K36 calculator

[Karol Klačansky]

Enthusiasm of @rothaer.

If you remove sub-regions, you are the closest to Czechs-Slovaks-Slovenians-Hungarians, in this order.

German average will not be in the top-10 for you, because of Western Germany.

If you are going to make a calculator with modern populations than they should be populations that more or less exist today and not based off of mixed populations that dont and share to much overlap with other modern populations. You can create sub-regions but based off of modern regions that actually exist otherwise it becomes messy and you get funny and strange admixture results. K36 seems to work very well for people who are not mixed. For those who are its even more vague than if you used less components.

[Karol Klačansky]

You can remove it from the "SOURCE" tab.

Yes I know, but this was a problem even when I paid for the report from Lukasz it was littered with German sub regions that dont even correlate to modern Germany and it probably still is.

[Selene]

...K36 seems to work very well for people who are not mixed.

True.

[Luke35]

Yes I know, but this was a problem even when I paid for the report from Lukasz it was littered with German sub regions that dont even correlate to modern Germany and it probably still is.

It seems to me that K36 works well for you when the source is tuned to your needs.

Target: Karol_Klacansky(Slk+US)
Distance: 977.5935% / 9.77593468
79.0 Czech
21.0 English_Southwest

Target: Karol_Klacansky(Slk+US)
Distance: 915.8221% / 9.15822077
68.2 Slovak
16.4 English_East
15.4 English_Southwest

Cornish,0.13,0,0,2.25,0,6.83,0,0,1.59,0,5.49,0,4.63,7.82,7.42,14.7,0,8.68,0,0,0.01,20.39,0.94,17.40,0,0.01,0,0,0,0.04,0.21,0,0.34,0,0.16,0.88
English_East,0.07,0,0.16,2.33,0.02,7.14,0,0,2.33,0,5.81,0.14,4.39,9.32,6.83,13.77,0,8.62,0,0.03,0,14.70,1.53,20.26,0,0.00,0.01,0.01,0,0.20,0.17,0,0.24,0,0.50,1.30
