Dance Party or PDP. Paragon Dance Party was a disco-like nightclub that looked like an abbreviated warehouse map with disco balls. It was created to give players a place to socialize at a time where the fad was parking characters in Atlas or in train stations and permanently /em dancing (this was before auto logout). It wasn't heavily used: oddly the devs didn't consider that the reason why people were parking in train stations while dancing was because they wanted people to see them parked in train stations dancing, not because there wasn't a designated place to afk-dance.

Pocket D has other uses and content now, but its also a legacy of a fad that doesn't really exist anymore. Now, when people get the urge to dance, they just do a Lambda.

The train is my canonical noob glitch as well. I'd played MUDs in the past and played around with other people's MMO accounts, but CoH was the first MMO I was basically entirely on my own on. And I soloed exclusively for the first few weeks or so, so I was literally on my own. My first trip to a train station, I waited for the train to arrive, waited for the doors to open, then tried to run into the train. When that failed, I thought I aimed wrong and tried again. And again. And again. Of course I tried to click on the doors, I'm not stupid. Always when they were *open* because that's when you can use doors, right? Right? I also tried both doors, because why not?

The sad part is that I specifically recall being there long enough for several other players to come and go, probably all wondering just why the heck was someone in the train station basically trying to molest the train doors.

Honestly, there is no one place to go. If there was, even the MMO producers would go there and get their answer. Its very hard to determine.

However, one empirical way to determine if the majority of players genuinely believe that an F2P conversion means the kiss of death to an MMO is to see how recent F2P conversions have gone. Does it increase subscriber count, activity in-game, revenues for the company, and most importantly enthusiasm within the MMO community.

People talk. The devs at Paragon Studios collectively probably hear a lot of inside chatter from Turbine, Cryptic, and every other MMO development house out there. If its working or not working, if it seems like things are improving or not improving. In every profession the grapevine is a critical source of information. And if F2P was a disaster everywhere, you'd notice in the tone of MMO developers and producers. If it was wildly successful, you'd see that in the tone of MMO developers and producers as well. If it was somewhere in the middle, you'd hear that also.

Given the fact that industry people don't usually just give interviews randomly, what do you think this interview says about the mood and sentiment about F2P as a model within NCSoft and the MMO development industry as a whole? Jeremy Gaffney probably isn't the best person to ask about what the sentiment is within the MMO player community at large, but he might be a useful proxy to figure out what MMO developers think the sentiment is within the larger MMO player community, which is a useful data point.

Without comparing, contrasting, or even discussing any actual game specifically, both of Turbine's fantasy role playing MMOs that switched from a subscription only model to a sub+F2P model have been stated to have greatly boosted their revenues after the conversion.

Actually, this is the singular character concept - at least in general - that I have on hold from the very beginning. You can play games here and there, but there's fundamentally no way to make Dr. Strange or similar characters in City of Heroes, and not because of the tank-mage aspect. In City of Heroes terms, the kinds of stuff Doc used to do in terms of magical combat in his own titles probably comes closest to Control/Ranged, sort of like a ranged-focused dominator. But with a stronger focus on personal powers and some personal defense and less on pets.

I actually agree the best way to do this would be with a separate archetype like the EATs, primarily because the abilities you want in such an archetype don't map well to the currently existing archetypes. They are more like controlly peacebringers and warshades.

I wonder if the "sorcerer" as you call it would work as an anti-dominator. Dominators have strong control plus domination, and both melee and ranged attacks. They are designed to get into the thick of combat using control as their damage mitigation while they employ their melee attacks (which is not to say a ranged-focused dominator isn't viable). The anti-dominator would be stronger at range and weaker in melee: perhaps it would have more single target and less AoE control, but more ranged protection. Maybe the best starting point would be control set primaries and control-heavy ranged set secondaries. Imagine Mind Control/Ice Blast, for example. But it would have to be modified in interesting ways so its not just a rehash on existing powerset combinations and doesn't steal thunder from Dominators.

