A practical and transparent method to sort

In this short post, I introduce a method that might interest readers from this blog. I present a practical method that randomly pick people without the need for a trustworthy third party. Before describing it in more details, I want to insist on its two main features.

The method is practical because it only requires a basic computer (a smartphone largely suffices). Computers cannot generate random numbers because they are deterministic, however they can produce long sequences of numbers that do not repeat themselves for a long time. If you don’t know how the computer generated sequence work it is difficult to predict the next outcome. The numbers in this sequence will constitute our set of random numbers.

How could we generate the set without a trustworthy party? A malevolent individual can let the computer run until they obtain the desired results. If you know the initial number of the sequence and how the sequence works the result becomes entirely deterministic. We call this number the seed and this number will entirely determine our set of random numbers. We use this property to generate an impossible to forge sorted list of members. But how do we pick this seed?

Our method uses a stock market index. This has two interesting properties, firstly it is impossible to predict (otherwise tell me so we can become rich fast :) and every market agrees on it. This guarantees true randomness. Secondly, everybody will be able to reproduce our sortition if they have the seed value, the member’s list and the script we used. This guarantees transparency and accessibility.

We will employ this method in our association (l’Association Française pour la Sortition) to constantly renew our board. I already wrote the code using the CAC40 index to perform this task, it is free, open-source and you can access it you can follow this link: https://framagit.org/PersonnePirate/sort.

13 Responses

  1. Romain,

    Generating a trust-worthy allotment procedure in a high-stakes situation like selection a national parliament is certainly a non-trivial task, and it seems to me that the associated difficulties are often under-estimated. It is true that ultimately what matters is not whether the procedure is “truly random” pr whether the samples it generates are “truly representative” (whatever those phrases mean), but rather whether the public is happy with the policy outcomes generated by the political system. However, if the allotment system can be rigged it is possible that it would be rigged and as a result policy outcomes could be undesirable. Certainly in the political atmosphere of division and suspicion generated by the decades of abuse and corruption inherent to electoralism, it is important to make sure that the allotment system is perceived as well-functioning a-priori, even before long term policy outcomes can be assessed.

    The procedure that you are proposing attempts to address items 3 and 4 of the list in the post linked to in the previous paragraph. I think it is doubtful whether it properly addresses either. The smaller problem is with item 3: To what extent a mapping based on a fairly hard to define and understand pseudo-random number sequence can be considered and publicly perceived as “an equal-probability allotment mapping” may be a matter of dispute.

    The more problematic issue is with the seed generation. The “practical randomness” of stock indices is quite doubtful in my mind. Even if we trust that the computers calculating the index are not themselves rigged (something that would be very difficult if not impossible to verify in practice), there is no reason to think that a powerful actor could not manipulate the values of the least significant digits of the index purely through well chosen market transactions. (It is worth noting that essentially the same objections can be leveled at using a blockchain system as a source of randomness – a proposal that has been fashionable over the last few years.) In view of these epistemic difficulties, I would not be surprised if doubts regarding the randomness of the seed would be raised as soon as some powerful political actor has any use for those doubts. And in view of the opaqueness of the system it would be essentially impossible to dispel such doubts.


  2. The American chapter of the sortition foundation build their directors board, by drawing papers from a hat and film it. Don’t get me wrong it is not a criticism, they did a world premier and they are doing truly amazing work!

    What I want to say is that the sortition method depends on your need. The one I propose here is not perfect (as the most advanced cryptographic algorithm), but it is a quantic leap compared to the hat method.

    The USA chapter could use the hat method because they had less than a 100 members who knew each otehr well. I hope one day we will be 1000 – 100 000 in this case the hat method does not hold a modicum of scalability and trustabily test. The method I propose here will work I think however.

    Like all forgeable things (like money or official papers or designations) it is a cost/benefit balance. Yes, an individual like Jeff Bezos might be able to influence the stock market but i would require an immense amount of money/power and elections are already heavily influenced by these kind of people. Plus an extremely tiny variation in the seed will dramatically change the outcome (this is how modern cryptography works), so the outcom would be still had to predict


  3. I am more worried about the CIA than about Jeff Bezos. But I do have an alternative in mind that I think will be more conspiracy resistant and which I intend to offer in an upcoming post.


  4. A market index cannot be controlled at the decimal level even by whales (big market player), financial markets are inherently unpredictable because they are multi-actors.

    I left a comment on your great post about the subject: https://equalitybylot.com/2019/07/10/criteria-for-the-acceptability-of-an-allotment-procedure/#comment-36887


  5. I am not sure why your faith in the unpredictability of the stock market is so strong. Surely just saying “multi-player” is not good enough, since it is easy to set up multi-player systems where a strong player can control the output.

    For example, let’s say that each of a N players chooses a digit (0 to 9) and the output is the sum of the N digits modulo 10 (i.e., also 0 to 9). Is this an inherently unpredictable multi-player system? Yet a single player who simply gets to see the sum of the other N-1 digits can instantly generate a digit that will produce the output they want. Surely the least significant digits of the stock market can be controlled in a similar manner by a trader who is in a privileged position.

    Again, to this risk you must add the risk that the price-setting algorithm itself, which is presumably a very complex piece of software, would be tampered with.


  6. To convince you of the unpredictability of a financial market just make a reasearch on internet. The wikipedia article https://en.wikipedia.org/wiki/Stock_market_prediction states that in the second paragraph. Market index behave randomly.

    Someone that knows everything might be able to predict everything (google Laplace Deamon). A stock market work with millions of actors dealing with thousand of numbers.


  7. Also about your last remark.

    > Again, to this risk you must add the risk that the price-setting algorithm itself, which is presumably a very complex piece of software, would be tampered with.

    The very short (10Lines without a user interface) program is impossible to tamper because we distribute it with the list of people to be alloted. The allotment can be done by everybody separately, no need to refer to a central authority.


  8. A little bit of trivia: if I recall correctly, according to Malcolm X in his autobio the illegal numbers gambling racket in New York City (called “the numbers”) used some number in the stock market on the day in question to determine the winning number. Clearly an excellent precedent.

    Liked by 1 person

  9. It seems to me that the “WallStreetBets” affair, involving multiple layers of manipulation in the stock market by a whole host of players, should give pause to anyone who believes the stock market provides a manipulation-resistant publicly-observable source of randomness.


  10. Hi Yoram,
    So you are concerned that a huge attacker (very rich) or multiple attackers with the same agenda (thousands at least) would manipulate the stock market, isn’t it? I agree it is possible, buying a lot of a stock will for sure increase its value, but do you think it is possible to have the chosen exact value? for instance, you want a stock share to exactly value: 1.1568799653 (this is the kind of precision we are talking about at least).


  11. Hi Romain,

    My point is that the system is opaque and open to manipulation in a variety of complex and undetectable ways. The notion that the public should or would trust the numbers coming out of this system is, it seems to me, highly unrealistic.


  12. I understand two things in what you say: 1) the stock market can be manipulated and nobody would notice it 2) People don’t trust what the stock market tells. About the second point, people do trust the stock market, it is not my opinion it is how foundational for the stock market the everybody agrees about it. On the first point, we both agree that markets can be manipulated but as I said previously it is impossible to reach a high level of control.


  13. Sorry, what I said and how I said it was clumsy


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

%d bloggers like this: