Forging Data
So, 11/17/2013 - 08:34
Link to the up-to-date spreadsheet: HERE
Current list of things to record (NOTE: Click to open the submission spreadsheet below and fill in all the data!):
1. Fire crystals applied (1x, 2x, or 3x)
2. Current item (star and level)
3. Bonuses achieved (25% heat, double level, forge box)
4. Knight rank (Vanguard, or any of the others)
5. Knight name (Credit who contributed!)
6. Item that was heated
7. Date when the item was heated (to see if the rates change)
A note with the data:
Please only post 1st hand information. I've noticed that secondhand information is often unreliable.
Link to the data spreadsheet: HERE
To make submissions easier I'm going to test out having an editable spreadsheet (with the master-list only editable via me).
Link to the submission spreadsheet: HERE
If I notice vandalism I will password protect it and PM that to the people interested in supporting this study.
The data submitted here will be deleted and added to the data spreadsheet every time I check it (typically weekly to monthly).
To help with this study you can either post in this thread, or fill out the submission spreadsheet.
Up to date with post #41.
Link to the older google docs version of the spreadsheet: HERE
--
Notes on the calculations
--
25% Heat Bonus and Double Levelup are only accessible for heating items of level 1 through 8. At level 9, these boosts become unavailable.
As such, level 9 forges do NOT contribute to 25% and double levelup attempts in calculations.
Forge box chances are available on any heat level.
For more information about the forge, visit the wiki page.
So, 11/17/2013 - 09:16
#2
@Kathrine-Dragon
I'm focusing on the 3x 100% for my data. They don't give the %'s for the heat bonus, double levelup, and forge boxes.
I will include 1x and 2x in my data collecting for this thread, but I personally will not be adding to it/analyzing it.
*Edit: You answered your own question before I made this post :P
Mo, 11/18/2013 - 20:29
#3
Hmm, I don't remember how
Hmm, I don't remember how many times I got double level ups or the 25% heat bonus but I've won 2 shinning, 1 glowing and 2 radiant forge boxes.
Since the forge release I've heated up 19 items, all level ups I used the max fire crystals. Three stars items I leveled up till 5 so I could upgrade to 4*.
17 3* x5 = 85 forges ~ 1 box
17 4* x10 = 170 forges ~ 2 boxes
17 5* x10 = 170 forges ~ 2 boxes
Note: some items were already 5* and some I upgraded to 5* but didn't/won't heat yet. If you want a full list of items and such to make it easier for your data, send me a message in game.
Mo, 11/18/2013 - 20:46
#4
@Mininalice
Unfortunately, I wont be able to use the data for one reason: Double level up reduces the number of forges that actually occurred. We can definitely use it as a sanity check to make sure newer data is pretty close to older data. I also added a "date" column, just in-case OOO ends up changing the percentages.
Mo, 11/18/2013 - 22:14
#5
Derp
Oh trudat, didn't think of it, but for what is worth, I rarely ever get double level ups, most times (when I get anything extra at all) is just the 25% heat bonus. But if I decide to make any extra stuff I'll keep a log of them.
Di, 11/19/2013 - 13:32
#6
right
Right. Do not count forgings by the number of items. Instead, just literally count the forgings. For example, here is my data for one item (yes I know that I should heat 3-star more; I was in a hurry).
3-star 1-9 forgings: 4 total, 0 +25%, 0 skip, 0 box
3-star 9-10 forgings: 0 total, 0 +25%, 0 box
4-star 1-9 forgings: 7 total, 2 +25%, 1 skip, 0 box
4-star 9-10 forgings: 1 total, 0 +25%, 0 box
5-star 1-9 forgings: 8 total, 3 +25%, 0 skip, 0 box
5-star 9-10 forgings: 1 total, 0 +25%, 0 box
Edit: Edited based on the next post. Arrgh! :)
Edit: Edited based on two posts down. Scepticraven is certainly correct that we must differentiate forgings from 9-10 from all other forgings (1-2, 2-3, ..., 8-9), because 9-10 forgings have no chance of skip.
Di, 11/19/2013 - 07:38
#7
@Bopp
With that one skip on 4*, wouldnt that mean 8 forgings total?
Di, 11/19/2013 - 12:41
#8
Also...
It is important to have the current item level for finding the correct averages for 25% heat bonus and double levelup.
Reason being: level 9 will still allow for a forge prize box, but will not allow for either 25% heat or double levelup.
Also added a modifiable spreadsheet link to the first post to perhaps make it easier for people to record data.
Mi, 11/20/2013 - 20:09
#9
Count me in, I've got a few
Count me in, I've got a few things I'd like to craft and level up anyway.
