One hundred trials later...

This is still disputed. Don't risk selling the soul of your firstborn under that assumption. (-:

There isn't really anywhere to go from that. Well, except that I know of one person that told me that after 50 trials without a very rare or rare that they intend to quit because the game isn't treating his participation equally with other players.

Trial rewards shouldn't be lotteries.

Triumph: White Succubus: 50 Ill/Emp/PF Snow Globe: 50 Ice/FF/Ice Strobe: 50 PB Shi Otomi: 50 Ninja/Ninjistu/GW Stalker My other characters

Isn't that what raiding is all about? That's how the 8-million-pound gorilla works.

--NT

I pretty much stopped doing trials because of this crazy crap which isnt fair. This should have never been a random thing ever. Now some one has posted numbers on how screwed up this is will there be more changes to make the system actually fair?

I know someone long ago who told me they were going to quit because they couldn't stand missing in combat and hitting a target shouldn't require luck.

It is what it is.

If trial rewards weren't random I would sit down, calculate how long it will take me to get one I want, get depressed and quit running them. Maybe I'm weird that way but deterministic rewards always end up depressing me. I have to very, very carefully avoid thinking to much about the Empyrean Merit cost of Very Rares since if I do then I start planning how many trial runs I'll need to get T4s at which point I get bored and wonder off.

Instead I'm able to take an attitude of get T3s (which is quick and easy) then run trials when I want to and it would be cool if I got a Very Rare component. If my only method of getting Very Rares was deterministic I'd never want to run a trial ever again (especially if the number of trials to a very rare was the 20+ indicated by Arcana's 5% estimation).

Trial rewards should not be deterministic.

Firstly, it is fair. You may not like it but, if everyone has the same odds, it's fair.
Second, there is randomness THROUGHOUT The game. When you defeat an enemy, it's random *IF* you get a drop and *WHAT* you get.
And, of course, hitting in combat is random.
If you don't want randomness, go play chess. Games like this are built around random results.

Whether a system is 'fair' or not is not based on whether everyone has the same chance. If I wrote a system that randomly deleted 1 character out of a million every month, that would not be 'fair' just because everyone has the same odds.

Randomization just means you are making the results unpredictable. It's the end result of what is being randomized that determines whether it is fair or unfair, not the fact that it is randomized in itself. Every random roll is not the same. It depends entirely on context.

I'm going to leave my personal opinion of whether this system is fair or unfair off the table, but I will say people who it randomly selects to hold back have my sympathies. It is not how I would design a progression system, although as far as I know the original model was to force us into always using Empyraens and it was decided to put a random system on top of that in order to... reward people at random with a get out jail free card? I really don't know, but I don't know much of anything these days about what decisions are made regarding reward systems.

And yet, they're both. Funny that.

The component drop is random. The merits are deterministic. Both can get you where you need to go.

Getting to a reasonable target (rare) doesn't take a whole lot of time to get even if you somehow only get threads for your end of trial reward. The components have a chance to make that go faster, and a chance to let you get to Very Rare. As Arcana has noted repeatedly, Very Rare is a pursuit goal.

Nobody should have any expectation they have a right to it in any particular span of time any more than they have a right to purples IOs in any particular span of time. (Of course, there ARE clearly documented Maximum Effort Paths for both).

Originally Posted by Arcanaville Softcapping an Invuln is fantastic. Softcapping a Willpower is amazing. Softcapping SR is kissing your sister.

Indeed, which is why I have to focus on mentally convincing myself that E Merits cannot be used to buy Very Rares at all lest I compulsively start calculating how many trials I'll need to run, causing me to freak out and stop playing. In fact I'm having to fight the urge right now to calculate that value for inclusion in this very post (bad brain!).

That being said my post should really have said said "Trial rewards should not be solely deterministic" I only used the wording I did to reference Snow's post. In reality I think a mix of deterministic and random works the best (which is what we have). The deterministic portion serves as a minimum progression level for the desperately unlucky while the random portion serves to keep my brain in check.

Which is exactly, precisely what a reward system for this sort of content should be. The randomness offers an element of chance, and the deterministic path offers an absolute floor. You can quibble over the precise level of reward of each path, but I consider it practically axiomatic that any system that offers only one or the other consistently is intrinsically broken.

