Reducing Moonbeam self bond back to 100k GLMR

[legendnodes]

Abstract:

In order to keep the Moonbeam network diversified, sustainable and healthy, I recommend reducing the self bond to 100k GLMR (or to any value that’s significantly lower than 2m GLMR).

Details:

~2 years ago we decided to increase the Moonbeam self bond to 2m GLMR (proposal 88). The main reason was because some collators didn’t perform as expected, didn’t produce blocks as they should’ve, weren’t responsive, while good collators found themselves out of the set (while the poor ones were in). We supported increasing the self bond in hope that with more skin in the game, we’ll probably get better collators.

In addition, back then, in order to be active the minimum pool size was roughly ~4m GLMR, therefore it didn’t make any difference whether the self bond is 100k GLMR or 2M GLMR - anyhow the collators need to raise at least 4m GLMR in order to become active, so better to be more restricted and demanding.

Why change it back now?

Currently the situation is significantly different. Here are some examples:

1. Today, in order to become active - collator needs much less than 2m GLMR. As you can see below, it’s around 1.2m GLMR (there were periods when it was even 600-700k GLMR only)

2. Some collators left the set or are winding down (simply staking is the last one who revoked a significant amount lately, GLMRHUB was before him), and the waiting list collator (as you can see below) is very short. Assuming all the collators on the waiting list are running active and updated nodes (and most likely they are not), there won’t be enough collators to fill the active set in case more active collators will drop out.

3. On the other hand, The 2m GLMR self bond prevents new collators from joining, especially the small one who may become active with a much smaller pool than 2m GLMR.

4. Last but not least, nowadays we have the possibility to chill inactive collators, but we can’t do it if there are no collators in the waiting list to take their position

How it goes with MOVR?

In MOVR we also had a suggestion back then to increase the self bond to 10,000 MOVR, but eventually it turned back to 500 MOVR (as it is today). It’s not a coincidence MOVR collators waiting list is much longer and with real collators that are definitely waiting to become active if someone drops.
Just to put it $ terms - self bond for MOVR is ~$3700 while for GLMR is $230k.

Summary

As a collator who supported increasing the self bond to 2m GLMR, it was known that this decision is reversible if the environment significantly changed, and I believe it did and therefore GLMR self bond should be back to normality.

I will be happy to hear what the community thinks and people’s insights.

Thanks,
Rafael

[Aleksandr]

hey Raphael, thanks for your proposal!

actually, the past decision to increase it to 2M was great at the time because there was chaos in the active set - collators were frequently getting knocked out since some had more total bonded, and delegators kept losing 7 days of rewards due to revokes, etc. even choosing a new collator didn’t guarantee stability, so the increase helped bring more consistency

but now, things have changed a lot - market conditions are different, and it might make sense to adjust the self-bond back to 100k. i see new collators joining discord / telegram, asking how to run a collator, and for them, a 2m self-bond doesn’t seem reasonable. lowering it could help onboard new technical community members. the only concern is whether we’ll run into the same issues that led community to raise it in the first place. not sure if that’s the scenario we’re looking at, but creating some competition and increasing demand for GLMR is definitely a good thing, so i totally support it

[legendnodes]

Thanks Turrizt for your support. Regarding your concern below:

@Aleksandr

the only concern is whether we’ll run into the same issues that led community to raise it in the first place

Don’t forget that unlike the last time when we decided to raise the bond to 2m, we can now chill collators that don’t perform (missing 1 round in GLMR & 2 in MOVR). Therefore, if we see this kind of behaviour - we can chill them out and hopefully better collators will join instead. However, for better collators to join - we need to have a reasonable self bond.

[tram.papam]

Looks pretty interesting.
I wanted to clarify some points:
- How many new validators are expected when the threshold is lowered? What are your thoughts on this?
- Are there any plans to strengthen monitoring to prevent repeated problems with unreliable validators?
- Do you think this will have a positive impact on the decentralization of the network?

[legendnodes]

1. It’s hard to tell how many new collators (or old with another node) will join, but it’s easy to tell how many joined in the last year for example and the answer is zero.

Stellaswap 2nd node became active only with 700k GLMR (they already had a collator with 100k self bond on the waiting list). I’m sure that if the self bond wasn’t 2m, few more potential collaotr could’ve join and try to get a spot.

Even now when the thersold is about 1.3m GLMR it can encourage new potential collators to join.

2. Today we can chill out collators that aren’t perfroming well (missing 1 round on GLMR & 2 rounds on MOVR).

3. Yes.

[Aleksandr]

ah, honestly, the issue wasn’t really about someone not producing blocks due to poor node management. It was more about the active slot being chaotic - one collator kept knocking out another every round because collators from the waiting set had a higher total bond, etc. so that was the main issue

[legendnodes]

Yes, indeed, this can definitely can happen as long as there is a battle at the bottom, but eventually when someone loses (the battle can’t go forever) then it returns to be more stable

[blackk_magiik]

Hey @legendnodes great points!..here is my counter argument

Why not set the minimum reserved bond to 500k GLMR? It’s still a significant reduction, improves decentralization but more importantly encourages only seriously committed validators to join the active set. IMHO, setting the bond amount too low, especially in volatile market conditions, can open the chaotic mess we saw few years ago, that @Aleksandr is referring to.

