Compiling data about historical execution block rewards in an attempt to provide an expectation of future rewards.
A previous article examined Ethereum execution block rewards. However, in a rapidly evolving and fluid environment, and with the move of the execution chain to proof of stake imminent, it is a topic worth revisiting.
Note that the background to block rewards is covered in the previously-mentioned article, which is worth reading to obtain an understanding of the terms used here.
Before evaluating block rewards, data needed to be gathered from the blockchain1. The start point for the data gathering was the London hard fork, which occurred at block 12,965,0002. The London hard fork introduced the current system of burning a portion of transaction costs, so this point is the earliest time at which the data can be considered relevant. The data was gathered to block 15,417,2833, and so this formed the end of the data set. For each block from 12,965,000 to 15,417,283 the following values were obtained:
To be clear: this information did not include direct payments made to the miner with non-Ether currencies such as ERC-20 tokens, nor any information about possible indirect payments made to the miner. As such, this was likely to be an underestimation of the total rewards paid to miners but it is impossible to say by how much due to the inherent nature of indirect payments (as explained in the previous article).
The raw data set consisted of 2,452,284 entries. 59,783 of these were empty blocks, with no transactions in them, and as such were considered irrelevant for this analysis and discarded. All blocks with direct Ether payments greater than 100Ξ were examined manually to understand the purpose of the payment, which excluded an additional 21 blocks where the payments were between different accounts of the miner which do not count as block rewards. This resulted in 2,392,480 blocks used for subsequent analysis.
The simplest figure to calculate is the average rewards per block. Taking the sum of the rewards and dividing it by the number of blocks results in a value of 0.244Ξ. However, rewards can vary significantly between blocks, and because proposing a block is a relatively rare occurrence it is important to understand the distribution of rewards rather than look at a single value. To calculate an initial distribution the rewards were placed in logarithmic buckets.
Displaying this information graphically gives us the following:
The graph highlights the fact that the vast majority of rewards are less than 1Ξ. For a closer look at the distribution in this range, we move to a linear scale, with each bucket having a size of 0.01Ξ.
Again, it is clear that blocks with higher rewards occur less frequently. The pertinent question is thus: "how likely am I to obtain a block reward worth at least Ξ?" This can be answered through use of a cumulative probability graph.
For example, from this graph it can be seen that approximately 80% of blocks had a reward of 0.25Ξ or less, so the answer to the question "how likely am I to obtain a block reward worth at least 0.25Ξ?" is "roughly 20% of the blocks you propose will be worth at least 0.25Ξ."
The information presented above shows rewards for blocks over a relatively long timeframe of approximately two years. An obvious question to ask is whether there is a trend in block rewards over time, or whether the rewards remained consistent throughout the data analysis period.
To explore this, block rewards were grouped on a daily basis and, for each day, an average reward per block was calculated. The day-to-day mean is shown below:
There are six areas where the daily mean block reward was significantly higher than the average. Investigating these times manually, these peaks can be ascribed to:
Ignoring these outliers (even though all but one were situations where miners did receive these rewards), an average block reward somewhere above 0.15Ξ looks to be achievable.
Validators are rewarded for their actions on the consensus layer securing the beacon chain (for further explanations please see the previous article). These rewards are defined in the beacon chain protocol, and are obtained using a relatively simple calculation4. These consensus layer returns, which are rewards expressed as a percentage of the 32Ξ validator deposit, are shown below:
At time of writing there are around 420,000 active validators, which equates to maximum returns of around 4.5%5. The reduction in returns as new validators become active is relatively gentle (for example doubling the number of active validators would result in maximum returns of around 3.2%).
For execution layer returns we take the average rewards per execution block of 0.244Ξ as calculated earlier in the article, and calculate the number of blocks each active validator is expected to produce in a year. This gives a total Ether reward based on the execution layer, and from this returns can be calculated as shown below:
This line shows a significantly sharper decline in returns as the number of active validators increases. This is because the number of blocks produced in a year is constant regardless of the number of validators. To contrast with the consensus layer returns, current projected returns are around 5.1%, falling to around 2.6% if the number of active validators doubles.
Combining the returns from both layers provides a projection for returns after the merge.
It should be noted that this is a projection, and, as mentioned above, many variables may cause this value to change over time. It should also be noted that these are the projected returns for the network as a whole, and will be unlikely to be reflected in the returns for any individual validator. When considering an individual validator, their returns are much more likely to trend to the median rather than the mean. The median reward of 0.08Ξ is a significantly lower value than the mean at 0.24Ξ and so corresponds to a significantly lower set of post-merge returns:
If a single block is expected to provides rewards around the median, and across all blocks the expected reward is the mean, this allows us (with a little creative license) to use these values to provide an idea of the range in which most stakers can expect to see their returns. Plotting this information, displayed as a percentage increase in returns that a validator may expect to see after the merge, is shown below:
So for the current number of active validators (around 420,000) a single validator would hope to see an approximate 30% increase in returns from today's values (from around 4.5% to around 5.9%), possibly ranging up to a 100% increase (from around 4.5% to around 8.7%) if they have a large number of validators.
Although these returns will remain dominated by the luck of how often each validator proposes a block, and the value of the transactions within it, over time validators would expect to see values in the ranges shown by the above graph. Further changes to the consensus layer, such as the introduction of withdrawals, could see a reduction in the number of active validators with a corresponding increase in annual rewards. The on-going growth of layer 2 protocols will also see Ethereum blockspace become more contested, with the expectation of higher rewards as a result. Validating is a solid method of obtaining returns on Ether whilst helping to secure the chain, and looks to continue to be such.