Sa, 11/23/2013 - 09:56
#10
/Cheer
Approaching 100 data points.
With Mininalice's data in post #3, we should be getting close to seeing our first forge box appear in data.
Sa, 11/23/2013 - 12:11
#11
nice
Nice. I have nothing to add right now, except encouragement. Especially when we're trying to estimate the frequency of something rare (forge prize boxes), more data are better.
So, 11/24/2013 - 11:32
#12
Skepticraven, I noticed that
Skepticraven, I noticed that you've put NA in some of the cells, do you want me to put NA when entering the data (ex. 0-2* items don't have any chance of winning a prize box) to make it easier for you to transfer the data over?
So, 11/24/2013 - 15:05
#13
@Saphykun
Short answer: No. Keep doing what you're doing.
Long answer:
I actually have automated all of the data analysis as of like an hour ago. [Lots of fun playing around with never-used before excel functions]
Any lvl 9 item now ignores 25% heat and double levelups as well as 0-2* items not giving forge boxes.
I have yet to add some error handling for the 1x, 2x, and 3x crystals applied (but for now it works 100% if everything is 3x).
The NA's I had before were for using the count() and average() functions - which ignore any text cells. [Am working on this as of right now. EDIT: Done. All calculations are done in a hidden sheet.]
Previously, the averages were a manual sorting of the data which I felt was getting tedious after dataset #100.
Mo, 11/25/2013 - 21:11
#14
SIDECHAIN WOLFGANG!!!
I feel the need for speed to vandalize it.
Di, 11/26/2013 - 07:33
#15
important point
Skepticraven, I hope that you're preventing known trolls from editing. Maybe you should allow access just to specific people. Or ask that everyone send you data, and enter it yourself?
Mi, 11/27/2013 - 17:00
#16
About the Trolls
Only I hold editing access to the main file (viewable by all).
I allow modification of the submission file so that everyone assisting can help with submitting data (I find it easier to type it into a spreadsheet, so I assume more people are willing if they can too).
Worst case, a troll can modify or add trolling data to the submitting spreadsheet. However, if we find this out its an easy removal from my master sheet. (I check and migrate the data just about every day, so hopefully catch everything before a trolling attempt occurs.)
Edit:
Google also allows for me to view revision history (who and what was changed). The only risk is trolls submitting fake data, which I can mark as being strange in my master file (by someone new that hasnt posted in the thread, or someone using a name or someone helping but having a different identification color). Google is really great though and won't let me see the IPs or identifications of editors (keeping them safe from me) - BUT they will tell me if an edit is from the same anonymous editor (giving me a color for who edits it that changes per person).
Mi, 12/04/2013 - 16:37
#17
One more data point
I thought I had more crystals than I really do....
Dark Thorn Shield, at level 6, using 33 crystals (max): Bonus 25%. (I'm a Vanguard, if that makes any difference)
Di, 02/25/2014 - 08:12
#18
I'm happy to announce that since I started this research project, I have gotten my first recorded forge prize box.
The averages with that one prize box are... not good.
Currently 0.36%. I would like to believe that we hit the dryspell of boxes prior to this box but... a relatively reliable estimate should not be expected until at least 10 prize boxes are recorded.
It contained:
3x Spark of Life
20x Glowing Fire Crystal
1x 2 Variant Ticket
1x Advanced Orb Of Alchemy
Since why not record forge box drops as well?
Di, 02/25/2014 - 12:56
#19
great
Yes, please do record such information. It's also great that you recognize the large uncertainty implicit in the data so far.
Di, 02/25/2014 - 13:31
#20
Koniec, Kropka.
IGN: Oski-Boski
Could I join the crew guys?
I'm currently heating lots of stuff so I can help I guess?
Sorry, for not remembering the level skips and bonuses :/
In my history of forging I have only came across one shining forge prize box...
While heating-
Strike Needle: 10 times Max fire crystals, one prize box.
Nightblade: 5 times Max fire crystals, zero prize boxes.
Needle Shot: 5 times Max fire crystals, zero prize boxes.
Fused Demo Suit(Armor and Helmet): 10 times Max fire crystals, zero prize boxes.
Next stuff I'll be heating:
Swiftstrike Buckler,
Heavy Demo Suit,
Silent Nightblade.
Fr, 02/28/2014 - 21:58
#21
@Oski-Boski
Unfortunately, it is unlikely I can use that data.
I discussed this with Bopp earlier in the thread - if we don't know the double levelup, we don't know how many forgings actually occurred.
Additionally, unless you were heating multiple Strike Needles and Fused Demo Suits, they would be 9 forgings each (lvls 1-9 need forgings, but not lvl 10).
I would be very interested in you helping.