And I really hope power customization will one day add more "magical-like" visual effects in general.

Had I designed the incarnate slots either both of the level shifts would be in Interface and Judgement, or one would have been in one of those slots and the other would have been in the opposite side of Lore/Destiny.

The point to the shifts is not to be the reward for completing your run through the trials per se, but rather to allow players to build power usable *in* the trials. They don't work anywhere else. So logically, we should either be unlocking them early so they can be used for at least half our core path through the system, or they should be made available throughout that path, one earlier and one later. Making both of them arrive later satisfies someone's itch to make the later powers stronger than the earlier ones, but fails to account for why level shift was even created in the first place.

In their current configuration they are mainly useful to a player that decides to shoot past rare powers and grind for the very rare powers. But that makes them more of a grinding tool than a progressional tool, and in my opinion that is an error.

Aren't trial accounts currently restricted in terms of things like Global Channels?

True, non-subscribers could clutter up broadcast and local. Suppose we were to add this extra restriction: if you want access to broadcast and local (and maybe some other stuff) you don't have to pay, but you *do* have to either have a credit card on file or alternatively buy some minimum of in-game store currency. That way, the ban hammer has teeth, but players don't have to actually pay anything to use those services. If they buy store currency that is at risk, and if they have a card on file that card itself can be banned limiting the damage a single player could do.

I could imagine five tiers of players:

1. Free to play trial players. Absolutely free, nothing required to join. However, these accounts are more like trial accounts and they won't have access to broadcast, local, or global chat channels. They would be essentially locked out of the in-game economy: no buying or selling from the market, no global mail transfers. Things like supergroups would probably be blocked. And all higher end stuff would be excluded: no alignment merits, no incarnate powers, no epic archetypes, etc. But they could at least run the character creator, roll up a hero or villain, and try the actual gameplay out.

2. Free to play starter players: No subscription necessary, but either credit card must be filed with master account or minimal store currency purchased (doesn't have to be spent). Basic access to most core gameplay including basic chat channels. Tier 1 can be upgraded immediately to tier 2 at any time.

3. Free to play veteran players: No subscription necessary, enhanced gameplay awarded based on a combination of being player in good standing for length of time and by amount purchased from the in-game stores.

4. Basic subscribers: subscription necessary, enhanced gameplay due to subscription. Some elements of the game still ala carte.

5. Veteran subscribers: subscribers in good standing with many veteran badges and/or in combination with ala carte purchases. Increased access to optional gameplay items that are otherwise ala carte for other subscribers and F2P players.

Assuming non-subscribers have access to the forums. I would assume that under an F2P model where people can make an unlimited number of accounts and even a large number of master accounts, this would be highly restricted.

Posters are self-selected advocates and smart alecs. Myself included. If you want to know what the average person thinks about politics, you don't read political bulletin boards. You select people randomly who aren't themselves self-selected for highly skewing personal motivations.

That's what makes MMO boards the *worst* place to take the pulse of the MMO player community at large. You're more likely to get a representative sample of people standing in line at Taco Bell.

That's true, although its tricky to quantify how fast and effective that is since it relies on a team mate moving into an optimal position to affect all or most of the targets.

Random postings in MMO bulletin boards are the *last* place I would go to take the pulse of the MMO gaming community at large. I mean literally the last place, after cemeteries and stadium rest rooms.

Its not like Atari actually released numbers capable of determining whether Cryptic's shipping titles were incrementally profitable, and whether the net losses were due to internal development of new titles.

Oh, wait.

I don't exactly know why, but I know it was deliberately set that way in beta and declared working as intended.

