Strategy to make a 91+ LP genie

Adroit - Lost City

HrunsPanda - Archosaur wrote: »

Ill upload it to my email what i have now. Im pritty sure its correct actually.

Uploaded. I have recalculated and my formulas manually. It works correct really. See edit in previous post for more explanation

The big difference in SP cost from 91-100 and 81-90 also plays a big role. This causes the amount of lvl 80 genies that make it into the succesfull lvl 90 selection to be less important than the distribution of them in this lvl 90 range and thus the amount of genies need to be leveled from 90 to 100

found something good for sharing

I'm having a little trouble following your spreadsheet, can you give me a really brief summary of what/how your calculation works?

soundslegit

CheetahWoods - Raging Tide wrote: »

what if those 2 dots between the 1 and 10 said "OR", as in [1 OR 10]

Your point is that if he had said something other than what he really said he would have been right?

HrunsPanda - Archosaur

Adroit - Lost City wrote: »

I'm having a little trouble following your spreadsheet, can you give me a really brief summary of what/how your calculation works?

Ignore for now all the pheripheral information and go straight to the big table with all the columns. The vertical axis of these collumns is the lucky points. Sorry, the table is way to big for convenience so it is hard to see right now. I will take care of the estethic aspects after the functionality is done. To know what the right numbers are without seeing the scale, just look at the length of the collumn or remember how much you need to substract from the row number.

You see 9 sets of collumns: a bleu unlabeled one, then chance,SP, Mirage, Time, Total.
The bleu unlabeled one is just a flag. It tells if this genie is going to be leveled further or not. For now they are set manually.

Then there is the chance. This calculates the chance that this genie is gonna make it to 91+. The last collumn just grabs the correct number out of the tables that are on the left. The other collumns take the 10 chances for getting 1,2,3...10 luckpoints and multiply these chances with the chances of the genie that you get by adding these luckpoints in the next collumn.

The other collumns calculate the average cost of these genies. They work very similar tothe chance ones. The last collumn only takes the cost to level 90-100 an devides it by the chance of getting to 91+. So if you pay 10m for a try and you have 20% chance, the average cost is gonna be 50m. Again the former collumns take the apporpriate fraction of the costs by adding 10 different luckpoint outcomes and taking the costs from the next collumn. Again it is devided by the chance.

After a good night sleep, i do see however a mistake in how i add the extra SPs for leveling from 80 to 90.(although that does not affect this oddity i spoke about) I can explain what, but then its gonna get really complicated. I think i better not try to share unfinished sheets and try to make sence of it at this stage. Gotta do different things now, get back to it tonight

HrunsPanda - Archosaur

Adroit - Lost City wrote: »

(and just for reference)
91+/100
Average LP value: 93
Average Mirage Cost: 23097
Average SP Cost: 971m (with a standard deviation of 766m SP)

Hmmm, fixed some errors.

23097 / 3 = 7699 .....100/7699 = 0.0129
My sheet sais 0.0128 % with the level criteria you used, so that seems to agree.

Unfortunately my SP cost seems to be off by a factor 60
**** i know whats wrong, mayor haulover needed..... current system works only when no genies are excluded from leveling further.

Edit: fixed ! 730m SP it calculates now. getting there......

And about this strange thing i explained before, they happen at many different levels.
For example, these are the average spirit costs for 17/20, 18/20, 19/20, 20/20 genies.
4,43E+08
3,99E+08
4,52E+08
4,02E+08
As you can see, the 18/20 genie costs on average the least to get to 91 when using this system. Similar results are found for most levels. Usually the cheapest result is gotten from genies a few points below maximum. This is because the ones on the bottom of the range that we keep leveling are the most expensive ones (they are most often a waste of money as they will fail a lot) These mid range genies avoid getting into that area because they need to score medium scores to get there. Their more common high and low scores either get them ditched immeadiately or get them into the favorable range on the next decade.

And if you need 20 seconds to level a genie 10 levels (lag if you click too fast), you are going to be clicking for 72 hours on average to make a lvl 91 genie.

HrunsPanda - Archosaur

Ok, i finished the basics. The sheet can calculate the chances and cost (mirages, SP, time). You can change the target genie level and you can choose which genies to continue or not.

I have added explanations for using the sheet, but not for edditing it.

Right now you have to manually choose which genies to continue or not. I may someday try to add some AI that will show you the most cost effective selections for you. But that is a maybe...
The sheet should be usefull already as it is though.

