I am considering starting an ambitious project of collecting data on Legendary drop rates. My question to people is--if you had an Excel spreadsheet that asked you to record the legendaries you get, where they came from, and your log on/log off times (I would say rift start and finish times, but I think that would cramp peoples' style efficiency wise), and it was a fairly simple format, would you use it? This would give two benefits...one, you could either have data to back up your arguments or see for yourself that your arguments are debunked, and two, I could accumulate the data and analyze it for larger trends. Please vote on the poll!
I *might* be willing to chime in (you can anonymize the stuff by just making a nice Google Forms submission field with optional authorization token and feed that into a Google Spreadsheet).
I doubt that a random spreadsheet where anyone can contribute will yield good data (there are always those that exaggerate, both on no drops or on many drops, especially if it's not anonymized), but you need a large sample set to make it worthwhile and a nice interface to keep it simple for people to enter lots of data.
I had an idea to re-create the legendary tracker with extended functionality: 1) better usability (start/stop all, save session, export to Excel), 2) distinguish legendaries/sets/plans, 3) log rift start (and possibly end times), 4) specify details of the run (Torment level, full rift clear or just up to rift boss, optional text field for other details such as "server lag issues". Haven't found the time to do that yet, and might not get around to do that within the next few weeks, but if someone wants to have a go I'd be happy to contribute both input and development and possibly some coding as well.
If you really want to go with it I strongly suggest you take a moderated approach and create a Google Forms sheet for people to enter their data, such that you can filter out obvious BS submissions. Otherwise a few people can skew your entire data. But it's different to distinguish fake data from extreme RNG streaks (both good and bad). I have a few samples of well over hundred hours logged in said legendary tracker above (unfortunately only tracked rifts for 4 days and not directly tied to the tracker).
I would be interested in something that tracked starting monster / elite kills and end totals for both rather than time. I this no the data would be more useful towards seeing an overall droprate than drops / time if you're interested in the mechanics of drop rate. +1 would participate.
@Ruluku: Great idea. Maybe it would be worthwhile to start compiling a list of what kind of data would need to be tracked in order to get a comprehensive overview (and if OP creates his spread sheet, can include the right information from the start).
You definitely want
Start time of the run
End time or duration of the run
Either number of rifts completed or even separate start and end time for rifts (only possible with a really good tracking tool like the one I linked, not in a spreadsheet/form)
Monster/elite kills before/after run
Number of legendaries dropped (optionally separated into sets, plans, legs), if in combination with a tracking tool drop time
Optionally, for OCD folks: legendary name
Difficulty
Quick sample of group composition (very minimum: number of people)
Optional field for additional input (again, for crazy ones you wanna track group composition and what not)
What we did was tracking legendaries with said tool above for 7 days (about 120 hours or so if all players combined) and 4 days of rifts (about 100) with time and Forgotten Soul drop rate. All on T1. May post that later. But felt that you definitely need to measure more to be complete, even number of rifts is not enough, as we close rifts as soon as boss is down, many people complete rifts which doubles the time. Elite/monster kills might be best measure here actually.... as it completely rules out efficiency and gives pure "legendary per monster/elite kill" ratio. You can even subtract MF if you know Torment level and number of players.
Create a program that has hotkeys to track the information. I dont see myself being motivated enough to tab to a spreadsheet and enter a time I start/finish a rift, when a leg drops and when my play session starts/ends. If you write a program that lets me start it up when I play, hit a key to start a rift timer/finish it and another key to show a legendary drop then sure.
That is a good point and an interesting idea. I wonder what kind of program would be 100% safe - no one wants to risk getting banned. In theory the program should do nothing than have ~5 global hotkeys for certain commands (start/end rift, found leg, ...). Anyone knows if something like that can ever be 100% safe?