I'll simplify. Based on looking at the code, the assumptions built into the calculations are:

Base Damage/Damage Cap:

Kinetics buffers: 20/4.0
Scrapper: 75/5.0
Brute: 50/8.75
Tank: 52/4.0
Resistance debuffer: 30/4.0

Note: there is a bug in the code for Brutes: all damage caps are expressed in the code as 1 + buff, not as absolute damage strength caps. So Scrappers are listed as 4.0. However, Brutes are listed as 7.75 and they should be listed as 6.75, because their absolute damage cap currently sits at 7.75 or 775%. Not important to Scrapper vs Tanker comparisons though.

Damage buff assumption: +400% damage continuously per Kinetics buffers (fixed), +800% total damage buff continuous
Resistance debuff assumption: -60% resistable debuff continuous per debuffer

Combat modifier assumption: -3 (65% effective strength)

We can then state that the damage output of the team given those numbers is:

(2 * 20 * 4 + A * 30 * 4 + (8 - A - 2) * 75 * 5) * (1 + 0.6 * A) for the team with 2 Kins + A Debuffers and the rest Scrappers. Which is (0.6 * A + 1) * (2410 - 255 * A) = -153A^2 + 1191A + 2410. And then solving for the zero we get -306A + 1191 = 0; A = 1191/306 = 3.89.

The optimum would be closer to 4 than 3 for even con. Against +3, the debuff shifts downward and we get:

(160 + 120A + (6 - A) * 75 * 5) * (1 + 0.39A) = (2410 - 255A) (1 + 0.39A) = -99.45A^2 + 684.9A + 2410. The zero then becomes -198.9 A + 684.9 = 0; A = 3.44.

By a very thin margin 3 debuffers is better than 4 at +3 given these numbers.

I didn't include the 0.65 combat level multiplier because its just a constant multiplier through the equation: it doesn't change the balance points.

Lets see if I can recreate these numbers:

Without a tank and with 3 debuffer(s), total DPS is 2315
With a tank and with 3 debuffer(s), total DPS is 2272
Without a tank and with 4 debuffer(s), total DPS is 2310
With a tank and with 4 debuffer(s), total DPS is 2191

3 debuffers and no tank:
(2410 - 255A) (1 + 0.39A)0.65 = (2410 - 255 * 3)(1 + 0.39 * 3)0.65 = 1645 * 2.17 * 0.65 = 3569.65 * 0.65 = 2320.27.
4 debuffers and no tank:
(2410 - 255A) (1 + 0.39A)0.65 = (2410 - 255 * 4)(1 + 0.39 * 4)0.65 = 1390 * 2.56 * 0.65 = 2312.96.

3 debuffers one tank:
(2 * 20 * 4 + 1 * 52 * 4 + 3 * 30 * 4 + 2 * 75 * 5) * (1 + 0.39 * 3 + 0.2) * 0.65 = (160 + 208 + 360 + 750) * (2.37) * 0.65 = 1478 * 2.37 * 0.65 = 2276.86
4 debuffers one tank:
(2 * 20 * 4 + 1 * 52 * 4 + 4 * 30 * 4 + 1 * 75 * 5) * (1 + 0.39 * 4 + 0.2) * 0.65 = (160 + 208 + 480 + 375) * (2.76) * 0.65 = (1223) * 2.76 * 0.65 = 2194.06.

I'm getting slightly different values because the code defines total_dps as as an unsigned long, which means it is incrementally rounding off all intermediate sums to integers. My calculations above are exact closed form calculations and are a bit more precise as a result.


These calculations presume some very specific values for everything, though. They presume you will always be damage capped regardless of archetype or circumstance, which is not a good assumption even in the strongest steamroller teams. They presume a very high resistance debuff value *per debuffer* of -60% resistance. No one has a -60% resistance debuff: it requires stacking two -30% debuffs to get that high, and I don't think anything can stack that much debuff in less than about 4.5 seconds of total execution time.

I believe the constants in the system presume things simplify as neat as they imply, but in actuality I don't think they do: the effectiveness of the resistance debuffs drops as kill speed rises, because their set up time becomes an increasing percentage of the total, for example. And the damage ratios affect the balance between base damage and resistance debuff. I think if you pick simple numbers you'll get a simple answer, but it might not be an answer that adequately factors in critical complications. And I think that given the large uncertainties in these calculations, a 1.87% difference in team damage has to be called a dead tie, not a calculated loss for the tanker.


