Several Questions About Selfish Mining

[butka]

I have some questions about selfish mining. I could have researched this topic myself, but I want to hear what the community has to say and it is
obviously easier to get a direct answer (so thanks in advance). Just to reiterate for the beginners, selfish mining is a situation where a large pool
or group of miners decides not to broadcast a newly mined block and keep it for itself (for some time at least) and then selectively release it. That way
they will have a comparative advantage of building on top of it, which may or may not pay off.

Questions:

1. Has there ever been a case of selfish mining of bitcoin so far? Or are we are talking about a purely theoretical situation (at least up to date)?

2. How does the selfish miner decide when to release the block and when to keep it? The rest of the network may find a new solution in
the first millisecond after theirs, and they have no way to tell when it could happen.

3. Can selfish mining be performed covertly? Because today we have only a couple of really large pools that could pull this off. The pool operator
may decide to selfish mine, but then all the users of that pool will know, so it would have to be done openly.

4. Associated with the previous question, how could one detect selfish mining? Can an increased number of orphaned blocks be an indicator?

5. If I understand it correctly, the theoretical threshold for selfish mining is somewhere around 33 percent of total network hash power.
But it seems to me that theory doesn't take into account electricity bills, salaries, volatility of bitcoin, or any other expenses the miners may have.
Is there an estimate what would be the practical limit for selfish mining? If it approaches 50 percent, selfish mining shouldn't be a problem in itself.

6. While the amount of power needed to perform selfish mining in Bitcoin is huge, for some of the countless altcoins it is not. Is there a documented case
of selfish mining for some of the altcoins?

[HCP]

6. While the amount of power needed to perform selfish mining in Bitcoin is huge, for some of the countless altcoins it is not. Is there a documented case of selfish mining for some of the altcoins?

Not sure about "selfish mining" per se... but there have been numerous "51% hashrate attacks" on various altcoins. A 51% attack is essentially just extended selfish mining, where the miner creates a nice long(er) chain in private until their goal is achieved and then they broadcast it... triggering blockchain re-organisations etc.

[butka]

Not sure about "selfish mining" per se... but there have been numerous "51% hashrate attacks" on various altcoins. A 51% attack is essentially just extended selfish mining, where the miner creates a nice long(er) chain in private until their goal is achieved and then they broadcast it... triggering blockchain re-organisations etc.

Thanks for the reply. Yes, it seems that compared to 51% attacks, selfish mining belongs mainly to the theoretical domain. Furthermore, to answer my point number 3, selfish mining seems to require a sort of cooperation of a good portion of the neutral network (so it cannot be done covertly), as described in the following 2013. article by Vitalik Buterin:

Furthermore, unlike a standard 51% attack, which only becomes obvious after the fact, this economic attack would need to be announced in advance to let neutral miners know that they have the opportunity to join the attacking coalition for their own benefit.
Source: https://bitcoinmagazine.com/articles/selfish-mining-a-25-attack-against-the-bitcoin-network-1383578440/

This article also outlines a strategy as to when and how to decide to publish blocks (point number 2). The end result depends on the portion of the network (Z) willing to join the attacker's chain:

Suppose the attacker’s portion of the network hashpower is X, and when there are two competing public chains the portion of the network that picks up on the attacker’s chain is Z.
...
As Z decreases, the attacker’s advantage goes down; at Z = 0.5, Eyal and Sirer showed that the attacker becomes more efficient than the public network at X > 1/4, and if X > 1/3 the attacker is more efficient than the public network at any Z.