• Welcome to the Speedsolving.com, home of the web's largest puzzle community!
    You are currently viewing our forum as a guest which gives you limited access to join discussions and access our other features.

    Registration is fast, simple and absolutely free so please, join our community of 40,000+ people from around the world today!

    If you are already a member, simply login to hide this message and begin participating in the community!

All-Round Rankings (KinchRanks)

kinch2002

Premium Member
Joined
Dec 22, 2008
Messages
2,504
Location
Guildford! UK!
WCA
2009SHEP01
YouTube
Visit Channel
I devised KinchRanks in May 2015 as a new system to rank all-round ability. Read all about it in the post below, or go to the website where I'm hosting the info and rankings,

https://wca.cuber.pro/kinch/persons

Homepage content:

1. For each event, you get a score of WR/PB x 100 (this will be in the range 0-100)
2. The average across your 18 scores is the final KinchRank score.

The "events" used are:
Averages for 3x3x3, 4x4x4, 5x5x5, 2x2x2, OH, Feet, Megaminx, Pyraminx, Square-1, Clock, Skewb, 6x6x6, 7x7x7.
Singles for 4bld, 5bld, Multibld
Averages or Singles for 3bld and FM (whichever is best for you)

If you haven't done the event then of course you get 0 for that event.
For multibld, your result is adjusted to a single number Points + ProportionofHourLeft. At the time of writing, Marcin Kowalkcyk has a score of 41/41 54:14, which equates to 41.0961. Someone with e.g. 11 points in 45:00 would get (11+0.25)/41.0961 = 0.2737. This calculation ensures that more points always equals a better score, no matter the time spent. Less time spent still gives a better score.
- Higher scores are of course better.
- If you're WR holder, you'd get a score of 100 for that event (WR/WR x 100)
- No matter how slow you are, you will always get some score
- Being double the WR gets you 50 for an event. Same in every event. Treble the WR would be 33.3 points.
- The maximum theoretical final score is 100, although you'd require WR in every event for that.
- For regional rankings, e.g. KinchRanks UK, the regional records are used as the benchmarks instead of the WR (basic calculation of NR/PB x 100).
- For country rankings, the NRs are used instead of PBs (basic calculation of WR/NR x 100).
For a while I had been dissatisfied by the "Sum of Ranks" Single and Average calculation used to show an all-round ability. That's not to say it's a bad system, but I had in my mind that there were some underlying issues with it when trying to assess all-round ability.

After a few other attempts to eliminate the issues I had perceived, I think KinchRanks has finally hit the nail on the head.

Here are some things that I see as drawbacks of the current Sum of Ranks system:
1. There are 2 lists (single and average).
2. Not competing in an event can "kill" your ranking, even if you're amazing at everything else (e.g. Feliks not doing feet, Vincent not doing 7x7x7 for ages).
3. Massive event biases. For example, you have to be really good at 3x3 to get high up. Bigbld barely matters in the slightest. This is due to the depth of events and the participation rates.
4. Personal improvement doesn't always equate to any gains - indeed if you're the World Record holder then there's absolutely no chance of you gaining anything from improvement!

I believe my system has removed all 4 of these perceived issues, as explained:
1. There is one list, that uses a combination of averages and singles as appropriate.
2. No result in an event just means you don't gain any points for it, rather that essentially losing an unlimited number of points.
3. All events have a score range 0 to 100 and are weighted equally in the final result. Your scores do not rely on how good anyone else is (apart from the WR that is the benchmark)
4. Every PB (in average or single as appropriate) gives you points!
What is the reasoning behind the use of averages and/or singles?
Averages are generally a better indicator of ability than singles, so those are used where available. For 3bld and FM, the average rankings are not yet established enough, so single ranks are allowed to be used instead if advantageous for a person. FM may be restricted to averages only in the future if the rankings reach a good point to do that.

Doesn't gaining a second while near the WR gain you a lot more score than gaining a second while far away from it?
This is a good thing, as a second is much harder to cut off when near the WR!

Is it fair that everyone loses score when a new WR is set?
Yes I believe it is. The rankings are supposed to be equal across events, and to do that we must use benchmarks that are as equal in ability as each other. Think of a WR as a recalibration of the system. Over time, it should converge towards an even more equal system than it is currently!

Is FM a little easier to get a good score in?
Yes I believe it is slightly. This is due to the completely different nature of the event compared to the others. The effect is not thought to be big enough to warrant implementing an exception.