The server keeps it for 7 days. If you happen to read this and want to download after that, just reply and ill upload it again.

Also consider it open source, freeware or whatever. put it on your site if you want, try to sell it if you find a nuthead whod pay for it. All fine with me

Link

Fissile - Archosaur

There seems to be an error in your implementation of the spirit cost factor for values other than 1. If you put 1,000,000 SP/coin the spirit cost value should reflect the number of millions of SP. Increase or decrease this by factors of 10 and you will see the output doesn't make sense.

Also, it doesn't properly calculate the spirit if you put a 1 in the column for a genie that (a) it is possible to make given the constraints from lower levels and (b) has a 0% chance of reaching the target. This is because you set the spirit cost to 0 for these genies, even though there is a real cost associated with leveling them (i.e. the number of spirit points it cost to level it until you hit a point where it stops leveling).

Anyway I ran it through some tests and the output seemed good, so I made some tweaks to it so I could perform an optimization on it, wrote a macro to do a conjugate gradient search for local minima and then Excel decided to crash while the macro was running so I don't have anything saved! f:weep

But I did get 1 result out of it before Excel kicked the bucket:

Level 10 - 10
Level 20 - 18
Level 30 - 26
Level 40 - 35
Level 50 - 44
Level 60 - 53
Level 70 - 62
Level 80 - 71
Level 90 - 81

Average Spirit Cost: 955m (For a level 105 player)

HrunsPanda - Archosaur

Thanx

There seems to be an error in your implementation of the spirit cost factor for values other than 1. If you put 1,000,000 SP/coin the spirit cost value should reflect the number of millions of SP. Increase or decrease this by factors of 10 and you will see the output doesn't make sense.

I noted that error, did something to fix it, forgot to check if its fixed I suspect it may be something as simple as copying my fix to all the cells. Ill look into it tonight

Also, it doesn't properly calculate the spirit if you put a 1 in the column for a genie that (a) it is possible to make given the constraints from lower levels and (b) has a 0% chance of reaching the target. This is because you set the spirit cost to 0 for these genies, even though there is a real cost associated with leveling them (i.e. the number of spirit points it cost to level it until you hit a point where it stops leveling).

You mean it is not right if you are going to waste SPs on genies that are impossible to get there ? Indeed, it assumes you stop leveling any genies that cant get there, even if it is flagged 1 (which is why i have whole collumns flagged 1, since it is irrelevant)

Anyway I ran it through some tests and the output seemed good, so I made some tweaks to it so I could perform an optimization on it, wrote a macro to do a conjugate gradient search for local minima and then Excel decided to crash while the macro was running so I don't have anything saved! f:weep

Wow, you seem a lot more experienced with excel than i. I was only vaguely thinking of how i can add any AI functionality to it. You already have done things that i dont understand the words you use for. (and i am sure that unlike serbetal you know very well what you are talking about b:chuckle)

But I did get 1 result out of it before Excel kicked the bucket:

Level 10 - 10
Level 20 - 18
Level 30 - 26
Level 40 - 35
Level 50 - 44
Level 60 - 53
Level 70 - 62
Level 80 - 71
Level 90 - 81

Average Spirit Cost: 955m (For a level 105 player)

Arwia - Momaganon

I kinda wanted to read the whole thread before replying but....

serbetel wrote: »

Stochastic is no easy topic and doesnt work like ppl think it does without learning it.

It starts with the chance a number out of [1..10] will be 10 is exactly 1:2
because the possible solutions are only TRUE and FALSE.

This statement is 100 % correct,
while the statement "the chance is 1:10" is non-proofable.

.. is just loony xD

Just because you ask the question "If i pick a number between 1 and 10, will it be 10?" and it has 2 possible answers (true or false), it doesn't mean that true will have a 1:2 possibility. If you randomly chose true or false (1 value out of a set of 2), yes, the probability would be 1:2. But you're randomly choosing a number from out of 10 values and that is BY DEFINITION 1:10.

What's more, as shown before, 10 doesn't even have a 1:10 chance when it comes to genies.

So you just started from a very, very false premise. You attach a true or false answer to your problem (1 out of 2 possible solutions) based upon a 1:10 choice. Your logic is flawed.

P.S. Sorry if you guys already established that, had to say this.

Arwia - Momaganon

So i had this idea...

What if you would try all possible LP thresholds for a level, and see out of them, on average, how many mirages and spirit you need to make a 91+ genie.