I'm pretty sure you could generate a set of closed form equations to figure out, given these values, what the optimum mix of buffers, debuffers, scrappers, tankers, and brutes would be (assuming that's all there is). However, I never got around to installing Mathematica on my new computer and I burned my linear algebra text books upon graduation, so someone else can eigenwhatever those.

That's actually what it was originally conceived to be. Someone somewhere, maybe Castle, maybe someone else, overrode that and added the timer, which not only made it more complex, it also made it twitchy to where you need to queue your next attack to have a guaranteed chance of it going off.

The problem with the 2 second duration is that its a bit misleading. Eagle's Claw has 2.53 second cast time. All of its effects take effect 1.3 seconds after it activates. That means the first 1.23 seconds of the crit chance buff take place *within* Eagle's Claw's cast time. The buff only lingers for 0.77 seconds after Eagle's Claw finishes casting, and you can't even activate another power until an additional 0.24 seconds elapses due to Arcanatime lag, which means you really only have 0.53 seconds to get that next attack off. If you have an attack queued that's not hard, but if you are clicking an attack after Eagle's Claw hits something, you might have 0.25 seconds of lag or more between pushing the button and the server getting the signal. That is very thin timing. Because *no* Martial Arts attack takes less than 0.83 seconds to cast and less than 1.0 seconds of Arcanatime expended, they could always at least increase the buff duration by at least 0.25 seconds with zero chance of the buff stacking onto two attacks. But it would be better to use a mode setting to just crit buff the next attack and only the next attack within a more generous window, like say five seconds.

Its on my list of things to twist Black Scorpion's arm about.

That's not *quite* true. Although in practice this doesn't happen usually, a tanker could build defenses less than a scrapper and end up with the same survivability: if survivability was not an issue the tanker could devote more slots to offense and still end up with the same defensive strength as a scrapper, which should be fine. They could also do this in an alternate build so as not to hurt their ability to tank in any other part of the game.

Its also possible the tank might be able to do certain things in their build that a scrapper couldn't do, even when you attempt to hold as many factors constant as possible, like choosing analogous primary and secondary powersets. For example, a tanker could slot the Perfect Zinger proc into its melee attacks: a scrapper couldn't (a brute could). Given what I've seen people squeeze out of builds, I wouldn't make a blanket statement like that.

That's probably problematic for a number of reasons. It hampers a primary selling point of the game which is customization of characters. Crippling that to just a small fraction of the powersets would be harsh: many powersets serve simultaneously as mechanical differentiators and conceptual touchstones. Blasters for example have weapon sets (AR, Dual Pistols, Archery) and projectile variations (Ice, Fire, Energy, Electrical). It would be very hard to pick two to four to represent the set.

Also, this would funnel all the F2P players into a much tighter set of "flavors." Suppose you're one of the Fire/Fire blaster subscribers out there and you discover 30% of all F2P blasters are Fire/Fire because the devs forced them all to be. Not so good.

I could see designating certain powersets as being special and unlockable - Stone Armor, for example, or Demon Summoning. But they would be the exception not the rule. And I'd be more inclined to introduce new powersets for this purpose rather than restrict most of the pre-existing ones.

The last problematic issue is that by taking powersets away, you are intrinsically reducing impulse shopping opportunities. If AR is locked, you lose a significant amount of the potential to sell unlockable rifles. If War Mace is locked, you lose a significant amount of the potential to sell unlockable maces. You don't want to lock the impulse shopping stands in a part of the store the shoppers don't get to see. You need to give the players X, and hope they accessorize with Y.

Hypothetically speaking:

Q1. Why buy more slots if you can just create more free accounts?
A1. City of Heroes currently, and probably moreso in a hypothetical F2P model, would have account unlockables. Having multiple accounts means more effort to unlock in-game unlockables and having to spend more money to unlock purchasable unlocks.

