Log in

View Full Version : Should I use G25 scaled or unscaled?



knez01
07-04-2020, 09:04 PM
I was wondering which one should I use, scaled one gives me better and more realistic results, but I've read that the unscaled version is supposedly more accurate?
What do you think?

Luke35
07-04-2020, 09:13 PM
My results from scaled tend to resemble more closely my known ancestry. Seems to me a lot of people here have come to favor scaled as well. Maybe Davidski has written in his blog about the pro's and con's of scaled and unscaled idk, but I don't recall any discussions here on the forum about it.

Luke35
07-04-2020, 09:21 PM
Okay, one of the "similar threads" linked below (thread founder, marco), contains in it a detailed, technical assessment. But still, there is no definitive conclusion drawn in the thread as to which method has the higher accuracy.

knez01
07-04-2020, 09:29 PM
Okay, one of the "similar threads" linked below (thread founder, marco), contains in it a detailed, technical assessment. But still, there is no definitive conclusion drawn in the thread as to which method has the higher accuracy.

Aha, I find unscaled results very odd, I get lots of Japanese and South African for some reason, while in scaled I get none.
I don't know what to believe, after being on this site for months I've come to the conclusion that so much of genetic testing is very unreliable. From the same raw data I get wildly different results depending on the calculator, it's all so tiresome.

Luke35
07-04-2020, 09:36 PM
Aha, I find unscaled results very odd, I get lots of Japanese and South African for some reason, while in scaled I get none.
I don't know what to believe, after being on this site for months I've come to the conclusion that so much of genetic testing is very unreliable. From the same raw data I get wildly different results depending on the calculator, it's all so tiresome.

I gathered from that other thread that unscaled can pick up on those distant ancestries more efficiently.

It's unreliable and inconsistent yes I agree. However, I really take a very light hearted approach to it nowadays, and just enjoy the process of picking apart my results, examining under a microscope, and then doing it all over again the next day. I find the "hobby" of ancestry analysis oddly cathartic.

knez01
07-04-2020, 09:49 PM
I gathered from that other thread that unscaled can pick up on those distant ancestries more efficiently.

It's unreliable and inconsistent yes I agree. However, I really take a very light hearted approach to it nowadays, and just enjoy the process of picking apart my results, examining under a microscope, and then doing it all over again the next day. I find the "hobby" of ancestry analysis oddly cathartic.

The amount of tests available is ridiculous, you have to buy several packages to get a realistic representation of your ethnicity. And it makes me think that this whole genetic testing industry is nothing but a scam which is made to milk people out of their money. Just like how they removed the v3 chip from 23andme, what could be the reason except greed? Absolutely shameful.

knez01
07-05-2020, 10:01 AM
Well I asked David which one to use, and he told me to use the scaled version, that's what I am going to do.

Chris596
07-05-2020, 10:19 AM
I gathered from that other thread that unscaled can pick up on those distant ancestries more efficiently.

It's unreliable and inconsistent yes I agree. However, I really take a very light hearted approach to it nowadays, and just enjoy the process of picking apart my results, examining under a microscope, and then doing it all over again the next day. I find the "hobby" of ancestry analysis oddly cathartic.


Well I asked David which one to use, and he told me to use the scaled version, that's what I am going to do.

He also told me to use the scaled version more often cause it's supposedly more accurate.

Mr.G, knez01, etc. what are your results here (Marques K64) ? https://www.theapricity.com/forum/showthread.php?325515-NEW-G25-calc-Marques-K64

This one gives me a ton of Germanic but it's probably an error because the rest of it is accurate, sometimes it happens idk why, there's another calc where my first match is a British sample, which is, of course impossible. The German there can be replaced probably by East European and/or Balkan in my case.

Luke35
07-05-2020, 01:04 PM
The amount of tests available is ridiculous, you have to buy several packages to get a realistic representation of your ethnicity. And it makes me think that this whole genetic testing industry is nothing but a scam which is made to milk people out of their money. Just like how they removed the v3 chip from 23andme, what could be the reason except greed? Absolutely shameful.

I personally think it's more incompetence rather than scam on their part - but yeah they're out to make a buck ;)

For me 23andMe was the only company where I felt the ancestry report was very good. They did predict that I am split about equally between Germany and Hungary.