English_East-Midlands,0.01,0,0.05,2.25,0,7.80,0,0,2.51,0,5.79,0.03,4.11,8.88,6.38,13.22,0,9.62,0,0.00,0.01,15.89,1.31,20.17,0,0.03,0,0.01,0,0,0.25,0,0.52,0,0.13,0.91
English_Lancashire,0.09,0,0.02,2.52,0,7.89,0,0,1.72,0,6.03,0,3.63,9.83,7.37,14.07,0,7.99,0,0,0.00,16.52,1.08,19.89,0,0.00,0.00,0.02,0,0.03,0.20,0,0.23,0.00,0.03,0.70
English_Northeast,0.07,0.00,0.16,1.85,0,7.29,0,0,2.01,0,6.52,0,4.21,8.67,7.95,13.23,0,8.31,0,0,0,18.23,1.09,18.73,0,0.01,0,0,0,0.03,0.28,0,0.68,0,0.09,0.52
English_Southeast,0.05,0.00,0.17,2.19,0.00,6.57,0,0,1.96,0,6.14,0.07,3.61,9.84,7.27,13.63,0,8.99,0,0.01,0.02,15.97,1.65,19.92,0,0.04,0.01,0.00,0,0.03,0.39,0,0.31,0,0.23,0.76
English_Southwest,0.07,0.00,0.09,2.40,0,6.38,0,0,2.17,0,5.22,0,3.64,8.93,6.76,15.81,0.00,7.72,0,0,0.02,17.64,1.24,19.50,0,0.03,0,0,0,0.15,0.48,0,0.54,0,0.11,0.98
English_Yorkshire,0.08,0,0.05,2.64,0,6.96,0,0,1.46,0,6.20,0.02,3.81,10.15,6.87,12.91,0,8.22,0,0,0.03,17.29,1.11,20.69,0,0.02,0,0.01,0,0.02,0.39,0,0.24,0,0.26,0.42
English_West-Midlands,0.02,0,0.15,2.63,0.00,7.00,0.00,0,2.05,0,5.50,0.00,4.41,9.26,7.80,13.28,0,8.84,0,0,0.05,16.10,1.42,19.88,0,0.02,0.00,0.00,0,0.00,0.23,0,0.41,0,0.25,0.60
Czech,0.05,0.02,0.87,2.06,0,9.06,0,0,6.25,0,16.81,0.40,10.49,10.76,5.12,6.75,0.00,6.11,0.00,0.59,0.14,8.20,1.97,11.48,0.01,0.02,0.03,0,0.00,0.11,0.19,0,0.76,0,0.43,1.19
Czech_Bohemia,0,0.13,1.91,1.61,0,9.64,0,0,5.04,0,18.48,0.13,10.27,10.61,4.28,6.17,0,7.76,0,0,0,7.73,2.28,10.54,0,0,0,0,0,0,0.52,0,1.28,0,0.51,1.11
Czech_Moravia,0.00,0.04,0.72,1.49,0.00,8.88,0.00,0.00,5.55,0.00,17.28,0.31,11.49,9.20,4.54,8.15,0.01,8.61,0.02,0.26,0.03,8.41,1.61,10.47,0.00,0.01,0.00,0.00,0.01,0.01,0.14,0.00,0.77,0.00,0.59,1.42
Slovak-2,0.02,0.05,0.86,0.68,0.02,8.54,0.00,0.00,6.86,0.00,18.42,0.74,12.38,9.70,4.03,6.78,0.00,7.45,0.00,0.44,0.02,9.03,1.48,8.98,0.00,0.01,0.00,0.00,0.00,0.04,0.46,0.00,1.29,0.01,0.35,1.36
Slovak,0.00,0.00,0.41,0.86,0.00,8.66,0,0,6.32,0,18.63,1.88,13.72,10.57,4.05,5.80,0,6.40,0,0.41,0.00,8.92,1.16,9.40,0,0.00,0,0,0,0.00,0.14,0,1.24,0.00,0.08,1.22

