Member
Hello riders!
To whom it may concern: iOS/Android players/Top 1000 who cares about wrong leaderboard
I am sorry for my English if it's difficult to understand, as I am originally from Russia. Currently I am on 44 total rank iOS - nickname RUS-Uziel
Unfortunately it's long time since Ubisoft support does not make corrections for the global leaderboard and it's results are very confusing, there are a lot of people in top 100 who got millions of fake points and even got highest ranks like #2 absolutely undeserved, nobody knows why RL are not making normal recalculations. One player from our Russian team was able to collect data with 1000 best results, which are currently in Excel file (it's 200 000 results, to get more is much difficult)
Due to the complications with calculation of total global score with people who are worst than top 10 000 on some tracks, or just did not receive new events tracks or others by some reason - I have made a selection of top 1000 riders with criterion "an average rating on all tracks" - which means. the rating (rank) on each track is summarized and devided on number of the finalized tracks which you have (from a top 1000 only). It gives 100% true result only for those who has no results worst than rating 10 000+.
in general I consider this criterion of sorting of a rating quite indicative and I call this as a "skills leaderboard", for example from iOS Carterclem and Alexbyrne in my file are on 2-3 rank, however they don't have several tracks, therefore they are on 1000-2000 in global rating while they are really top players. This sorting is rather on "ability", than on "total result" i.e. if you are everywhere a top 5, but you have no 10+ maps (so your rating at top 1000+), then excel shows you at top 5 result.
Actually if you have all available tracks and no results below top 10 000 then this file is for you, but all others are also quite close to their real place as per their skills.
For instance - mlagadms currently is on rank 2, however his results on tracks are on rank 33. There are many others like him who has 3-6 millions of fake points which were taken by a game error.
Uploading of all information was in this way:
1. People who are rank 1-500 in LB - uploaded all results an all tracks
2. Peope who are above rank 500 in LB - uploaded results only 1-2000 rank on tracks (so all your resuls which worst 2000+rank just missed)
Here are the results:
Android on 31/01/16
Here is simple file where you can sort only results:
https://docs.google.com/spreadsheets...it?usp=sharing
Raw file 33 MB with separation by tracks (on Russian "Скачать" - means download)::
http://depositfiles.com/files/20kd4a5yc
https://ru.files.fm/u/rn7qxpcp#_
iOS on 26/01/15:
Here is simple file where you can sort only results:
https://docs.google.com/spreadsheets...it?usp=sharing
And this one is a whole heavy file 33 MB with all data and how it's calculated-connected:
(on Russian "Скачать" - means download):
https://ru.files.fm/u/hn29fxgx
http://depositfiles.com/files/jp297mcn9
People who are addicted to Excel can play better with numbers on a big raw files 33 MB.
Hope this will show much more precise picture of global leaderbord
Once again - it's a skill leaderboard, not total final score. But total score could be also estimated by this results. It's not 100% precise, but very close to it and surely can show a big picture to those who understand the topic.
Probably this could even help to RL to recognize the problem and start normal recalculation of all results
BTW in a big file you may find almost all results on all tracks and sort it how it's convenient to define what need to be impoved
Member
+100
Member
well done!!
happy to see i'm still in top10 (actually i expect higher)
mlagadms is my friend, and he also knows that he once got the boost on points.
he also wants the correct leaderboard so he will know which rank he's on
btw, i cannot open the 2nd link. it says file exceeds 5M.
Member
I have updated the links for downloading a big excel file, as it's almost 8 MB and system on Microsoft webiste s not working correctly with files above 5 MB
Senior Member
Personally, I dont think its a more precise picture, but its certainly an interesting/worthy way of looking at it. I think RL already knows they have a leaderboard problem, and I dont know that this will help them, but that doesnt mean this isnt good in its own way.Hope this will show much more precise picture of global leaderbord
..
Probably this could even help to RL to recognize the problem and start normal recalculation of all results
Take all of this as constructive criticism...
My biggest problem with this is; Rank is not a very good way to judge; especially when you compare across all tracks and especially with low LB participation. We have a LOT of low participation tracks (50+ tracks at 20K or less)
KarTank is more than a full second ahead of EVERYONE on Chilled Cliffs. Rank 2-10 sounds sooo good next to rank 1... but its not in this case.
Compare that The Cave, where rank 20 and rank 30 are separated by .028 secs. Seriously, is rank the way to judge that?
Being rank 10 on a track with 3869 players (Dark Machine) is not nearly as impressive as being rank 10 on Pike Alley, with 3.7 million players. I cant tell you how many times I have run a track and been top 10 only to come back in am month and find I am more like top 500 once 200K people have played.
The number of tracks played and a players "body of work" should matter, and it doesnt with your current method. I have real doubts that a person who did rank 3 across 10 Tier 1 tracks and stopped playing (or moved to Android)is as good as a person who was rank 6 across 150+ tracks over all Tiers. I think you are being VERY generous to many people who havent had the "opportunity" to play Traction Testground, or did not spend gems to get the Tier3 Challenge tracks, etc. . . or just stopped playing the game a year ago and havent done any of the recent harder tracks You are giving those people rankings they might be able to earn... but have not. The people who did play all tracks will have ranks, and could very easily be unfairly penalized for this.which means. the rating (rank) on each track is summarized and divided on number of the finalized tracks which you have (from a top 500 only).