Remember there is no penalty slashing in Moon* networks. The worse that can happen to you is getting “chilled” and kicked out from operating negligence. At that point, you just collect the rewards and move on or come under a fake name and get back into the active set. Yes, there is a balance between opening to the community for larger adoption and decentralization but where has that demand been? I could be missing it, but I occasionally see a request to join as an orbiter or what’s needed to run a collator node. I haven’t seen high, strenuous demand to get in on the action.

This demand could be renewed if the path to Ethereum adoption works out, but for now, objectively speaking I say we reduce but stay the course in using the reserve bond as a sign of committing to Moonbeam.

Full disclosure to the community, yes I do operate a collator on Moonbase/Moonriver/Moonbeam as “Paradoxx”

[legendnodes]

Hi @blackk_magiik , thanks a lot for your feedback - it’s always better to hear different thoughts. Let me address your concerns one by one:

1. Setting it to 500k GLMR is definitely an option, and I suggested it as well when I said:

@legendnodes

or to any value that’s significantly lower than 2m GLMR

     500k is significantly lower than 2m and it's a valid option

2. You’re right that slashing doesn’t exist, but it’s for a different discussion imo whether we should consider it as well. However, chilling collators is a good solution that already helped to remove poor collators (both on Moonbeam & Moonriver) and get better ones instead. It’s true that chilled collator can become active again (if they have enough bond), but if they underperform long term, their pool will become lower naturally

3. The main reason people are asking about orbiter (and not how to become a collator) is because they already know they can’t be a collator with 2m self bond. Therefore, the only solution for them is the orbiter program

4. Imagine for example that one active collator decides to leave (or being chilled) - the first collator on the waiting list right now is Cosmoon with 287k GLMR, followed by GLMR-HUB1 with 186k GLMR. Hence, if a new collator wants to join he can’t because he needs at least 2m GLMR for self bond, while if the self bond was 100k GLMR, he could easily do that and try to become active if with 300k GLMR (including the self bond). This isn’t an imaginary scenario - it happened a few weeks ago with STELLASWAP-2 that became active with 700k GLMR only.
   The waiting list eventually turned into a close group where no one can really join. When the minimum to become active was about 4m GLMR, it made sense to have 2m self bond, but now when you can become active with much less, why not open it to the broad community?

5. I agree with the statement that if new collators will join, and there will be “battle at the bottom” then it’ll cause frequent changes with the active set. However, history shows (for both Moonbeam & Moonriver) that eventually it’s settled. The fight to become active can’t go forever. Eventually someone gives up, and the moment it happens - the stability returns

To summarize, I think the risk of not having new collators at all (meaning being a closed group due to the high self bond) is much higher than having too many that “fight” each other to become active. If we go back to the waiting list, there are 4 collators there right now, and I’m sure that at least 2 of them are off and not even aware they have a collator. It means that if 3 active collators decide to leave (which isn’t rare scenario - many left so far) - we won’t be able to fill the gap just because of the high self bond

[rafaelcastaneda]

I’ve understood the proposal and it seems fair, considering past and current conditions.

But out of curiosity, what happens with the unbonded 1900k in this case? Is it reverted back to a specific treasury or is there any purpose for this amount to be allocated?

[legendnodes]

It goes back to the collator account since its belongs to him. In order to launch validator, you need to deposit 2m GLMR in your validator account. If the theresold decreases to 100k for example, then the collator can withdraw 1.9m back.

[Jim_CertHum]

Thanks @legendnodes for putting up this proposal.

Once again the community is back at discussing the balance of keeping the collator set stable via a bond that favors conviction vs allowing new participants that may not have the capital to replace collators that are not focused on the responsibilities of their duty. But it makes sense, because the token price is dynamic, that this discussion comes up again.

I’m in favor of lowering the bond. I do share the concern that if it’s lowered too much, the staking will re-balance, and we’ll end up in a scenario where there’s a frequent rotation of nodes in the active set, with delegators taking the loss.

Therefore I would be in favor of a reduction to 500k GLMR, Any existing hodlers may be inclined to join the set – those that already hold high conviction in the network, non-hodlers would have a serious outlay, and it would help avoid a situation where speculators or those who would just like to cause some chaos for the lolz decide to disrupt the active set.

[legendnodes]

Thanks @Jim_CertHum for your feedback.

Yes, I agree that a 500k self bond is a good balance to start with and see how it goes. Eventually it’s reversible, and we can always keep lowering it or even bring it back up.

[aaron.mbf]

Thanks for bringing this up @legendnodes . It sounds like we have a good amount of community support to lower this to 500K.

I believe this must be done through a runtime upgrade. My suggestion would be for you to create a governance proposal on Fast General Admin with a System remark referencing the proposal in the forum.

Assuming this passes, then @Piotr_Dziubecki can get it added to the next feasible runtime upgrade.

The Foundation would be happy to support the decision deposit for it. Let me know if you have any questions on this!

[legendnodes]

Thanks @aaron.mbf for your guidance

I’ll create the governance proposal as suggested. If it’ll pass - Moonbeam self bond will be reduced to 500k GLMR in the next RT upgrade.

@Piotr_Dziubecki - I’ll contact you directly on tg about the referendum process.

Thanks everyone for your participation in this discussion! It was really insightful