[Karol Klačansky]

It seems to me that K36 works well for you when the source is tuned to your needs.

Target: Karol_Klacansky(Slk+US)
Distance: 977.5935% / 9.77593468
79.0 Czech
21.0 English_Southwest

Target: Karol_Klacansky(Slk+US)
Distance: 915.8221% / 9.15822077
68.2 Slovak
16.4 English_East
15.4 English_Southwest

Cornish,0.13,0,0,2.25,0,6.83,0,0,1.59,0,5.49,0,4.63,7.82,7.42,14.7,0,8.68,0,0,0.01,20.39,0.94,17.40,0,0.01,0,0,0,0.04,0.21,0,0.34,0,0.16,0.88
English_East,0.07,0,0.16,2.33,0.02,7.14,0,0,2.33,0,5.81,0.14,4.39,9.32,6.83,13.77,0,8.62,0,0.03,0,14.70,1.53,20.26,0,0.00,0.01,0.01,0,0.20,0.17,0,0.24,0,0.50,1.30
English_East-Midlands,0.01,0,0.05,2.25,0,7.80,0,0,2.51,0,5.79,0.03,4.11,8.88,6.38,13.22,0,9.62,0,0.00,0.01,15.89,1.31,20.17,0,0.03,0,0.01,0,0,0.25,0,0.52,0,0.13,0.91
English_Lancashire,0.09,0,0.02,2.52,0,7.89,0,0,1.72,0,6.03,0,3.63,9.83,7.37,14.07,0,7.99,0,0,0.00,16.52,1.08,19.89,0,0.00,0.00,0.02,0,0.03,0.20,0,0.23,0.00,0.03,0.70
English_Northeast,0.07,0.00,0.16,1.85,0,7.29,0,0,2.01,0,6.52,0,4.21,8.67,7.95,13.23,0,8.31,0,0,0,18.23,1.09,18.73,0,0.01,0,0,0,0.03,0.28,0,0.68,0,0.09,0.52
English_Southeast,0.05,0.00,0.17,2.19,0.00,6.57,0,0,1.96,0,6.14,0.07,3.61,9.84,7.27,13.63,0,8.99,0,0.01,0.02,15.97,1.65,19.92,0,0.04,0.01,0.00,0,0.03,0.39,0,0.31,0,0.23,0.76
English_Southwest,0.07,0.00,0.09,2.40,0,6.38,0,0,2.17,0,5.22,0,3.64,8.93,6.76,15.81,0.00,7.72,0,0,0.02,17.64,1.24,19.50,0,0.03,0,0,0,0.15,0.48,0,0.54,0,0.11,0.98
English_Yorkshire,0.08,0,0.05,2.64,0,6.96,0,0,1.46,0,6.20,0.02,3.81,10.15,6.87,12.91,0,8.22,0,0,0.03,17.29,1.11,20.69,0,0.02,0,0.01,0,0.02,0.39,0,0.24,0,0.26,0.42
English_West-Midlands,0.02,0,0.15,2.63,0.00,7.00,0.00,0,2.05,0,5.50,0.00,4.41,9.26,7.80,13.28,0,8.84,0,0,0.05,16.10,1.42,19.88,0,0.02,0.00,0.00,0,0.00,0.23,0,0.41,0,0.25,0.60
Czech,0.05,0.02,0.87,2.06,0,9.06,0,0,6.25,0,16.81,0.40,10.49,10.76,5.12,6.75,0.00,6.11,0.00,0.59,0.14,8.20,1.97,11.48,0.01,0.02,0.03,0,0.00,0.11,0.19,0,0.76,0,0.43,1.19
Czech_Bohemia,0,0.13,1.91,1.61,0,9.64,0,0,5.04,0,18.48,0.13,10.27,10.61,4.28,6.17,0,7.76,0,0,0,7.73,2.28,10.54,0,0,0,0,0,0,0.52,0,1.28,0,0.51,1.11
Czech_Moravia,0.00,0.04,0.72,1.49,0.00,8.88,0.00,0.00,5.55,0.00,17.28,0.31,11.49,9.20,4.54,8.15,0.01,8.61,0.02,0.26,0.03,8.41,1.61,10.47,0.00,0.01,0.00,0.00,0.01,0.01,0.14,0.00,0.77,0.00,0.59,1.42
Slovak-2,0.02,0.05,0.86,0.68,0.02,8.54,0.00,0.00,6.86,0.00,18.42,0.74,12.38,9.70,4.03,6.78,0.00,7.45,0.00,0.44,0.02,9.03,1.48,8.98,0.00,0.01,0.00,0.00,0.00,0.04,0.46,0.00,1.29,0.01,0.35,1.36
Slovak,0.00,0.00,0.41,0.86,0.00,8.66,0,0,6.32,0,18.63,1.88,13.72,10.57,4.05,5.80,0,6.40,0,0.41,0.00,8.92,1.16,9.40,0,0.00,0,0,0,0.00,0.14,0,1.24,0.00,0.08,1.22

yes but this is a good example of what Im saying. If I could change the source of any calculator to just czech/slovak and the British isles I will get more accurate results. K36 works well for people who are unmixed and therefore very similar to the source populations. With 36 components the model is very sensitive so once someone is a bit mixed then u see larger distances and kind of funny results. K36 isnt bad but its borderline overfitting. I actually think a smaller amount of components already acounts for the important amount of variation between populations, and is more flexible.

So sorry for going on but in conclusion more flexible models are needed for mixed people. Un-mixed people who are very similar to source populations benefit a lot from the detail.

