Fisher-Yates Shuffle Debate

Discussions center on the proper implementation of random shuffling algorithms like Fisher-Yates, critiquing biased or insufficiently random methods in software features such as lists, tests, or playlists. Commenters debate the importance of true randomness, potential loss of entropy, and alternatives like seeded shuffles.

📉 Falling 0.5x Programming Languages
3,448
Comments
20
Years Active
5
Top Authors
#2471
Topic ID

Activity Over Time

2007
1
2008
24
2009
68
2010
84
2011
131
2012
150
2013
123
2014
144
2015
218
2016
160
2017
187
2018
181
2019
337
2020
187
2021
209
2022
270
2023
337
2024
334
2025
295
2026
8

Keywords

NV worldcubeassociation.org RailsConf E2 DHH en.m github.com WCA RIF LFSR random shuffle randomness line probability weighted trick sampling generator order

Sample Comments

dwattttt Apr 24, 2025 View on HN

You can randomly not randomise it :)

ars Feb 1, 2011 View on HN

It might work, but you lose some randomness by doing this. You'd need to check the statistics carefully.

taxonomyman Apr 30, 2012 View on HN

Randomization is a good idea. Give everyone a fair change kind of thing maybe? reply

bean-weevil Dec 10, 2023 View on HN

I wish they had a setting to go back to the Fischer-Yates shuffle.

steckerbrett Sep 4, 2015 View on HN

You can't shuffle and end up with less random output.

Dylan16807 Dec 28, 2023 View on HN

You don't need to make it less random to achieve that. A 100% random shuffle won't have that issue, you just have to make sure it's actually a shuffle and not something else.

nonethewiser May 24, 2023 View on HN

Oh shit, youre right. Thanks for that. I think I saw that method but didnt realize it is implicitly random (like an idiot)

mike_hock Dec 6, 2023 View on HN

That might get them to rely on the randomization, though :)

Nab443 Jul 4, 2024 View on HN

To be fair it's still random from the user perspective.

folli Jan 1, 2022 View on HN

Is there a practical reason that they are randomized?