What i mean by this is this: for example, for lvl10, you set the threshold at 1/10LP, and you run the script to generate, let's say 5k or 10k genies with 91+ LP, using the rule that you will raise the genie until it becomes impossible to reach 91+ LP. Then you do the same for a threshold of 2/10, then 3/10, and so on.

It will take a fair amount of time to do this, but i think you can make a table with these values, and then using some other tool, you could specify an importance for mirages and spirit, and you would get the most optimum threshold values for each lvl.

There is only one question... Figuring out the x/10 thresholds data would be easy... but how about x/20? Do you start with 2/20 and work your way up? Do you somehow choose a value for the lvl10 threshold and use it to get data for lvl20? Or maybe do all this dynamically somehow, calculating all thresholds when the mirage/sp importance is given, and using the appropriate thresholds? Last option sounds the best, but how long will that take?

Think i'll try to implement this, if laziness doesn't get the best of me. f:strive

Fissile - Archosaur

If you want to minimize the mirage cost, then the answer is simply to level every simgle genie until it can no longer reach your goal. There is no way to use fewer genies than that. The problem with that is on average it would cost 100,000 MILLION spirit or more to get a 91+ genie which is obviously a waste.

The goal here was to come up with a strategy for creating genies within a reasonable amount of spirit and coin cost.

HrunsPanda - Archosaur

And its updated, now almost guaranteed bugfree

Kokki - Harshlands

Have you conisdered buying 30/30 genies for lets say 500K each to decrease time and SP cost?

I'm pretty sure alot of people would sell you 30/30 genies if you world chat for them.

I'm not the best at coming up with these types of things, but once i finally see what it is about i will understand how to use the sheet properly, so what is the converted coin cost ( SP coin cost+ Time coin cost + Mirage coin cost ) to get 1 30/30 LP genie if someone would try to make 1? ofcourse starting out with lets say 50 30/30 genies would significantly increase your chances to get a 91/100 genie with your amount of spirit that you have.

Hmm actually now i think about it, maybe it's actually best to reverse engineer some of the stuff that you come up with, with your simulations.

For example, on average, how many 30/30 LP genies do you need to create 1 91+ LP genie? if you do the same thing for all the milestone you might come up with a better converted coin cost then simply only spending your own SP.

Converted coin cost = MirageValue* Mirages used + SPValue*SP used + Time value* Time used

HrunsPanda - Archosaur

0,34% chance your 30/30 will make it to 91 when using the restrictions used by those who did the simulations. This 300 genies, if you pay 500k each, that makes 150m.

2,85E+08 Spirit needed on average to get your 30/30 to 91.

3,57 hour needed on average to click and level all those genies.

Unfortunately my sheet cannot calculate the cost to get from lvl 1 to 30/30.

Kokki - Harshlands

HrunsPanda - Archosaur wrote: »

0,34% chance your 30/30 will make it to 91 when using the restrictions used by those who did the simulations. This 300 genies, if you pay 500k each, that makes 150m.

2,85E+08 Spirit needed on average to get your 30/30 to 91.

3,57 hour needed on average to click and level all those genies.

Unfortunately my sheet cannot calculate the cost to get from lvl 1 to 30/30.

First off i would like to add that i'm not a mathmatician by any means, i understand most if explained in normal number, so do i understand correctly that with 2,85E+08 you mean 285,000,000 spirit?

i'm trying to set up a formula to calculate SP cost, time cost and mirage cost to get 1 30/30 genie but it will take me some time to figure that out correctly. Once the formula is set up it is possible to calculate per milestone how much SP you need to get to your treshold that you set in your simulations.

i'm pretty sure with this formula combined with that sheet and some smart buying you will be gauranteed to get a 91+ LP genie with your spirit.

HrunsPanda - Archosaur

First off i would like to add that i'm not a mathmatician by any means, i understand most if explained in normal number, so do i understand correctly that with 2,85E+08 you mean 285,000,000 spirit?

yes, thats it.

i'm trying to set up a formula to calculate SP cost, time cost and mirage cost to get 1 30/30 genie but it will take me some time to figure that out correctly. Once the formula is set up it is possible to calculate per milestone how much SP you need to get to your treshold that you set in your simulations.

Maybe you could modify my sheet to make it work for other genies than lvl 100 ones Its open source, free ware and stuff like that


i'm pretty sure with this formula combined with that sheet and some smart buying you will be gauranteed to get a 91+ LP genie with your spirit.