Ahem. I just ran my last 212 trials (over 4 characters) through Excel. (Edit: I should note that my recorded trials are from around the same time as Leandro's - mid-April.)

Overall:

Very Rare (14) = 6.60%
Rare (40) = 18.87%
Uncommons (94) = 44.34%
Commons (51) = 24.06%
Threads (2) = 0.94%
Failed Trials (11) = 5.19%
Total trials = 212

BAF Trial:

Snow Globe - Ice/FF Controller:
Very Rare (2) = 4.88%
Rare (6) = 14.63%
Uncommon (19) = 46.34%
Common (11) = 26.83%
Threads (0) = 0.00%
Fails (3) = 7.32%
Total trials = 41

White Succubus - Illusion/Empath Controller:
Very Rare (4) = 12.12%
Rare (9) = 27.27%
Uncommon (16) = 48.48%
Common (3) = 9.09%
Threads (0) = 0.00%
Fails (1) = 3.03%
Total trials = 33

Blastbot - Bots/Dark Master Mind:
Very Rare (1) = 5.26%
Rare (4) = 21.05%
Uncommon (9) = 47.37%
Common (4) = 21.05%
Threads (0) = 0.00%
Fails (1) = 5.26%
Total trials = 19

Little Imp - Fire/Fire Tanker:
Very Rare (0) = 0.00%
Rare (3) = 18.75%
Uncommon (8) = 50.00%
Common (5) = 31.25%
Threads (0) = 0.00%
Fails (0) = 0.00%
Total trials = 16

BAF Total:
Very Rare (7) = 6.42%
Rare (22) = 20.18%
Uncommon (52) = 47.71%
Common (23) = 21.10%
Threads (0) = 0.00%
Fails (5) = 4.59%
Total trials = 109

Lambda Trial:

Snow Globe - Ice/FF Controller:
Very Rare (1) = 2.86%
Rare (4) = 11.43%
Uncommon (16) = 45.71%
Common (10) = 28.57%
Threads (1) = 2.86%
Fails (3) = 8.57%
Total trials = 35

White Succubus - Illusion/Empath Controller:
Very Rare (4) = 11.43%
Rare (9) = 25.71%
Uncommon (15) = 42.86%
Common (6) = 17.14%
Threads (0) = 0.00%
Fails (1) = 2.86%
Total trials = 35

Blastbot - Bots/Dark Master Mind:
Very Rare (0) = 0.00%
Rare (3) = 23.08%
Uncommon (4) = 30.77%
Common (4) = 30.77%
Threads (1) = 7.69%
Fails (1) = 7.69%
Total trials = 13

Little Imp - Fire/Fire Tanker:
Very Rare (1) = 7.14%
Rare (2) = 14.29%
Uncommon (5) = 35.71%
Common (6) = 42.86%
Threads (0) = 0.00%
Fails (0) = 0.00%
Total trials = 14

Lambda Total:
Very Rare (6) = 6.19%
Rare (18) = 18.56%
Uncommon (40) = 41.24%
Common (26) = 26.80%
Threads (2) = 2.06%
Fails (5) = 5.15%
Total trials = 97

Keyes Island Trial:

Snow Globe - Ice/FF Controller:
Very Rare (0) = 0.00%
Rare (0) = 0.00%
Uncommon (1) = 50.00%
Common (1) = 50.00%
Threads (0) = 0.00%
Fails (0) = 0.00%
Total trials = 2

White Succubus - Illusion/Empath Controller:
Very Rare (1) = 100.00%
Rare (0) = 0.00%
Uncommon (0) = 0.00%
Common (0) = 0.00%
Threads (0) = 0.00%
Fails (0) = 0.00%
Total trials = 1

Little Imp - Fire/Fire Tanker:
Very Rare (0) = 0.00%
Rare (0) = 0.00%
Uncommon (1) = 33.33%
Common (1) = 33.33%
Threads (0) = 0.00%
Fails (1) = 33.33%
Total trials = 3

Keyes Total:
Very Rare (1) = 16.67%
Rare (0) = 0.00%
Uncommon (2) = 33.33%
Common (2) = 33.33%
Threads (0) = 0.00%
Fails (1) = 16.67%
Total trials = 6

Triumph: White Succubus: 50 Ill/Emp/PF Snow Globe: 50 Ice/FF/Ice Strobe: 50 PB Shi Otomi: 50 Ninja/Ninjistu/GW Stalker My other characters

Players can have the expectation that the disparity between the two is close. Given Oedipus_Tex's results, I don't think the average player will think it is close or fair.