You can either submit your data into a spreadsheet here or format the information similar to what Bopp did in this post.
@Thread Research.
I just re-looked at the bonuses.
It seems like the chance for them is dependent upon the item star count.
For heat bonus:
1* is at 50%.
2* is at 48%
3* is at 30%
4* is at 22%
5* is at 18%
I think there's a pattern. It is similar for doublelevelup, and I may also assume for forge prize boxes as well.
Do, 04/24/2014 - 09:54
#22
↓
Chesshirecat has been the first to start contributing data using the 2x crystal multiplier.
The data now currently has 3 forge prize boxes (each with a note of their contents in the datasheet).
The 25% boost seems to be fairly stable for 3* 4* and 5*.
The double levelup is still a bit volatile.
The forge prize box is still no where near getting a reliable estimate.
Sa, 06/07/2014 - 04:26
#23
↓
I'll still need more data to prove it, but a sharp decrease in the rates (for 25% boost) for 0* and 1* may be because I was doing quite a few 1x crystals. The rates may be different for different crystal multipliers.
That being said.
/joy
1000 datapoints so far!
Sa, 06/07/2014 - 04:51
#24
congratulations
Congratulations on reaching 1,000. Now please reach 10,000. :)
Fr, 07/11/2014 - 15:34
#25
I stumbled onto this project
I stumbled onto this project today, and I'd be happy to help achieve some better numbers. My question is, will you be collecting data for the contents of prize boxes as well? Either way, thanks for bothering to collect this information so that the community can have this knowledge.
Fr, 07/11/2014 - 16:31
#26
go ahead
If you plan to open prize boxes, and you don't mind recording what you got, then please do. When it comes to data, more is better. If you're interested in this aspect of the game, you might search for previous threads too:
https://www.google.com/search?&q=prize+box+chances+site:forums.spiralkni...
Fr, 07/11/2014 - 18:16
#27
↓
Yes. Any box I get I will open and record the contents (as a note in the data). You can either submit it in this thread or as I have notated it in the actual data (Data on line #333, the "1" on the forge prize box has a note on the spreadsheet cell).
I may make a new sheet later on to actually do stuff with that data when we get about 10 boxes per type (30 total).
Also, your poor username. Who would do that to a lizard. :(
Fr, 07/11/2014 - 22:21
#28
Ok, I will put the contents
Ok, I will put the contents of the box as a notation to my first entry. I actually ended up here because I was curious what might be inside the box and wanted to do a little research and found a link to this in another thread. I'll try to keep contributing to the data pool when I have the time, and see if I can point any interested parties I encounter to this thread in hopes of a larger data pool. One question about the notation however, is there any type of format I should follow to make this easier down the road? Not sure if you want to parse through that data by hand, even though it is relatively little compared to what you have already been collecting.
By the way, the answer to your rhetorical question at the end of your post is simply...natural selection. "legless lizards" surprisingly exist naturally.
If you're interested:
http://en.wikipedia.org/wiki/Legless_lizard
Sa, 12/06/2014 - 18:51
#29
↓
Bump.
Gunner update is out. I bet a lot of willing contributors are forging up some items. I know I am.
2 more forge boxes were added to my list of loots [bringing 4* box rate to almost 1%!].
Also, I just broke my 1k personal datapoints. Approaching 1.8k total datapoints.
Fr, 02/20/2015 - 19:26
#30
Can I Help.....?
I would like to help, but this seems a bit.... over. Well. The last post was in December, but I was wondering if I could help. I mostly do 2x crystal multiplier, but the more data, the better! This project seems discontinued, but I'm wondering if we could revive it.
Fr, 02/20/2015 - 20:02
#31
↓
Seems discontinued? Maybe.
Actually discontinued? No.
Want to help? Post data either in this thread or in the submission spreadsheet from the original post.
Be sure to list the 7 points of information for every forging attempt. If you want to compress the information in a forum post, you can use Bopp's post as a template for how to structure the data.
I personally like to use the spreadsheet, because I can copy-paste easily and keep information sorted nicely.
I shift over the data from submission sheet to the datasheets every once in awhile, and watch the wiki subforum for updates if people don't want to use googlesheets [I should have it set to anonymous submissions]. Cheshireccat and I have the two big data recorders [consisting of around 90% of the total data in this thread], but quite a few others have helped out [Pyrocover, Leglesslizard, Saphy, Jamesh, and Bopp].
We are currently at 2252 successful and 38 unsuccessful forgings for a grand total of 2290 datapoints!
Progress is slow, but steady. Around 400 were added since the last post in this thread.
Fr, 02/20/2015 - 20:25
#32
still collecting data
Yes, we're still collecting data, always.