PS, to modify my sheet for this purpose note that the last collumn uses a different forumula than the others. To make it work for different level genies, you should have both formulas in each collumn with an IF statement to check wich formula to use. (always use the one that is now in the lvl 90 collumn for the last collumn that applies) I might do this myself tonight or tomorrow if i feel like it.

Kokki - Harshlands

Added a sheet to it and named it genie cost that calculates converted coin for every combination of LP genie that you want.

Uploaded it genie costs 2.0.xls

i used asterelle's Genie rarity index. I've calculated the perfect genies and held my findings against that index and the margins i found where less then 1% off so the converted coin cost per genie is pretty accurate.

Not 100% sure yet to effectively combine it with your sheet but that's something i might come up with in a later stage if someone still needs it that refined by then

By changing the values in the first sheet in the blue box, your converted coin cost will change also. So for example you think your time is more valued, or your Spirit is less valued you can change it and it will change the coin cost in the second sheet.

If you find any flaws in it let me know and i'll change it, or just change it yourself

At this time my thought would be to incoporate this sheet into your sheet to come up with a converted coin cost to get to a certain LP genie, and then how much more converted coin cost you need to make it a 91+/100

The most important part of all this is how much value everyone puts for their time and spirit and that's something noone else can

HrunsPanda - Archosaur

Well it doesnt account for different strategies than just leveling every genie it seems. The cost is going to be different if you decide not to level for example 11/20 genies any further. That is kind of what this whole tread is about.

Kokki - Harshlands

HrunsPanda - Archosaur wrote: »

Well it doesnt account for different strategies than just leveling every genie it seems. The cost is going to be different if you decide not to level for example 11/20 genies any further. That is kind of what this whole tread is about.

I disagree

I'm not pointing out a specific strategy, but creating a valuta for a certain LP genie makes it easier to determine if that type of strategy is worth it to you.

for every treshold that you set, you now know what it will cost you to get x amount of genies to that treshold in order to get a 91+/100 LP genie.

if time is valuable to you, you want to buy alot of 30/30 genies for a reasonable amount of coins to save you ALOT of time of leveling up those genies. This factor is important in creating a strategy.

I'll give you an example of how to use this.

You take a treshold at level 30 of 27/30 which would have a converted coin cost of "x"
You take the converted coin cost that is still needed to create a 91+ LP genie which would be "Y"

X + Y = Z

Now you can start trying to set different tresholds at different points to see if they have a huge impact on "Z"

Combining our sheets will result in a perfect calculation on what path to take up to 91+

p.s. Also i discovered a flaw in my formula and updated the sheet and link, nothing crazy but i calculated the time value wrong.

Saeidon - Heavens Tear

serbetel wrote: »

No because 1/10 has nothing to do with following numbers,
its just the start sequence of a potentially valid combination.

We dont even are able to ditch it instantly, it is in your bag anyway
and if your really think it has a low chance - then you wont loose much xp,
but yeah you could waste that week for no reason and leave the chance unused
but maybe you could explain why you would.. b:sweat

lmao, I have been reading this thread desperately hoping that this guy didn't mean this... and eventually its there... he actually did mean this.

oh... my... lord...

You should not be allowed to use a number ever again.

Saeidon - Heavens Tear

How many 71+ and 81+ are we making per 91+ on average? To offset costs.

I believe on HT an 81-90 will average out to around 25mill (not sure, haven't looked to buy one). A 71+ probably wouldnt fetch more than 5mill.

Saeidon - Heavens Tear

In my opinion, if you use a good strategy, the major limiting factor is time.

Spirit: You can always get friends to help you level genies for the high end levels, removing a huge chunk of spirit cost, or even use some warsong essences which are basically free now if you just ask friends to give you theirs (not that they do a huge amount)

Mirages: good strategy will pay back a significant portion of the spend anyway via 71 and 81+ genies. Mirages however does correlate strongly with time, so I would want a fairly good mirage strategy for that reason.

Adroit - Lost City

I haven't checked in for a while, but just found the completed excel sheets.. great stuff (thanks HrunsPanda + Kokki)! I actually found a 93LP level 105 genie for sale.. and ended up just buying that (even best strategy I found was less than 50/50 chance to make one). I'll probably still try leveling genies, seems like a decent way to trade time + SP for some coin. These charts will be great for both deciding exactly how I want to level genies, and judge their worth. Thanks to everyone who helped in this thread!