Quote from: eminer001 on February 20, 2018, 09:37:44 AM

Quote from: hashrefinery on February 20, 2018, 06:26:57 AM

Quote from: GoRdiE on February 20, 2018, 01:20:09 AM

Hi all,

what are you using? 24 hour average estimate or 24 hour average current? Or maybe only current? and why? Someone is doing some test?

Thanks to all Grin

With yiimp based pools I use the current estimate. The 24 hour estimate is an average of the previous 24 hours. Yiimp does not attempt to predict the next 24 hours (if it could do that with any accuracy we would all be rich

).

As they say in the financial services industry, past performance is not indicative of future results.

I'm sure others will have different opinions.

Are you sure that mining current profit is better than 24h?

Absolutely.

The current estimate is based on all the information the pool has currently, such as difficulty of the coin(s) being mined and the latest exchange rates to BTC (if applicable). That information will be out of date very quickly because both difficulties and exchange rates change constantly but the pool will continually adjust the current estimate using the latest info.

The 24hr value is an aggregate of estimates that older than the current one and up to 24 hours old. It is specifically NOT taking into account current difficulties or exchange rates.

This is a simplistic example that assumes difficulty does not change:

The pool might be mining two coins on the same algo - A and B.
A is the most profitable coin and is mined most of the time, but B is popular with payouts and left on the pool primarily for that purpose.
The value of A is normally 2x the value of B.
Perhaps there is some scandal involving coin A and the exchange rate drops dramatically making B the more profitable coin on that algo.
The current estimate will immediately drop by 50% because the pool is now mining B.
The 24 hour estimate will gradually drop over the next 24 hours until it also reaches 50% of the starting value.

If the mining software is choosing the pool/algo based on the current estimate it would probably swap to something else within a few minutes. If it were using the 24h figure it may take several hours before the average value drops to a point where the rig switches.

Quote from: eminer001 on February 20, 2018, 09:37:44 AM

I did the test with 2 identical rigs using 2 separate btc addresses and it turned for me that current statistics had lower BTC income after 24 hours or 48 hours than the rig with 24h statistics. Anyone else tested with 2 identical rigs and separate payout addresses for accuracy ? All you have to consider is the actual BTC amount from your payout address, not current statistics that have nothing in common with real income and all you care is the amount you are getting paid, that is important.

I won't dispute your findings - let's just agree there are a lot of variables

Quote from: eminer001 on February 20, 2018, 09:37:44 AM

Pools are adjusting difficulty and also many coins have a lot of variation in difficulty, but the low difficulty that you see in statistics lasts just a few minutes and after that is decreasing a lot because of spiked hashrate. The Pool might not find any coins with low difficulty and also pool can crash because of the spike.

This is actually a separate issue but one worth mentioning - a big enough spike of new miners connecting can absolutely break or degrade a pool. There is nothing in the API currently that indicates the how busy the pool is as a percentage of total capacity or similar. The Awesome Miner dev is putting in some features which should help detect when this is occurring and temporarily disable to pool until it recovers.