[Luke35]

yes but this is a good example of what Im saying. If I could change the source of any calculator to just czech/slovak and the British isles I will get more accurate results. K36 works well for people who are unmixed and therefore very similar to the source populations. With 36 components the model is very sensitive so once someone is a bit mixed then u see larger distances and kind of funny results. K36 isnt bad but its borderline overfitting. I actually think a smaller amount of components already acounts for the important amount of variation between populations, and is more flexible.

So sorry for going on but in conclusion more flexible models are needed for mixed people. Un-mixed people who are very similar to source populations benefit a lot from the detail.

Ah okay, I can understand what you are saying, seems a valid point.

Interestingly, I believe it may be the fine-grained nature of this calc that allows me to get the (accurate to my paper trail) result below. I can't get a 3-way split (Hungarian/German/English) on K13/K15. But I can get it on both K36 and G25, and I had assumed the greater # of components was what allowed for the flexibility of modeling in my case. However, there are other far ranging variables (the limitations of working with averages, quality of the samples, the algorithm of the Vahaduo calc itself, etc), so it's possible I'm misinterpreting the reason why it seems to work better for me. I am just one case, after all.

I look forward to seeing how this calc performs for others as the population averages improve. Also, it may ultimately be very important to have a larger region spreadsheet and a smaller region spreadsheet, as has been done for K13.

K36
Target: Mr.G_Merged
Distance: 7.3495% / 7.34951461
53.1 Hungarian
36.3 Germany_West
10.6 English_Southwest

Target: Mr.G_Merged
Distance: 6.9713% / 6.97130624
51.4 Hungarian
41.3 German_South
7.3 English_Southwest

[J. Ketch]

Good quality (20-25 kits) Scandinavian averages made by Aren added, to replace the smaller samples I had for Danes and Norwegians.

Danish,0,0.00,0,1.43,0,8.52,0,0,1.57,0,8.30,0,4.62,15.00,7.12,8.85,0,6.18,0,0,0,15.46,0.50,21.52,0,0.00,0,0,0,0,0.47,0,0.09,0,0.16,0.11
Swedish_2,0.02,0,0,1.49,0,8.53,0,0,1.64,0,7.36,0,5.03,17.88,6.70,8.43,0,4.47,0,0,0.00,16.34,0.76,20.82,0,0.01,0,0.02,0,0.01,0,0,0.37,0.01,0.03,0
Norwegian,0,0,0,1.27,0,7.40,0,0,2.22,0,7.66,0,5.29,17.52,6.26,8.22,0,4.92,0,0,0,15.16,0.85,22.17,0,0,0,0,0,0,0.04,0,0.42,0,0.63,0.22

Distance to:	Danish_new
3.58866270	Danish
3.85167496	Icelandic
4.11324689	Norwegian
4.11324689	Norwegian-2
4.27141663	German_Schleswig-Holstein
4.30547326	German_Lower_Saxony_North
5.11185876	German_Westphalia
5.34577403	Dutch_North
5.51659315	Dutch
6.18962034	Dutch_Central
7.20698273	German_Lower_Saxony_South
7.47891035	German_Frisian
7.60013816	English_Yorkshire
7.80237784	Swedish
8.20040243	Dutch_South
8.34509437	English_Lancashire
8.50145282	English_Southeast
8.53515085	German_Mecklenburg_Center/East
8.73813481	English_West-Midlands
8.77572789	English_East
8.79562960	German_Mecklenburg_West
9.00607573	English_East-Midlands
9.27859364	English_Northeast
9.43727185	Scottish
9.63703274	German_Hither_Pomerania

