• Good day, Stranger! — Are you new to our forums?

    Have I seen you here before? To participate in or to create forum discussions, you will need your own forum account. Register your account here!

Sorcerers' Homecoming event

Sir Derf

Adept
Forgot that nuance from last year's implementation.

At 18/20, do you get 2 2 SK options, 2 1 SK options, or random?

The only difference this presents is, if given two 2 SK options to choose from, a Face strategy would see no difference between the two 45's, and so might pick the 45 without instead of the 45 with SK. Otherwise, both strategies will still make the same choices.
 

Hekata

Artisan
Touche. I was distracted for that 12 minute gap before hitting send.
We'll file that under "great minds think alike but some type faster" :D
 

Gargon667

Mentor
Even if you are a player experiencing 'bad luck', shouldn't you still make every effort to maximize your results? Doesn't it make sense to try to have the best bad-luck experience you can get? Unless you are theorizing that the RNG skews further from random when you try to follow the mathematically best courses.

How different is this strategy? Going by the face return means you are undervaluing chests that might return event currency.

Chest (sk)Expected Return (sk/flag)Face Return (sk/flag)Face difference
18 (1)1818
27 (1)1919
45 (sk) (2)19.7022.50-2.80
80 (3)21.6726.67-5.00
89 (3)21.6729.67-8.00
45 (no sk) (2)22.5022.50
54 (2)22.5027.00-4.50
32 (1)23.6032.00-8.40
30 (1)26.2530.00-3.75

If this event operates like the previous version of this did, then we are not being presented with three random chests out of nine, but rather with 1 of 4 1 SK, 1 of 3 2 SK, and 1 of 2 3 SK chests.

Under expected strategy, you would never pick the last 4 (45 (no SK), 54, 32 and 30), as there would always be a 3 SK chest available that is better. Under Face strategy, you would take 45 (no SK) ahead of either 3 SK, and 54 ahead of 89.

50% (32 or 30 as the 1 SK) * 33% (45 (no SK) as the 2 SK) you would take the less efficient 45 (no SK) over either 3 SK, for a reduction of 0.83 return.
50% (32 or 30 as the 1 SK) * 33% (54 as the 2 SK) * 50% (89 as the 3 SK) you would take the less efficient 54 over the 89, for a reduction of 0.83 return.

That's 16.67% chance reduction by 0.83, and 8.33% chance reduction by 0.83, for an overall reduction in return of about 0.2 SK/Flag. That's a less than 1% inefficiency, so small potatoes.

Huh. Didn't expect it to matter that little.


Comparatively, the underestimation being applied by only looking at Face value, means that if one were to look at the results at the end of the event , they would be left thinking that they came out way ahead of what they thought they would get.

Thanks for that! Makes it so much nicer to see what´s going on than my tiny little snippet of paper with my unreadable handwriting in 4 different directions ;)

I guess the main reason is that we don´t get big "upsets" meaning bad face value turned into awesome expected value, the bad ones are made a lot better, but not good enough to make a difference in choice.

But there is one thing that I am missing: why does your 27 beacon have a face value of 19? Shouldn´t it be 27 for face value and only the expected value of 19? If that is so, there will be the big upset that makes the difference between face and expected value.
 

Pauly7

Magus
But there is one thing that I am missing: why does your 27 beacon have a face value of 19? Shouldn´t it be 27 for face value and only the expected value of 19?
Yep. I expect just a typo or a missed formula in the spreadsheet. This is why people keep ignoring the 27 beacon, even though it's the second best. because they see that face value.
 

Sir Derf

Adept
And that's why I show my work...


Chest (sk)Expected Return (sk/flag)Face Return (sk/flag)Face difference
18 (1)1818.00
27 (1)1927.00-8.00
45 (sk) (2)19.7022.50-2.80
80 (3)21.6726.67-5.00
89 (3)21.6729.67-8.00
45 (no sk) (2)22.5022.50
54 (2)22.5027.00-4.50
32 (1)23.6032.00-8.40
30 (1)26.2530.00-3.75

Yeah, that makes a big difference...

25% (27 offered as 1 SK) * 66% (either 45 as 2 SK) you'll pick the worse 2 SK.
25% (27 offered as 1 SK) * 33% (54 as 2 SK) * 50% (80 as 3 SK) you'll pick the worse 3 SK.

That's 16.5% of an average 2.1 loss + 4.1% 2.67 loss.

That's an average 0.46 sk/flag loss.

And 25% (27 offered as 1 SK) * 33% (54 as 2 SK) * 50% (89 as 3 SK), you'll thin there's no difference between the 19 and the 54, and so might take the 54.
 

Gargon667

Mentor
And that's why I show my work...

Glad I could help :)
And even gladder that the expected value comes out ahead after all, I would have been grumpy if face value had just about the same result ;) All the pretty maths for no gain? Not in my town :p
 

Wibbly Woo

Spellcaster
Furthermore, if Elvenar were to use not a classical/purely random RNG system, but rather what I tend to call, in video games, a pseudo-random system, i.e. one in which each failed outcome causes the software to increase the odds of a successful outcome for the next random roll, leading eventually to the certainty of success, then I don't think we would see anywhere near as much discrepancy between individual players' Event reward results. I am no mathematician, though, and there are probably other meanings of the same term (pseudo-random) - or perhaps there is another term for what I mean? Any advice very welcome, since correct English usage is a particular love of mine, and I'd hate to be using the wrong term! :D

I would be wary of using the term pseudo-random the way you are here, as it is commonly used with a very specific meaning in computing to describe a sequence that is statistically indistinguishable to random to some level while in fact being completely deterministic if you know the starting state.

