Make sense , at the moment i think this will be beneficial for many T4 dApps so they don’t loose interest about dApp staking and can focus on building on Astar.
1 Like
First of thanks for the explanations on the updates!
Seems like some lower dApps are now getting some more reward. I think this is a good first step, but have to agree aswell with @you425 vision about the future of dapp Staking.
1 Like
The current model is in any case incorrect and not in line with what was promised to the dapps who made their proposals in v2.
Adding a few peanuts to peanuts still makes peanuts.
It’s an adjustment that doesn’t change anything. Posts in the forum asking for explanations about the v3 system have grown and several teams have expressed doubts.
@you425 Based on the current number of dapps, would it be possible to make (when you have time) also a comparison with the linear v2 system with a hypothetical drop in rewards of 50% or more for all dapps? You could also easily set a maximum cap of Astars that can be staked on a dapp, something that is slightly higher than the current tier 1. What would be the impact on inflation?
Tiers are not tied to voting and build to earn periods, which are a fair narrative, so just burn them instead of team support…
Yes, I think it is the same even if the increase in rewards is almost linear.
Even in v2, the developers would have been trying to get more staking. This became more pronounced with v3 Period. Of course there is a Tier effect, but that only works around the threshold and the effect is limited.
That is true, but I believe that the results of the test are to be reflected in the production. The difference in the parameters now probably doesn’t make sense.
Well, let’s end this discussion on the topic of this, since it is not the main topic of this issue.
Personally, I don’t like the classification by tier and even less adapting the thresholds according to the token price.
Here is my personal analysis regarding the Staking V2 and V3 dApps and my suggestions : dApps Staking v3 - proposal - #154 by GuiGou
I think it better to simplify the dAppStaking implementation with a logarithmic function (or more sophisticated function) instead adding additional rank or complicity in the tier classification.
Generally, the simpler the better:
- simpler to implement => less bug and easier to improve,
- simpler to understand => less misunderstanding for stakers and builders
1 Like
When dApp Staking was changed to v3, Tokenomics was also changed at the same time. As a result, the dApp reward allocation has been reduced by almost half, so let’s apply the v2 mechanism in this situation.
The result is as follows.
If we leave it as is, it will not reduce the excessive rewards for top dApps that we wanted to improve in v3, so we will need to devise a new way to do so.
Simply setting a staking limit is not a sufficient countermeasure, so if the Tier system is to be eliminated, it would be appropriate to revise the fundamental calculation formula.
We added two types of functions to the simulation.
- Current: Current calculation method (Tier system)
- Linear: Calculation method for v2 format (distributed by share ratio)
- Irrational: Reduce the difference using √
- Irrational+Sigmoid: Add a sigmoid function to 3 to put some limits on the bottom and top.
Graphically, the shape of 3 looks good, but allocating the current reward pool as it is would appear to be too much. Therefore, a calculation would need to be made that would vary the reward pool.
3 Likes
Since we did not have time yesterday, we proceeded with the simulation a little more today.
First, the amount of reward for dApps should increase or decrease with the number of registered dApps.
There would be a way to use the current slot system (up or down depending on the price of ASTR) and also TVS.
This time, let’s assume that the lower limit is 10% when the number of dApps is 50 or less, and the upper limit is 100% when there are 200 or more. The remainder will be burned. Any remainder will be burned.
If we apply this to the current situation of dApp Staking, it looks like this.
Irrational and Sigmoid require a lower limit of 1.5M ASTR.
Since the parameters are only created in a way that is close to the current dApp Staking situation, more delicate adjustments should be necessary, especially for Sigmoid.
Since there is currently a lot of discussion about dApp Staking, it is better to decide on a policy first and then work out the details. The current situation is a mixed bag.
That’s a great idea. I’m not sure if I’ve missed anything, but may I ask for some additional information? What’s the current situation regarding this matter? Thank you very much. =)
Thanks for your research @you425.
Has the Foundation align with these ouputs?
The ranked tier system was recently implemented, but it should not have reflected the parameter changes I suggested.
To avoid complicating the formula, I currently recommend changing the parameters related to slots.
