The total staked ADA
In a year, there are 73 epochs of 5 days on the Cardano blockchain, and approximately 21600 blocks are minted per epoch by staking pools based on a slot coefficient (one slot = one second) of 5%. Meaning, on average, every 20 seconds, a block is produced. The actual number is usually lower due to height and slot battles; you can read more about it here.
The number also varies between epochs due to the randomness of how a staking pool gets to produce blocks. Staked ADA only gives a threshold number that the pool needs to be lower than in a randomized function, meaning the number of blocks the staking pool produces in an epoch will vary around an average number. Historical data of blocks minted can be found here and gives an indication of how many blocks are actually produced in an epoch by all pools.
Given an average of around 21000 blocks per epoch, and 24,5 billion ADA staked, you would require around 1.16 million ADA in a pool to produce a block on average in an epoch. If 30 billion ADA were staked you would need around 1.43 million ADA to, on average, produce a block. As you can see, this influences the rewards by requiring more ADA to produce blocks if the total staked amount of ADA goes up.
The fees collected and rewards available.
ADA is distributed from the reserve (currently around 28 million ADA per epoch) during every epoch, also fees are collected from transactions each epoch. You can see historic data for issued supply here. The fees vary depending on the number of transactions, the value size, and the byte size of each transaction. You can see a historical overview of transactions here with a ballpark figure of around 0.3 million transactions per epoch with about 81000 ADA in fees. After that, a 20% treasury is collected for governance (estimated at approximately 5.6 million ADA). As you can expect, the more transactions and fees collected, the more rewards are distributed. The reserve issuing rewards lowers as time passes, affecting the total rewards available.
Pledge influence factor and saturation point.
The pledge influence factor (see this article) affects the probability of a pool producing a block, and in general, the higher the pledge, the more rewards. Saturation points depend on the desired number of pools (so-called K value and the total ADA in circulation - see this blog post by IOHK.) In general, the more saturated you are, the more stable your block production is and also the less effect the minimum cost of 340 Ada (or higher) per epoch a pool has. Currently, you can have around 70 million ADA before a pool is saturated. NBX also has a relatively high pledge with 350000 ADA currently. There are ongoing governance discussions in the Cardano community about further increasing the pledge influence factor and introducing other methods of pledge calculations, such as in the Cardano Improvement Proposal 7 suggestion.
Combining it all in a random simulation
To look at all these factors together and simulate the randomness of block production across multiple scenarios and calculate an average, you can use the site here with the complete formula explained for staking rewards. By entering the pool ticker ANP you can see we fall in the range between 3% (very unlikely), 3,5% (more likely), 4% (most likely), and 4.5% (very unlikely). Also, the more saturated a pool is the closer the numbers will reflect each individual delegator’s rewards compared with the pool estimated rewards. The pool has historically done around 4% rewards but recently has trended somewhat down from this. Interestingly enough, this also seems to be most pools in Cardano and could be due to fewer overall blocks produced and lower overall rewards distributed, however this data needs to be viewed over time. To complicated matters, you can add a compounding effect to delegators that will vary depending on a pool being lucky with an early high number of blocks to reward delegators. Your compounding effect will be more significant if you get more rewards early. Another factor that could be more important in the future is governance, where a portion of fees could be diverted toward governance instead of staking returns.
Hopefully, this explanation has been helpful in showing some of the factors that go into rewards on Cardano staking pools and why this has to be an estimate based on blockchain data. We at NBX look at theoretical data and actual pool performance to try to give a good estimate to our staking customers on expected rewards. Because of the uncertainties, we are careful in giving an exact number. We follow the development of the staking returns and will adjust yearly our estimates based on the data we collect from actual pool performance.