Are 2x2 and other short events a little harder to get a good score in?
Yes I think short events are generally a little harder to get good scores because the WR is likely to be a lucky average. Again, the effect is not thought to be a big one

Why are all the Multibld scores so low?!
Up your game everyone! If Maskow can managed 41 cubes, I'm sure if people put time into it like they do with other events, they could get good scores. Being able to 2/2 just means you've learnt 3bld and have some extra locations. Practising 3bld doesn't count as practising Multibld, so I'm not concerned that 2/2 barely gets you anything :)

Why am I ranked much lower than I was in Sum of Ranks?
The most likely cause is that you practised events that have a lot of depth to them a lot (3x3, 2x2 being the biggest culprits). You used to be higher because it's hard to get a low world ranking in those events, and many people were so far behind you in the rankings for them that their other events couldn't make up for it.

Why am I ranked much higher than I was in Sum of Ranks?
There are 3 main causes for this.
Firstly, you are missing some events completely, and were massively penalised for it before with huge ranking scores.
Secondly, you are relatively good at events without much depth to them (bigbld and feet the biggest culprits). You didn't gain much on anyone for your skill before, because others could just avoid them completely without a big penalty. Now you get a score for those events, and others get 0.
Thirdly, that pesky bad ranking in 3x3 no longer drags you down. Many people are very good at 3x3 but have a terrible world ranking because of the depth of the event.
 
Last edited by a moderator:

Kit Clement

Premium Member
Joined
Aug 25, 2008
Messages
1,631
Location
Aurora, IL
WCA
2008CLEM01
YouTube
Visit Channel
I'm not sure that I like completely ignoring singles -- it might be interesting to have a system where events with both single and average have their WR/PB (or PB/WR for MBLD) score computed, and then are given some appropriate weight (25% single, 75% average) that still reflects the value in averages. It's a bit arbitrary, but I do feel like it's better than throwing away this information about the potential for getting very fast solves.

Regardless -- I've always been a fan of the proportional ranks system, but this is an even better improvement on that.

Nice idea. It might be cool to do this for a single competition, to get the MVP of that particular comp.

Now this is something I'd really like to try in the future.
 

Pedro

Member
Joined
Mar 17, 2006
Messages
1,743
Location
Uberlandia, MG - Brazil
WCA
2007GUIM01
YouTube
Visit Channel
I'm not sure that I like completely ignoring singles -- it might be interesting to have a system where events with both single and average have their WR/PB (or PB/WR for MBLD) score computed, and then are given some appropriate weight (25% single, 75% average) that still reflects the value in averages. It's a bit arbitrary, but I do feel like it's better than throwing away this information about the potential for getting very fast solves.

Regardless -- I've always been a fan of the proportional ranks system, but this is an even better improvement on that.



Now this is something I'd really like to try in the future.

From what I understood, singles are included for the bld and FM events. If you check his table, you see for example that Feliks' result for 333bf is 0.526, which is 21.17 / 40.28.
 
Last edited:

Kit Clement

Premium Member
Joined
Aug 25, 2008
Messages
1,631
Location
Aurora, IL
WCA
2008CLEM01
YouTube
Visit Channel
For fun -- this is an individual competition result from Peach State using this ranking system. Events with multiple rounds were based solely on round 1.

