I would argue it's not even close enough. See, if you run a 144-sample on the Erlang distribution function, you would see how variance is very likely at numbers you might have thought were not possible. For example, there is a 26% chance that the network finds 5% or fewer blocks, this happens once every 4 times. That's a lot. There's about a 40% chance that the network finds 2% fewer blocks, all without a single miner dropping out of the network. Almost a 10% chance exists for 90% of blocks or more. So it's very likely that you look at the charts or the RPC calls and think, "Oh, the hashrate today dropped by 10%," when it has not.

The point I am trying to make here is that nobody knows what the hashrate is -- it's all just guesswork using the best tools we have at our disposal. You look at the chart posted above, it's a perfect representation of variance. These tools use the same logic you mentioned: assume the network should find x blocks per n period, and then make assumptions based only on that factor. Since they apply this on a small timeframe, like hourly and daily, the results are far off. We don't actually get 50 EH drops daily -- it's not possible, it's just variance.

Now, if we were to take the entire epoch of 2016 blocks and the difficulty drop we witnessed, the numbers would differ a lot. The drop was 7.4%, which is huge. The chances of that happening are just 0.02%. While that still is a possibility, it's more likely that some serious hashrate went offline during that epoch. Of course, it's by no means 7.4% worth of actual hashrate -- because, for all we know, it could be just 3% worth of hashrate and the rest is just variance, which is expected to happen at 3%. Or maybe it's a 4% drop and 3% variance, which is 8% likely.

Was this a large mining farm in Iran or a few large Texas players throttling back for the summer heat? We don't know what we don't know -- but we know that, based on statistics and probability, there was (very likely) a large shutdown somewhere in the ecosystem.


Hashrate is just a theory; nobody knows what the total hashrate is. In reality, no pool knows their hashrate with 100% accuracy. When a miner connects to a pool, the pool will try to "figure out" its hashrate based on the shares they submit at a given pool difficulty, and even then, the hashrate isn't exactly as it is on the actual miner. If you are wondering about a pool's hashrate, then there are two ways to figure that out.

1. The reported hashrate:

Many pools report their hashrate (again, based on the shares their miners submit). this could be inaccurate and misleading since the pool can report a lower number to show a better luck status or report a larger number to show dominance. It's hard to verify and pretty useless unless you are sure the pool operated is somewhat honest

2. How many blocks do they find at x difficulty?

This is the most accurate way since the difficulty is a part of the blockchain; it's public info and can be verified fairly easily. If you have time, difficulty, and number of blocks, you can calculate (or rather estimate) hashrate, but again, even that isn't 100% accurate and is very subject to variance.

Hashrate = Blocks solved × Network Difficulty × 2^32 / time in Seconds


If you take Foundry Pool as an example, they found 23 blocks in the last 100 blocks; that's 192 EH. However, their estimated or reported hashrate is 250 EH, which means without variance they should have solved about 29 blocks, so they are short 6 blocks. Since hashrate is a function of time, Foundry can claim they own 500 EH but it doesn't matter. The truth of the matter is their hashrate for the last 100 blocks is 192 EH as far as the blockchain is concerned.
​

​