If I rerun the simulation with current ASTR prices, this is what I get.
- Current: Current Parameters
- Slot Param Changed: Parameter changes I proposed
- Linear: v2-like (footprint at Tier 4 threshold)
The last proposal is deprecated at this time because it would be a complex change that could be called v4. It is only a new calculation method created based on the assumption that if we were to stop Tier.
Currently, it would be easiest and make the most sense to adjust the parameters of the slots to increase the small rewards of Tier 4.
However, I think that would require the community to redefine what dApp Staking should be.
Is dApp Staking support or reward for developers?
- Support: too little for Tier 4 in its current state?
- Reward: maybe the current state is fine?
2 Likes
There is nothing specific other than the implementation of ranked tier.
This implementation is also very recent, so we will need to discuss this further.
Very interesting to see the ranks within each tier, and it makes more logical for the ecosystem in general, I assume. Curious to know more about how the ranks are determined each tier.
It is a minor one, but for the last column in the table, ‘Increase’ on ‘Linear(v2-style)’, in general they are decifit from the current rewards point of view rather than surplus?
@pithecus
For ranks, the difference between the thresholds for each Tier, as previously defined, is divided into 10 parts.
To give a simple example, if the threshold difference to the next Tier is 100M, every 10M ASTR gained will increase the rank by 1.
When you are at rank 10, you have achieved the threshold for the next tier, so you will never be at rank 10 unless the next tier slot is filled.
Yes, this is the increase from the current compensation.
If we take a format like v2, where everything is distributed according to the share of staking (but with a minimum threshold set equal to Tier 4), this is what we would get.
This is the same bad thing as v2, so I am only putting this up as a reference.
If the Tier system is to be dropped, it needs to be designed more closely to avoid excessive compensation.
Examples given previously:
1 Like
Hi @you425 thanks for the details!
Regardless of the type of weighting (irrational, sigmoid), it seems beneficial for the dApps not too much to concern about coming back and forth when the amount of ASTR staked is at the edge between two tiers. And perhaps it could lower the barrier for the new devs / teams to join the Astar Network ecosystem.
For the dApps on the Tier 1, then there is no upper limit, so perhaps the upper limit could be either the highest staked ASTR or something dynamically adapting to the price as well.
Excellent work, thank you.
There are already some things that need to be changed in the tier system. For example, 10 Million astr is vital for tier4, while it doesn’t make much difference between 2 projects in tier2. We need to put more thought and consideration on this side.
@pithecus
As for tiers, I guess opinions will differ again depending on perspectives.
If there is a tier, the goal becomes clearer and you may be more motivated to improve, and on the other hand, if there is no tier, you may work harder to get as far as you can. Also, Tier might make it easier to understand the degree of popularity depending on the user.
Either way, eliminating tiers would require major changes and should be discussed carefully.
Yes, Rank is only available up to Tier 2, there is no Rank in Tier 1. you get the maximum reward when you achieve Tier 1.
@MrKarahanli-Emre
As you say, the higher the Tier, the greater the difference in thresholds, and therefore the greater the ASTR required to achieve the next rank.
This is a logarithmic function concept, meaning that the less powerful the dApps (lower Tier), the easier it is to increase the reward.
The current system divides ranks into 10, so it is possible to deal with this by making it more detailed (e.g., 100 divisions?). However, this is a symptomatic treatment and may not be a fundamental solution.
If this request is to be fulfilled, it would be better to eliminate tiers. However, as I commented earlier, this would require a major change in the system and should be considered carefully.
Hi @you425 ,
From my point of view, having ranked tier system (having both, if I understand correctly) makes more sense to motivate dApps in general. Or at some point, gradual steps might be taken into account, having them both and later the rank system eventually. But again, then it should be definitely discussed carefully.
By the way, can you, by any chance, refer any reference for the recently implemented ranked tier system?
Thank you very much. Looking at the information, I gained a lot of knowledge. =)
1 Like
Hi @pithecus !
Come to think of it, it doesn’t appear to be in the documentation yet.
As for the logic, it is possible to learn about it from a post on Forum.
1 Like