Name333444222333bf333oh333fmclockskewb666777333mbfSUM
Daniel Wannamaker0.8939544110.8459119500.8021120290.5686274510.7907692310.76959620106.670971271
Callum Hales-Jepp0.71872510.724231290.63593380610.8631422920.8529411760.7795753290.6022304830006.176779476
Kit Clement0.5922521340.53877140.5123809520.3044619420.6068079190.90625110005.460924348
Ben Cofield0.6307692310.7260566110.5512295080.4382258290.6673032850000.971559412014.985143876
Jacob Ambrose0.5716096320.5646863690.43387096800.499856938100.466187050.8071039560.62077883304.964093748
Anthony Brooks10.55687732310.4801471720.953602620000003.990627115
Sydney Weaver0.6207845840.7457188370.6314553990.5541401270.51671103200.4026109660.4625267670003.933947712
Thomas Kohann0.7940140850.6229208250.71164021200.765892153000.9878048780003.882272152
Clark Cheng0.7091194970.6582879240010000.8476322650.65828125403.87332094
David Ludwig0.7947136560.8432785410.6241299300.863568957000.3188976380003.444588722
Jonas Ventresca0.5609452740.5031573290.43527508100.378794449000.550084890.7413197003.169576722
Jared Stinson0.5427196150.5762424990.56751054900.402441834000.4883195180002.577234014
Chris Tran0.5782051280.674774775000.4437388870.604166667000002.300885457
Chris Foster0.6169630640.653920028000.662495260000001.933378352
Phillip Lane0.3796296300.553497942000.46774193500.4119516850001.812821192
Ramone Jackson0.6013333330.3806668020.36698499300.4417193430000001.790704471
Aussie Greene0000000010.63774780301.637747803
James LaChance0.5410917820.43724460.539078156000000001.517414538
Austin Pitchford0.4828693790.445780264000000.5543199320001.482969575
Parker Oxford0.46614987100.40695915300000.353904970001.227013994
Aarya Kapani0.39929172200.38872832400000.4369521240001.22497217
Nathan Graves0.4873041600.390420900000.3328197230001.210544782
Dylan Miller0.55644663800.64354067000000001.199987308
Raymond Goslow0.68075471700.504690432000000001.185445149
Chaewon Min0.4465346530.587450980000000001.033985634
Tristan Rappon0.43428021200.28495762700000.2794307890000.998668628
Robbie Inglett0.27250755300.33292079200000.3614054660000.966833811
Caleb Saiyasak0.40124555200.28286014700000.2147828970000.898888596
Ethan Botelho0.27865307400.39442815200000.2017434620000.874824688
Michael Lu0.47373949600.346649485000000000.82038898
Seth Bailey0.41662817600.321385902000000000.738014078
Kendal Reid0.43701550400.27962578000000000.716641284
Calvin Bernardo0.30248155600.313885648000000000.616367204
Eric Vande Linde0.35414212800.253057385000000000.607199513
James Bobbitt0.30985915500.270896274000000000.580755429
Brian Weinshenker0.27458143100.301907969000000000.576489399
Noah Joiner0.24944690300.304298643000000000.553745545
Eldridge Harris III0.2829360100.269809428000000000.552745438
Jackson McNab000.25717017200000.269102990000.526273162
Jack Hayhurst0.15345355600.335830212000000000.489283768
Jorge Luis Avila0.25559648600.214171975000000000.469768461
Chris Krueger0.450099800000000000.4500998
Reese McGee0.221785100.223980017000000000.445765116
Ty Fazenbaker0.23489583300.184246575000000000.419142409
Tejas Manem0.32728592200000000000.327285922
Neil Verma0.12787071200.163030303000000000.290901015
Tanner Corbett0.13660457400.066028473000000000.202633047
Agastya Gudipati0.12026666700000000000.120266667
Jacob Freed000.113790186000000000.113790186
Yonatan Nigussie000.107000796000000000.107000796
Ved Rao0.08455193100000000000.084551931

From what I understood, singles are included for the bld and FM events. If you check his table, you see for example that Feliks' result for 333bf is 0.526, which is 21.17 / 40.28.

I meant that within any event with both average and single rankings (like 3x3x3) it would be nice to have some sort of weighted average of the two WR/PB scores.
 
Last edited:

tseitsei

Member
Joined
Jan 12, 2012
Messages
1,374
Location
Tampere, Finland
WCA
2012LEHT01
Me too :). Even if it is just the opposite of what Matthew says/wants, one list!

I just summed up BLD and finds that Grzegorz is the best, followed by Oliver and Marcell.
(And you are in 31:th place)

Yeah I thought Grzegorz and Ollie would be 1 & 2 :D

Maskow wins amazingly much at MBLD but doesn't really do bigBLD so...

EDIT: In a month I'll hopefully be much higher :D we have a comp coming with all events. Hopefully I can get sub-4 4bld sub-11 5bld and sup-15 points mbld
 
Last edited:

DanpHan

Member
Joined
Oct 6, 2013
Messages
645
Location
House
WCA
2011WANN01
YouTube
Visit Channel
For fun -- this is an individual competition result from Peach State using this ranking system.