Triumph: White Succubus: 50 Ill/Emp/PF Snow Globe: 50 Ice/FF/Ice Strobe: 50 PB Shi Otomi: 50 Ninja/Ninjistu/GW Stalker My other characters

All I can say is that if it was my decision to make, I would bet the entire game against the supposition that the absolute best and absolute worst case statistically likely scenarios would have to have a very small gap, which implies the randomness factor would have to be trivially small. I don't say that lightly, because I would not only be betting the game, but the future employment of a lot of people on that judgment.

I don't know if you're referring to my iteration results or not. If so, can you explain what you mean by "absolute best and worst cases?" The cases I present are not "absolute," they are the numbers that resulted from 30 runs of the program from the perspective of the player who got low rewards. I wanted to know how often does that happen and from his or her perspective how much further behind are they exactly. I did not evaluate Rares so that leaves an obvious hole in the design. IMO knowing the difference between the highest and lowest rewards and how often they manifest is absolutely critical to understanding how the rewards are actually distributed.

If you are not referring to the numbers I posted, then my apologies.

Normally I don't have to ask this of you, but can you untangle what you have said here? Are you saying that the intent is that the absolute best and absolute worst cases should be very large?

Are you saying that one player getting 30 very rares & rare in a row while another player on the same trials getting only common & uncommon is an acceptable margin? One of the players would say "yes", while the other would say "no".

I would think that kind of margin would be something to be avoided as it would be designed to lose subscribers that feel they are getting the shaft.

Triumph: White Succubus: 50 Ill/Emp/PF Snow Globe: 50 Ice/FF/Ice Strobe: 50 PB Shi Otomi: 50 Ninja/Ninjistu/GW Stalker My other characters

Just wanted to point out I did a few iterations to see if my numbers came up around yours from 199 trials (I ignored threads and failures). Your numbers for the Rares look about right. Here is a fairly typical run of 199 trials with results for 24 players (this time it's a thumbnail instead of a huge pic since I just figured out how to do that):



Basically your results appear to fit pretty easily within the window of 30/40/20/10 guessed at by Leandro. Like others I suspect there is also a downward modifier from the team score. Are there people out there who have done 200 runs and actually have 25+ Very Rares who can post numbers? We'd expect to have at least a few of them running around by now if the percentages are correct.

The players who like randomness are intrinsically willing to take a chance of getting below average rewards for a chance to get above average ones. They know the difference between those two possibilities can be high: that is the point. The players who like deterministic rewards aren't willing to make that trade: they do not want the possibility of getting sub par rewards even if it offers the chance for higher rewards.

I believe in most circumstances, and especially the high reward ones, the reward system should address both desires by having a predictable component that is itself highly uniform, and a random element that has the chance for high variability between the best and worst case. This will satisfy people who want one or the other and are willing to concede the other to the rest of the player population.

Who this will piss off are *both* the people that want all random rewards, and all deterministic ones: the people who aren't satisfied with having a component they can count on to behave the way they want, but cannot tolerate any other part acting contrary to the way they want. If you're asking me if I'm aware I might lose such players, I am.

You're a strong advocate for all deterministic rewards, and you either are aware that will cost customers that find that boring, or you're unaware that there even exist such people in large numbers. I'm giving you the benefit of the doubt that it is the former, which is why I find your question just a little puzzling. It comes down to the fact that I think you are assuming that people advocating randomness just don't understand why their preference is unfair, when its actually that people who prefer randomness have a different definition of fairness. Your definition of fairness is everyone gets the same thing. The other definition of fairness is everyone has the same chance and/or opportunity to get anything, but the person by person results can be different.