Certainly a real program running in the background (written in C++/Java, prefer the first though) would take quite a bit longer than writing a simple web-app, at least for me, but would have the advantage that the interface doesn't need to be pretty (any rough thing built with Qt would do as it's mostly in the background). And Qt even has CSV export features as far as I'm aware. Hm, I'm tempted... but certainly can't do that before end of May. But before I'd even start I really wanna make sure this is absolutely safe.
That is a good point and an interesting idea. I wonder what kind of program would be 100% safe - no one wants to risk getting banned. In theory the program should do nothing than have ~5 global hotkeys for certain commands (start/end rift, found leg, ...). Anyone knows if something like that can ever be 100% safe?
Certainly a real program running in the background (written in C++/Java, prefer the first though) would take quite a bit longer than writing a simple web-app, at least for me, but would have the advantage that the interface doesn't need to be pretty (any rough thing built with Qt would do as it's mostly in the background). And Qt even has CSV export features as far as I'm aware. Hm, I'm tempted... but certainly can't do that before end of May. But before I'd even start I really wanna make sure this is absolutely safe.
It should be 100% safe since it doesn't interact with diablo at all. It would be along the same lines as the stuff that lets you control media with hotkeys (like play/pause, next track, stop etc)
Rollback Post to RevisionRollBack
To post a comment, please login or register a new account.
My answer is not in there, it's three-fold:
I *might* be willing to chime in (you can anonymize the stuff by just making a nice Google Forms submission field with optional authorization token and feed that into a Google Spreadsheet).
I doubt that a random spreadsheet where anyone can contribute will yield good data (there are always those that exaggerate, both on no drops or on many drops, especially if it's not anonymized), but you need a large sample set to make it worthwhile and a nice interface to keep it simple for people to enter lots of data.
I had an idea to re-create the legendary tracker with extended functionality: 1) better usability (start/stop all, save session, export to Excel), 2) distinguish legendaries/sets/plans, 3) log rift start (and possibly end times), 4) specify details of the run (Torment level, full rift clear or just up to rift boss, optional text field for other details such as "server lag issues". Haven't found the time to do that yet, and might not get around to do that within the next few weeks, but if someone wants to have a go I'd be happy to contribute both input and development and possibly some coding as well.
If you really want to go with it I strongly suggest you take a moderated approach and create a Google Forms sheet for people to enter their data, such that you can filter out obvious BS submissions. Otherwise a few people can skew your entire data. But it's different to distinguish fake data from extreme RNG streaks (both good and bad). I have a few samples of well over hundred hours logged in said legendary tracker above (unfortunately only tracked rifts for 4 days and not directly tied to the tracker).
@Ruluku: Great idea. Maybe it would be worthwhile to start compiling a list of what kind of data would need to be tracked in order to get a comprehensive overview (and if OP creates his spread sheet, can include the right information from the start).
You definitely want
What we did was tracking legendaries with said tool above for 7 days (about 120 hours or so if all players combined) and 4 days of rifts (about 100) with time and Forgotten Soul drop rate. All on T1. May post that later. But felt that you definitely need to measure more to be complete, even number of rifts is not enough, as we close rifts as soon as boss is down, many people complete rifts which doubles the time. Elite/monster kills might be best measure here actually.... as it completely rules out efficiency and gives pure "legendary per monster/elite kill" ratio. You can even subtract MF if you know Torment level and number of players.
That is a good point and an interesting idea. I wonder what kind of program would be 100% safe - no one wants to risk getting banned. In theory the program should do nothing than have ~5 global hotkeys for certain commands (start/end rift, found leg, ...). Anyone knows if something like that can ever be 100% safe?
Certainly a real program running in the background (written in C++/Java, prefer the first though) would take quite a bit longer than writing a simple web-app, at least for me, but would have the advantage that the interface doesn't need to be pretty (any rough thing built with Qt would do as it's mostly in the background). And Qt even has CSV export features as far as I'm aware. Hm, I'm tempted... but certainly can't do that before end of May. But before I'd even start I really wanna make sure this is absolutely safe.