Name333444222333bf333oh333fmclockskewb666777333mbfSUM
Daniel Wannamaker0.8939544110.8459119500.8021120290.5686274510.7907692310.769596200.80736260206.478333873
Callum Hales-Jepp0.71872510.724231290.63593380610.8631422920.8529411760.7795753290.6022304830006.176779476
Kit Clement0.5922521340.53877140.5123809520.3044619420.6068079190.90625110005.460924348
Ben Cofield0.6307692310.7260566110.5512295080.4382258290.6673032850000.971559412014.985143876
Jacob Ambrose0.5716096320.5646863690.43387096800.499856938100.466187050.8071039560.50119361404.844508529
Anthony Brooks10.55687732310.4801471720.953602620000003.990627115
Sydney Weaver0.6207845840.7457188370.6314553990.5541401270.51671103200.4026109660.4625267670003.933947712
Thomas Kohann0.7940140850.6229208250.71164021200.765892153000.9878048780003.882272152
Clark Cheng0.7091194970.6582879240010000.8476322650.53147166603.746511352
David Ludwig0.7947136560.8432785410.6241299300.863568957000.3188976380003.444588722
Jonas Ventresca0.5609452740.5031573290.43527508100.378794449000.550084890.7413197003.169576722
Jared Stinson0.5427196150.5762424990.56751054900.402441834000.4883195180002.577234014
Chris Tran0.5782051280.674774775000.4437388870.604166667000002.300885457
Chris Foster0.6169630640.653920028000.662495260000001.933378352
Phillip Lane0.3796296300.553497942000.46774193500.4119516850001.812821192
Ramone Jackson0.6013333330.3806668020.36698499300.4417193430000001.790704471
James LaChance0.5410917820.43724460.539078156000000001.517414538
Aussie Greene0000000010.51489372601.514893726
Austin Pitchford0.4828693790.445780264000000.5543199320001.482969575
Parker Oxford0.46614987100.40695915300000.353904970001.227013994
Aarya Kapani0.39929172200.38872832400000.4369521240001.22497217
Nathan Graves0.4873041600.390420900000.3328197230001.210544782
Dylan Miller0.55644663800.64354067000000001.199987308
Raymond Goslow0.68075471700.504690432000000001.185445149
Chaewon Min0.4465346530.587450980000000001.033985634
Tristan Rappon0.43428021200.28495762700000.2794307890000.998668628
Robbie Inglett0.27250755300.33292079200000.3614054660000.966833811
Caleb Saiyasak0.40124555200.28286014700000.2147828970000.898888596
Ethan Botelho0.27865307400.39442815200000.2017434620000.874824688
Michael Lu0.47373949600.346649485000000000.82038898
Seth Bailey0.41662817600.321385902000000000.738014078
Kendal Reid0.43701550400.27962578000000000.716641284
Calvin Bernardo0.30248155600.313885648000000000.616367204
Eric Vande Linde0.35414212800.253057385000000000.607199513
James Bobbitt0.30985915500.270896274000000000.580755429
Brian Weinshenker0.27458143100.301907969000000000.576489399
Noah Joiner0.24944690300.304298643000000000.553745545
Eldridge Harris III0.2829360100.269809428000000000.552745438
Jackson McNab000.25717017200000.269102990000.526273162
Jack Hayhurst0.15345355600.335830212000000000.489283768
Jorge Luis Avila0.25559648600.214171975000000000.469768461
Chris Krueger0.450099800000000000.4500998
Reese McGee0.221785100.223980017000000000.445765116
Ty Fazenbaker0.23489583300.184246575000000000.419142409
Tejas Manem0.32728592200000000000.327285922
Neil Verma0.12787071200.163030303000000000.290901015
Tanner Corbett0.13660457400.066028473000000000.202633047
Agastya Gudipati0.12026666700000000000.120266667
Jacob Freed000.113790186000000000.113790186
Yonatan Nigussie000.107000796000000000.107000796
Ved Rao0.08455193100000000000.084551931

Wait, shouldn't my 7x7 be 1?
 

Tim Major

Platinum Member
Joined
Aug 26, 2009
Messages
5,381
Location
Melbourne, Australia
WCA
2010MAJO01
3x3: 0.535
4x4: 0.468
5x5: 0.443
2x2: 0.478
3bld: 0.199
3oh: 0.539
fmc: 0.714
mega: 0.363
pyra: 0.684
sq1: 0.326
clock: 0.389
sk: 0.678
6x6: 0.350
7x7: 0
4bld: 0
5bld: 0
multi: 0

Total: 6.166

Nice effort Daniel, I really like sum of ranks related stats. It's funny how my results suggest my best events are FMC->Pyra->Skewb->OH->3x3, when regionally 3x3 is one of my worst events. I guess it suggests Oceania is strong at 3x3. These sorts of stats whilst flawed are really good for comparing regions.

My only suggestion would be to make the value used the top 1% result, instead of the WR result. I think the top 1% result is more indicative than the WR, as several events have WRs as outliers at the moment (OH, 5x5, multi). I can't do fancy enough excel stuff to try this, but would be grateful if someone else did.

Also personal request, top 100 for Oceanic "Danielranks"?
Data: https://www.worldcubeassociation.or...Id=_Oceania&eventId=&years=&separate=Separate
People: https://www.worldcubeassociation.org/results/misc/sum_of_ranks/?regionId=_Oceania&average=Average
 
Top