I know its tempting to give a break to those people who have not played all the tracks due to the excessive grinding of the game. . . .but really, if you want to race for the Global LB, you should be penalized for not having all tracks. GO GET THEM. The LB will be waiting for you when you get there. Likewise, if someone isnt in the top 500 on a track and not being counted in your method or others... BETTER GO GET TO WORK. Or maybe just give them rank 1000 for every track where they arent in the top 500. Penalize them for NOT playing... dont penalize the people who DID play for ranks they actually earned.
(Last week I reviewed all my non-donkey track ranking and was horrified by what I saw over rank 500; if you arent paying attention for a few months you WILL drop out of top 500 on a track. Way out!
Member
Hey there, constructive critisism is always goods. As I said - it's average rate. To those who does not understand math and counting it in excel - I will place an example:
For instance if we take 3 tracks where you are rank 20, rank 25 and rank 10 here is a formula - (20+25+10)/3 = 18.33, which means that on 3 tracks you are in average 18.33 total rate. In total in my file there are 197 tracks counted (even included 3 bunker tracks), so results of all tracks are summarized and deleted on 197 (for those who does not have 10 tracks - it's deleted on 187 etc).
Xitonner - I counted only people with 120+ tracks which is already quite significant result, you can see it on mili-seconds column, there are no people like you mention with results on 10 tracks. I repeat, that it's not a total global leaderboard score, it shows only potential skill of the players and in general it's close to leaderboard. because we can easily define people who got millions of fake points - therefore in my file they jump from ~#15 to ~# 90.
Just think about it logically - basically for 1 track where you got platinum are a getting approximately 800 000 points in total, all rating above will give you additional ~40 000, and it's variation is placing you on some rank - therefore, you can easily connect this rank to the points you got, there is almost line proportion for it. Therefore ranks could be considered and counted almost the same value as a total score.
Your example with the Cave vs Cliffs does not make any deviation, it's still counting between ranks, not the time and between 197 tracks results in general 2-10 exception things does not make any big sense on statistics, they are leveled on average counting. If you are riding on top 200 in average - then surely the Cave you'll do approximately the same, it will not be a top 1000-2000 rank which will spoil all the picture! You can read math statistics to understand it better
Regarding Traction Testground - there are only 800 people who finalized it. Approximately 200 people got there a platinum medal. 1 mistake (fall) starts from 289 rank. Traction Testground is approximately like a Quagmire as per difficulties rating, probably a bit easier, but results as per the platinum rank are mostly similar and fair enough for comparing between 2 of them, surely a bit higher than it should if we opened the track for everyone, but as per my view - this is also important, because this track is not coming by a coincedence or easily by the event, so people who got it and finished should get some small benefit in general, even if this benefit between 197 tracks taking in average is almost nothing
P.S. for sure some results are not very fair, like between Carterclem and Abylons - Carter has very good results on all maps which he finished, but he finalized only 160 tracks out of 197, which is surely good, but in comparison with Abylons who finalized all 197 almost with similar average rating - it's not fair. So I would rather put Carter right between Abylons and DanielP, but it's all demagogy. Counting system almost for all top 100 in my file is 95% correct, probably could be some deviation for 1-3 ranks, but it's not significant, so it's up to each one of you to decide how you are performing in average. It could be really strange if mlagadms (top #2 currently) who got many fake points and finalized all tracks will say that he is playing much better than Alexbyrne (top #1723 currenly - due to the lack of new tracks and long vacation period).
P.S.S. And once againCounting and ranks are correct according to the criterium of average rank, you cannot argue with it. I repeat - that it's not total global leaderboard score, it shows only potential skill - how people are performing on tracks, which they have
Member
Thats Great work !!!!!! Amazing!!!!
Rank 16 in Global LB for me 😊
Community Developer
Hi everyone,
I am really not good at calculation, so what you did is even more impressive!
Just to answer your concerns about the LB recalculation, I can assure you that our devs are working on it right now (ok, maybe not now now because it's a bit late for them). Unfortunately, it takes some times to recalculate everything, the formula is complicated, a lot of factors are involved.
But it will come!
JollyCharly
Member
poor mlagadms
i think mlagadms and abylons both got fake points.
but abylons is too good. he ranks high even without the fake points.
so mlagadms becomes so typical in this situation
Member
I just got new data with top 2000, so it's now counted more preciesly and last information from today! Unfortunately Gregory have beaten youOriginally Posted by Roger-_-Lee Go to original postpoor mlagadms
i think mlagadms and abylons both got fake points.
but abylons is too good. he ranks high even without the fake points.
so mlagadms becomes so typical in this situation
https://docs.google.com/spreadsheets...it?usp=sharing
BTW all bunker players at normal rank now can find easily their opponent and see for what they worth
Also one more column to compare - current rating on 12/01/16 versus average rating