Mi, 02/25/2015 - 13:39
#33
↓
I noticed Dracora-Speaking has been updating the forge prize box wiki page.
I felt it was a good time to update the search and display of information pertaining to forge prize box contents.
Since there are only 6 datapoints for this currently, I cannot draw many conclusions past...
1. 20-50 fire crystals of the type of box opened.
2. 3-12 sparks of life.
3. 1 Orb of alchemy.
4. 1 UV ticket or SL key.
The odd part is that orbs don't match the star count of the box all the time. I also suspect that the percentage of UV ticket distribution will match the Equinox prize box.
Overall, I think the wiki page is too specific on the contents already if this is the only data source... We don't know if a 3* box can give 3UV tickets or if only the 5* box can give SL keys. I can say from personal experience that the SL key does come bound from the box [I got one from a 5* box before starting this thread].
Mi, 02/25/2015 - 14:41
#34
@Forge Box
Added link to pool research on Forge Box page.
Just add items as you confirm them to the glowing/shining/radiant lists. Specific %data can remain ??? until we get better approximates. My sources for the basic "is it in the box in the first place" included my own boxes as well as youtube videos. Let's hope there's no stealth patches that have modified the drops since data collection began.
Thanks for confirming shadow key, and bound status of key.
EDIT: added more "guys don't rely too much on this page yet" stuff to note at top.
Mi, 02/25/2015 - 15:25
#35
my current data
Here are my updated data. Please note that the data reported in post #6 are part of these data. So subtract those out if you've already counted them. All of the data I'm reporting here are at 100% chances.
In some cases I got a skip and a +25% in the same forge. In one case I skipped heat level 9 and got +25%, even though +25% has no effect at heat level 10. So that's possible, in case anyone was wondering.
3-star 1-9 forgings
#total: 82
#box: 0
#25%: 24
#skip: 17
3-star 9-10 forgings
#total: 8
#box: 0
4-star 1-9 forgings
#total: 49
#box: 0
#25%: 13
#skip: 3
4-star 9-10 forgings
#total: 6
#box: 0
5-star 1-9 forgings
#total: 30
#box: 1 (25 Radiants, 5 Sparks of Life, 1 2-UV ticket, 1 Elite Orb)
#25%: 11
#skip: 2
5-star 9-10 forgings
#total: 4
#box: 0
Fr, 04/17/2015 - 07:50
#36
↓
Eep. I've been slacking on heating stuff enough to have Cheshireccat almost have more datapoints than me. At the time of this post, hes about 300 forgings behind.
Anyways, just under 3k datapoints with around 100 3*, 90 4*, and 60 5* items fully heated+recorded since the start... That's a lot of heating.
Mo, 07/24/2017 - 04:13
#37
↑↑↓↓←→←→ba
At the time of this post...
Just under 4k datapoints (3962).
20 forge boxes (12 with contents recorded, 10 of them 5*... putting relative chance around 1.3%).
A couple players have asked about "Is there a difference between 3x and 2x double levelup chances".
The data can't quite answer that just yet, because we don't have a ton of 2x.
That being said, I started to heat my 2* gear with 2x instead of 3x (because I can't get a forge box... so doesnt matter to me).
Mi, 10/18/2017 - 04:38
#38
↑↑↓↓←→←→ba
Just breaking 5k datapoints now, with most of the new additions being 1* and 2* 2x data.
Numbers for 1*/2* 2x crystal failure rates are converging towards the expected values (each getting close to 100 samples).
As for the "are the bonus chances different for 2x or 3x" question, the numbers are getting better significance suggesting that they are different.
I may need to put more samples into the 3x bin, since the 2x data is roughly twice as large.
Di, 10/31/2017 - 13:42
#39
Error
I've noticed small error in the Calculations sheet, specifically in the Totals section. Every field sums various counts from heat level 1-8 rows + unused N/A heat level row at the end of each star level section (except fields in Total Forge Prize Box column).
Do, 11/02/2017 - 05:00
#40
↑↑↓↓←→←→ba
I'm pretty sure I took all the situations into account correctly -- The omission of level 9 for 25% boost and double levelups is intentional.
As noted in my other post, the 25% and Double Levelup are only available for lvls 1-8 (+NA). The NAs are specific placeholders for "Unknown 1-8". Forge box still gets the chance at lvl 9.
Cheshireccat was also confused with this through ingame messages and Bopp earlier in this thread... Maybe I should place a brief explanation of the calculations on the first post.
Also, thanks for the data :D.
Do, 11/02/2017 - 11:28
#41
My bad!
Ah, yes, I see now. I remember I read something about it, but it didn't come to my mind when I was looking on that chart.
And you're welcome. :)
(Apologies for misunderstanding)