Q2. What about booster packs?
A2. Booster packs in their current form would almost certainly not escape unscathed in any hypothetical F2P conversion. F2P requires the convenience of an in-game store, and generally requires much more ala carte granularity. Ergo, Booster Packs would become at best a legacy item, replaced with whatever the new ala carte in-game store had to offer.

Q3. What about PvP?
A3. As a general principle, F2P players should not be barred from participating in content you have an explicit interest in encouraging. If the devs wanted to encourage more participation in PvP to revitalize PvP, they would not likely require F2P players to purchase an unlock for it.

Q4. What about Pocket D
A4. Pocket D lies in the general realm of nice to have but not critical to have, with a few noteworthy exceptions. Its probably the sort of service likely to be awarded either through some form of activity unlock or purchasable unlock or both.

Q5. What about Architect/Fort Trident/Crucible/Market?
A5. As a general principle, F2P players should not gain immediate access to problematic services including the markets and the architect. They should be unlockable through purchase though, in nearly all cases.

Q6. What about Incarnates/end game?
A6. Good question. I'm of the opinion that the end game should be at least time gated if not purchase gated. Meaning: no matter how fast a PLer you are, you can't simply jump into the end game in the first week of play. I would think you would either need to purchase an unlock for this, or unlock it through some form of in-game activity which cannot be trivially performed in a short period of time.

That's a good question. Google does. Their primary business model is to give services away for free, presuming that advertising on those services pays for them. Its the network television model, without the televisions (technically, you have to buy computers also). Many open source software companies use a variation of this model: you can use the software for free, but if you want enhanced versions or support you have to pay. Redhat, for example, succeeds on the free to play model for Linux, and supports themselves from the smaller percentage of users willing to pay for enhanced services. That is an even closer analog to the F2P one than Google.

VMware, the eight *million* pound gorilla in the virtualilzation space that makes billions of dollars a year on its software has almost always offered a free version of their product that included many, but not all advanced features. Their investors didn't seem to mind that particular strategy, and EMC, the company that purchased them, doesn't seem to mind either. You can still get a free to play edition of vSphere.

A much closer analog to a service delivery company that is comparable to an MMO would be Skype. Free to play service, with purchasable extras. Their investors did ok with that model, although everyone is wondering what Microsoft was thinking there.

Can I do free to play at McDonalds? Actually, yes I can. McDonalds has free to play contests that anyone can play basically for free. You don't need to even shop at McDonalds to get game pieces: you can just request them via mail. They presume most people will come to McDonalds and buy food, getting the game pieces as essentially a reward for shopping there, but you can in fact play that game for free (minus the cost of postage).

Although it costs me postage to play that game, it also costs McDonalds money to print and send them to me, and they make exactly *nothing* from me when they do so. But they think the benefits of that situation outweigh the costs, or they wouldn't have done it for several decades.

I'd be interested in seeing how those calculations were derived.

You're stating that your teams are doing overlapping contradictory things, so its difficult to say what this scenario is supposed to be: they are steamrolling with concentrated buffs and debuffs, but they are also splitting up a lot, for example. If we assume both are at their respective damage caps we have to assume we're in the scenario where everyone isn't all split up. In that case, the difference would be the difference between 1.125 * 1.1 * 5 = 6.1875 vs 0.8 * 4 * 1.2 = 3.84. The difference would be between 10 hits and 16 hits, which would be about 8-9 seconds on average.

Except that is the case where both are buffed to the damage cap, which implies they are not targeting those alone. The difference in damage would be diluted by the rest of the team being there also shooting at the same target. If even just one other person shoots at each Cyst, things change dramatically. The first one I'm going to assume both players deal the same amount of damage: 6.1875 * 2 = 12.375. In the second, I'm assuming that the other player deals the same amount of damage as the scrapper: (6.1875 + 3.2) * 1.2 = 11.265. Assuming that the two scrapper case drops from 10 attacks to 5, the scrapper + tanker case drops to 5 * 12.375/11.265 = 5.49, or less than one attack.

