It's easy to do worse than random. Trying to do better than random often ends with doing worse than random, while aiming for randomness usually ends with randomness.
> They thought they could randimize their way to fairness too, it looks pretty silly and ineffective in retrospect:
I see information about the process but none about how and why it failed.
Cases normally are assigned a case number anyways already. So lets start from case number 0.
First lets choose a nothing up my sleeve number. Let say like the fractional part of the PI with the bit length of some secure hash function. We can treat this case -1 or the initial state.
The plaintiff and defendant each generate a private and public key pair. They then give each other their public keys. The plaintiff and defendant each sign the what will become the previous state with their private key. XOR the the two signatures together and then hash the XOR-ed signatures. This gives the new state. We then take this new state and take it to the modulus for the number judges. Just repeat this for each case new case.
Neither the plaintiff nor the defendant should be able to directly influence who their judge is in this system. Also both can verify that other party performed a proper signature. Although there could be gotchas depending on the underline primitives from hash function to signing algorithm.
Also, your method is biased towards lower-number judges when the number of judges is not a factor of the range of the output of your hash function. Use an existing library instead.
I could sit down try to figure out a construction that adjusts/accounts for any given N judges. Again though like I said this was just quick thought off the top of my head showing it seems possible to make a construction that can choose judges. Not something with every possible detail refined.
However, modular bias becomes smaller the larger your max number gets. We are not talking a small 32 bit random generator here. While it would be ideal to eliminate the bias it's not exactly easy. One could try the method of re-rolling/generating till you get an output in the target range. This has a massive problem because it is improbably at the scale of these numbers. Secondly, one can't control the modulo N in this case. There are things we can try to smear the bias, but that does not eliminate it.
Moreover, as mentioned before the scale of max output works in our favour here. The larger your max output/range is the smaller the bias is. Anyways there is a reason the CPRNGs will just truncate their output when you tell one to return X bytes. The range of possible bias is so small because how large cryptographic numbers are. Although some algorithms max output is a power 2 so the modulo N works out nicely.
The worst a bias like this can do is make partitioning cryptanalysis stronger, but unless you made some really bad choices generally should only be a theoretic break. At which point your designing primitives and not a protocol. However, that part we are using the modulus on is not secret so we are not gaining/losing anything there. For this instance it's just extremely tiny bias towards a certain judge.
If even one of the parties submits a fairly chosen random number, and manages to conceal it from the others until they've submitted their numbers, then the result will be a fairly chosen random number. Then no one involved will have standing to complain that the result wasn't chosen fairly enough.
Exactly.
We can't trust judges. And we're in a lot of trouble.
What specific problem were the Venetians trying to solve with that design, and how did it turn out to be ineffective?
Alexander Hamilton wrote in the federalist papers that we need juries to protect us from corrupt judges.
We should trust judges.
We should also create a system that strives to reduce the chance that individual judges could get to a point where they're less than trustworthy. Judges are people too, and respond to incentives. And yeah some individual judges suck, and are bad at their jobs. Like some doctors, and some lawyers, and some engineers.
A system that reduces their impact and influence, like randomization does, minimizes the harm of a bad apple, and also reduces their blast radius. They can't advertise that they'll adjudicate your case.
After all, isn't this why we have appeals courts? If you could trust your individual judge implicitly, why allow an appeal?
Trust, sure, but verify.
> Or what happens if I accuse the randomizer of being untrustworthy?
Open source the RNG. Let people see for themselves. Publish the stats that show it's a uniform random distribution.
The default assumption should be that people are corrupt given the right incentive which in this case shows. It's not perfect but it's better that someone is not certain that they are able to attract certain cases for future.
They're just humans. They are not to be trusted by default until they prove otherwise.
Moreover, what if the randomizer picks one out of 297 judges, 235 of whom are known to be biased? Then it doesn't help much.
You may want to check out how elections were done in Venice in centuries past. They thought they could randimize their way to fairness too, it looks pretty silly and ineffective in retrospect: https://www.venetoinside.com/hidden-treasures/post/the-elect...
The legal system is for the exceptions, the places that automated systems break down, the places that need human judgment.
That's the importance of a free press. We can't work without trust, we need it, even though it introduces certain risks.