It is also important to remember that very few methods of generating a random number in a computer are truly random. Most are in fact pseudo random generators. I suspect, without knowing anything about how the game is written that they are going to be using one of the standard rand() implementations, because that is the obvious solution and should be sufficient to protect the company from any serious attempts to claim bias in the random generators, which would be a fairly legal problem if players are paying: if the odds don’t match the presented odds that would be fraud!

So yeah, my assumption would be the game is scrupulously “fair” with the random generators, which inevitably means you get outliers on both the lucky and unlucky sides.
 

Twigjam

Spellcaster
Just remembered another thing INNO didn't think trough.

Why do we have tournament quests during the dailys that are on days when there are no tournaments! Inno can you put them on tournament days and make sense of them please?
 

DeletedUser

Guest
Why do we have tournament quests during the dailys that are on days when there are no tournaments! Inno can you put them on tournament days and make sense of them please?
Well, if i remember correctly then you have either regular provinces encounters or Spire ones combined with tournament ones, so you can chose :)
 

Sir Derf

Adept
Yes, you can choose... between a task that is scheduled to be be impossible to perform on the day it is unlocked, and a task that is not. Is that really a choice?
 

DeletedUser

Guest
Yes, you can choose... between a task that is scheduled to be be impossible to perform on the day it is unlocked, and a task that is not. Is that really a choice?
Well, it is a choice because you can wait ;)

I didn't plan well regarding those two so got to fast to the top of the Spire. But i can wait, another Spire starts tomorrow, and the last one with tournament encounters can be done when that starts. So there's only a problem if you're impatient :cool:
 

Bind

Seeker
Very interesting stuff here folks! I have a particular issue that has not been mentioned. For the past week I have only received one quest per day and as a result not getting many scrolls. I am barely hanging on to bronze now. Is anyone else having this problem?
 

CrazyWizard

Shaman
And that's why I show my work...


Chest (sk)Expected Return (sk/flag)Face Return (sk/flag)Face difference
18 (1)1818.00
27 (1)1927.00-8.00
45 (sk) (2)19.7022.50-2.80
80 (3)21.6726.67-5.00
89 (3)21.6729.67-8.00
45 (no sk) (2)22.5022.50
54 (2)22.5027.00-4.50
32 (1)23.6032.00-8.40
30 (1)26.2530.00-3.75

Yeah, that makes a big difference...

25% (27 offered as 1 SK) * 66% (either 45 as 2 SK) you'll pick the worse 2 SK.
25% (27 offered as 1 SK) * 33% (54 as 2 SK) * 50% (80 as 3 SK) you'll pick the worse 3 SK.

That's 16.5% of an average 2.1 loss + 4.1% 2.67 loss.

That's an average 0.46 sk/flag loss.

And 25% (27 offered as 1 SK) * 33% (54 as 2 SK) * 50% (89 as 3 SK), you'll thin there's no difference between the 19 and the 54, and so might take the 54.
Got a bit lucky, did all quests, picked pretty much most city drops, used the mail bonus of 100 currency. no outpost, no ash phoenix, no bought currency (yet)

got 310 staffs now
Opened:
18: 53x
27: 10x (1 return of 80 @ 10%)
30 1x (accident)
45: 35x (with return)(2 returns of 80 @ 7%)
54: 1x
80: 31x (3 returns of 300 @ 5%)
89: 27x (5 returns of 200 @ 12%)

Spot on average on the 27 chest (10% vs 10%)
Slightly unlucky with the 45 chest (5.7% vs 7%)
Lucky with the 80 chests (9.7% vs 5%)
Lucky with the 89 chest (18.5% vs 12%)

Being lucky in the big return chests can make a huge difference as you get a huge chunck of extra currency, obviously it's also risk as if you are unlucky it can be a big blow as well.
but this time with 23 dailies as well it payed off really well.
I knew that if I wanted that portrait, I had to make a gamble reach high or go low
So I picked all 18's, skipped many 27's, and gambled quite a lot on the 80/89 chests, essentially if there was no super cheap 18 or the correct 45 chest I grabbed the 80/89 chests.

Looks like for my main account that wizard avatar is within my reach, maybe they really wanted me to have it ;)
btw I want to have it too :cool:
 

Pauly7

Magus
Why do we have tournament quests during the dailys that are on days when there are no tournaments! Inno can you put them on tournament days and make sense of them please?
There is nothing to suggest that a certain quest belongs on a certain day. Admittedly the closer you get to the end the more likely it is most people will see it today, but certainly if it was a week or so ago a people will be catching up to different quests at different times. There are probably still some people who haven't seen this quest. In my main city I caught up with the quests literally 5 minutes before today's one unlocked. I have another city where I've made it up to about quest 40.
 
So I have just done quest 71 out of 74.
And I just noticed that I got 65 scrolls for it on two of my cities, but only 63 scrolls for it in Arendyll.

Elven Gems of Knowledge site says that I should get 70 scrolls for it.

So why am I getting less for it and different amounts in different cities?
Has this happened to anyone else? A couple of people in my FS in Felyndral noticed it has happened to them too.

Thanks.
 
Top