If you want to maximize the number of ICP you have in total (regardless of whether they are staked or not), then you should lock up your ICP in staked neurons with the longest dissolve delay possible. This will guarantee that you will get the best return on your ICP.
If you want to maximize the number of ICP you receive in 8 years (assuming no compounding), then you should set your dissolve delay as high as possible, 8 years in this case, and immediately click dissolve. This is better than a 4 year dissolve delay, keeping it locked for 4 years (to accrue age bonus), and then unlocking for the last 4 years. Any other combination of holding and dissolving is still inferior to setting your dissolve delay as long as possible. The dissolve delay bonus is better than the age bonus, so it seems to win out every time. Always lock up your ICP for as long as possible to maximize returns.
The really interesting question is how compounding affects returns, specifically broken down by locked ICP vs liquid ICP. By compounding I mean reinvesting ICP staking rewards in staked neurons. There are 4 different scenarios we will look at.
Each condition assumes 100 ICP are staked with an 8 year dissolve delay.
1. Locked up compound: Locking up your ICP with an 8 year dissolve delay and compounding, meaning reinvesting your ICP rewards into your staked ICP neuron.
2. Locked up no compound: Locking up your ICP with an 8 year dissolve delay and not compounding, meaning you withdraw your ICP rewards and do whatever you would like with them.
3. Dissolving compound: Locking up your ICP with an 8 year dissolve delay, immediately clicking dissolve, and then compounding, meaning reinvesting your ICP rewards into your dissolving ICP neuron.
4. Dissolving no compound: Locking up your ICP with an 8 year dissolve delay, immediately clicking dissolve, and then withdrawing ICP rewards to do whatever you would like with them.
Let’s jump straight into the results! The table below shows how many ICP are still locked/staked in a neuron and how many ICP are liquid across the 4 conditions at the end of the 8 year period. The total is the sum of the locked and liquid ICP.
Locked up compound: Notice how locking up your ICP in a neuron with an 8 year dissolve delay, never clicking dissolve, and compounding all ICP rewards back into staked neurons will result in the most ICP. This makes sense because the NNS was designed to reward those who have a longer time horizon on their staked ICP. Also note that all of the ICP in this condition are still locked in an 8 year dissolve delay neuron, so even though it results in the highest amount of ICP, you can’t access any of it as there aren’t any liquid ICP. And you would have to wait 8 years for that ICP to dissolve before you could access any of it.
Locked up no compound: Looking at the “Locked up no compound” condition really illustrates the power of compounding. The “locked up compound” condition resulted in 2.65x the returns (454.1 / 171.3) of the “locked up no compound” condition. This is a pretty significant compounding effect. Also note that in this condition we have 171.3 liquid ICP and 100 staked ICP, which is more liquid than the “locked up compound” condition, but at the expense of a large chunk of returns.
For simplicity I used a daily compounding model rather than an every other week compounding model (which is about what you would get with an 8 year dissolve delay neuron with 100 ICP right now). So the real compounding effect would be slightly lower than what I am quoting here. Although, the NNS team is working on an automated compounding feature which might make this analysis a little bit more realistic (if they implement daily compounding that is).
Dissolving compound: Interestingly, the “dissolving compound” condition results in the most liquid ICP. All of your ICP is liquid because your neuron is dissolving and you are continuously adding reward ICP back into your dissolving neuron. So in the event that you want to maximize your liquid ICP 8 years down the road, you want to set your dissolve delay as high as possible, immediately click dissolve, and then compound your reward ICP back into your dissolving neuron. You will be sacrificing some rewards to have liquid ICP (rewards are 2.05x if you keep it all locked up), but in the end all of your ICP will be liquid.
Dissolving no compound: Finally, the “dissolving no compound” condition is predictably the lowest yielding condition overall, but results in the second most liquid ICP.
If you want the most ICP and don’t care if you have access to it or not, use an 8 year dissolve delay and reinvest all of your ICP rewards back into your staked neuron.
If you don’t want to compound, and you want the most liquid ICP in 8 years, it still makes the most sense to set your dissolve delay as high as possible and immediately click dissolve. Dissolve delay reward > anything else.
If you want to maximize the amount of liquid ICP in 8 years, stake your ICP with a dissolve delay of 8 years, then compound your ICP rewards back into your dissolving neuron.
Finally, even if your timeframe is 2 years, 4 years, 20 years, or anything in between, these same principles still apply, although as your timeframe gets shorter you will see less of a difference between conditions.
Want to dive deeper into my analysis? I created a nice spreadsheet to share. Feel free to make a copy and play around with it. Let me know if I missed anything or if there is another analysis you wish you could see.
If you liked this article, you might also like my fun analysis on seed round ICP, and why we might not see the price stabilize for a few more months.