We've covered this before, but my design requirements for a system analogous to the trial rewards is that the rewards should have some built-in minimum reward level that meets some earning standard (which might itself be debatable) and it should have some opportunity for luck to play a significant role in what the player can achieve beyond that minimum. The deterministic path satisfies the minimum requirements (or should) and the random drops satisfies the element of chance, and they also present a random and deterministic reward path for people who prefer one or the other (and aren't strictly opposed to the alternate one). Such a system satisfies my requirements for fairness, it satisfies my requirements for being sufficiently appealing to a sufficiently large percentage of the player population, it satisfies by general principles of diversity, and it is consistent with my general philosophy that you should design MMOs for people who want X in some form, and not people who need to avoid Y in all forms: people who want something for everyone and not require everything for anyone (note that none of this addresses whether the specific *numbers* are correct, just the structure of the system).

If this was just my personal preference, it would stop there. But I also believe this is the correct strategy to create a healthy game and appeal to the widest possible audience. Which is why I don't just say its my preference, I also say that if it was my decision to make, that's the call I would make and I would override subordinates to make it without hesitation.

For some context, I mentioned missing. Personally, I think missing is the price you pay for having a game that allows for evasion, a fun mechanic unto itself. You could replace it with an autohit system and pseudo-resistive crap, but you'd lose the mechanic's intrinsic quality if you do so. I'd fight personally to keep evasion in this game, and to add it to any other game. *However*, if I were in Black Scorpion's position, I would *not* apply that personal preference in an overriding nature. If Black Scorpion removed Defense and made all attacks autohit tomorrow, I'd be disappointed, and I would probably at least break one of his less useful fingers, but I would understand: I know my opinion is a minority opinion. I would respect the business decision every second that I was toilet papering his car. And if it was my call to make, I would not override the collective decisions of the other people involved: I would leave that one up for discussion. And if the collective decision was to remove missing, I would allow the group to go that route, and I would only accidentally mail their paychecks to Tahiti once or twice at most. Consensus would be more important here than my own personal preferences, particularly when there is a contrary business interest involved.

So its not enough for me to believe in something strongly enough to actually be willing to fight for it to state that if it was my call then screw everyone else we're doing it my way. But if it was my call to make on the subject of random and deterministic rewards, screw everyone else we're doing it my way. That is how absolutely essential I believe that balance to be.

I actually agree with you about this but I would have gone about it a different way. Actually, exactly the same way merits work right now. Basically:

Everyone gets the deterministic reward. You can choose to hang on to it or spend it for a chance at something better or a chance to lose it.

Right now when you roll well you not only get the main reward, you keep the Empyraens. Because the Empyraens are used for things other than just the salvage but salvage is worthless for anything else, you lose a lot if you end up spending anything on Empyraens.

There is a more optimistic interpretation too though, I suppose. Most of what you buy with Empyraens are random drops of some type anyway. Getting salvage for free is sort of similar to getting a PVP IO drop in PVP or a rare Purple in PVE if thats what the saved Empyraens from a random V Rare end up letting you buy. I would still prefer the merit system's version of opt-in gambling though, which does in fact penalize you heavily for taking the deterministic approach but at least asks you to buy into it first.

For someone who absolutely loves how reward merits allowed you to avoid the market, I find it surprising you completely avoid the fact that you can spend Emp merits to get the components you may need. Assuming you don't fail a trial, you would get a guaranteed 3 Emps a day running each trial once, with whatever drop you got at the end being a potential windfall.

Please remember that the drop rate was adjusted on April 26th. Any drops before then are no good.

The trial reward tables were also adjusted on June 28th (Issue 20.5), so are we to ignore everything before then as well?

June 28:
April 26:
I have not had the thread table since the June 28th patch, just to name one change that affects my results AFTER April 26th.

This is my distribution for the first 180 runs (included in those 212 runs):

If anything, I think I can safely say that I've been getting consistently random rewards. Bottom of the grey sections are a fail, 1st line up is threads, 2nd line is common, 3rd line is uncommon, 4th line is rare, 5th line is very rare. Each of the 4 bars represents 50 trials.

Triumph: White Succubus: 50 Ill/Emp/PF Snow Globe: 50 Ice/FF/Ice Strobe: 50 PB Shi Otomi: 50 Ninja/Ninjistu/GW Stalker My other characters