Why am I comparing two scrappers to one scrapper and one tank? Because the question as posed was why take a tanker over a scrapper. Thus, I substituted one tank for one scrapper. Why assume the *other* player was a scrapper? Mostly because its irrelevant: whatever it is, assuming it deals the same amount of damage as a scrapper the numbers would come out basically the same.

And there is still the question of how everyone is managing to stay simultaneously pegged to the damage cap *and* separated so they aren't all fighting the same things. The problem is that the Tanker is having *all* of its situational disadvantaged summed up simultaneously when they cannot possibly be simultaneously experienced. In situations where its damage cap is lower than scrappers, you're also in highly damage buffed steamroll situations where kill-speed is so high *everyone's* damage is devalued by sheer overkill. When you are doing more than enough damage to defeat things, its impossible to notice a small drop in damage: you are supersaturated. Conversely in situations where kill speed drops to levels where the gap between tankers and scrappers would magnify over longer periods of time, you cannot possibly be, in the general case, permanently in the damage capped situation. These two things don't happen simultaneously except in very weird corner cases.


But then the question becomes why take *anything* except a team of mixed debuffers and maybe one kin? Why does your optimum team have any "DPSers" at all, when resistance debuffs in a full team *always* beat anything that only delivers "personal damage?" You're saying your teams never have a survivability problem, and you build for maximum kill speed. There is no way for any amount of personally generated damage to beat AoE resistance debuffs. That is the nature of the beast. So why bother?

I'm on record as saying if I was in charge, I would probably not take the risk of going F2P: I would expand significantly the current ala carte model exemplified by, but poorly executed within, the booster pack system. I might do something about trial accounts. But I'm aware there is risk involved in such a move.

But having said that, as a player my playing decisions are based purely on value propositions: does the game give me enough value for my money. And as I'm also on record as saying, if the F2P model works, it should be possible to deliver more game to subscribers than they are currently getting now for the same amount of money, given higher net revenue overall. And if the players are as smart as you think they are they will decide based on whether they get more for the same amount of money, not if the game offers even more things for sale than their subscription costs. The latter is not an intelligent perspective.


In other words, suppose it was possible to quantify the amount of work within the average issue to date, and lets say its something like 10 man months of development time. If the F2P conversion ends up giving me 8 man months of content for my sub and the rest is ala carte, I would rightly think I was subsidizing the free players: that would be a worse value proposition than what I'm getting now. On the other hand, if after the F2P conversion I end up getting 12 man months of content, and then *on top of that* there's an additional 3 man months of content that is ala carte that is released in the same interval of time, on the one hand to get *everything* I would have to pay more. But I'm getting more game for what I was paying before, which means I'm better off. If we're talking about the intelligent choice the intelligent choice is to pick the choice that gives the most value for the money. Abandoning an MMO that goes F2P without knowing for certain which situation you're in isn't intelligent: its a knee-jerk reaction without foundation.

I wouldn't say that.


And this contradicts the above.


The idea of the F2P model is that the playerbase as a whole pays more, so the company makes more. But the average player pays the same or less as the typical MMO subscriber for the same level of content. That makes the model individually attractive to individual players, and attractive on the whole for the company.

The trick is an execution trick: to ensure that the relative value you get at the different price points is attractive enough to attract players and convince a high enough percentage of them to pay. While its not an easy balancing act, the target is certainly not extremely difficult to hit because many games seem to hit it successfully.

I find it interesting that the only two possibilities you acknowledge are not possibilities anyone has tried yet.


I'd take that bet. I don't think its realistic to assume that the most likely result of a game going free to play is the game shutting down in eighteen months, when that does not appear to currently be the most likely outcome of that conversion when it actually happens.