Distance to:	Swedish_2
2.77928048	Norwegian
2.77928048	Norwegian-2
4.22177688	Icelandic
5.50527928	Swedish
5.50533378	Danish
6.05424644	German_Lower_Saxony_North
6.67985030	German_Schleswig-Holstein
7.04216586	Dutch_North
7.81959718	German_Westphalia
8.17596477	Dutch
8.23027339	Swedish_North
8.67957948	German_Frisian
8.74297432	German_Mecklenburg_Center/East
8.95822527	Dutch_Central
10.10032178	English_Yorkshire
10.34219996	German_Mecklenburg_West
10.44874155	German_Lower_Saxony_South
10.74925114	Dutch_South
10.78018553	English_Lancashire
10.94397551	German_Hither_Pomerania
11.08715924	English_Southeast
11.29830961	English_West-Midlands
11.42772068	English_East
11.45231418	Scottish
11.54634574	German_Brandenburg_Northwest

Distance to:	Norwegian_new
2.19045657	Norwegian
2.19045657	Norwegian-2
4.89167660	Icelandic
5.35171001	Swedish
5.39505329	Danish
5.48903452	German_Schleswig-Holstein
5.64967256	German_Lower_Saxony_North
7.06743235	Dutch_North
7.19716611	German_Westphalia
8.31900835	Dutch
8.34671792	German_Mecklenburg_Center/East
8.54517993	Swedish_North
8.57722566	German_Frisian
9.06336030	Dutch_Central
9.88922646	German_Mecklenburg_West
9.94731622	German_Lower_Saxony_South
10.08379889	English_Yorkshire
10.40377816	German_Brandenburg_Northwest
10.88651000	Dutch_South
10.91986722	English_Lancashire
10.93413920	English_Southeast
11.00845584	German_Hither_Pomerania
11.09682387	English_East
11.21833767	English_West-Midlands
11.54890038	English_East-Midlands

Also Moldovan averages from Ion Basecul (banned) added.

[Token]

Distance to:	Token
8.20829459	German_Mecklenburg_West
8.48517531	Danish
8.68404859	German_Westphalia
8.72013761	German_Lower_Saxony_South
9.34842233	German_Schleswig-Holstein
9.54715664	English_East
9.65669198	German_Brandenburg_Northwest
9.88516565	Norwegian
10.08441372	Dutch
10.09262602	German_Lower_Saxony_North
10.23945311	German_Hither_Pomerania
10.29560586	English_Yorkshire
10.29571756	Dutch_Central
10.37240088	English_Southeast
10.42685955	English_East-Midlands
10.47304158	English_Lancashire
10.54029411	Icelandic
10.56840575	German_Saxony
10.58358162	English_West-Midlands
10.76864894	German_Mecklenburg_Center/East
10.79570285	Dutch_South
10.87699867	German_Frisian
11.03160007	Swedish_2
11.10784408	German_West_Bohemia
11.51722623	German_East_Prussia_Central

Target: Token
Distance: 589.0294% / 5.89029423
49.4	German_Mecklenburg_West
30.2	German_North_Rhine
20.4	German_Frisian

[mariusz99]

I decided to try K36 only with Poland, Sweden, Norway, Denmark and Lower Saxony North and South.


Target: mariusz
Distance: 740.3023% / 7.40302319
91.4 Polish
8.6 Swedish_2


Target: mariusz
Distance: 743.3332% / 7.43333153
91.8 Polish
8.2 Norwegian


Target: mariusz
Distance: 747.8833% / 7.47883321
92.2 Polish
7.8 Danish

Target: mariusz
Distance: 749.8594% / 7.49859445
92.6 Polish
4.0 German_Lower_Saxony_South
3.4 German_Lower_Saxony_North

I'm unmixed, I have family only from eastern Poland and Hannover. So K36 is perfect.

[Sebbo]

Hi everyone and especially Creoda!

I have been playing around with the K13 vs K36 results of a number of British Isles people.

Some of them have results that are quite similar across all calculators, however some individuals are significantly Southern shifted in the one calculator but not the other.

What could be the reason for this?

Is it possible that some people just get better results with one calculator?

If you were to rate K13 , K36 and Dodecad K12b out of 10 what would you give them?

Thanks!