https://www.dropbox.com/s/6i4nu2mjviw3y67/Captura%20de%20pantalla%202019-04-13%20a%20las%206.35.30.png?dl=0You will see the Xfox currency and its estimate of more than 7000 coins in a day (rigti 1080ti). But it's wrong, I think the formula of your profit is not very accurate and I can say it because we measure and test all the currencies / pool and elprofit is between 0.70 to 0.95 90% of the time, but it's not what same mine to 0.80 that to 1.
In this case, this Xfox coin has a blocking time of 2 minutes and has a reward of 2.79 (removing fee from the pool), the maximum coins per day is from 2008, it does not matter how many machines and the hash is the same because the difficulty it adjusts itself, the maximum is 2008 coins per day.
With which his estimate is completely wrong in this currency and slightly deviated in many others, but it is no longer a question of data, it is a question of formula.
The formula itself is correct and gives the exact same results are WhatToMine and CoinCalculators in general. There can however be specific coins that do something differently where you manually need to adjust the profit factor as it's not possible to compute the number of coins per day the standard way.
You must take into account block time, reward and nethash, to have a more accurate profit. It is a more complex formula, more data to take into account and we must take into account the maximum coins per day for the estimates offered by AM
It is only a constructive criticism, we customize all fields, difficulty, recommensa, price of exchanges directly (that's why it would be good to cache results between requests)
It's just an example where the formula is not right. Neither can it be to mine ZEL when its real profit is 0.65 and so many currencies. You imagine mining ZEL to profit 1, but then only receive 65% of the mined. That's why I recommend everyone to do pool and coin tests, if you change the pool you have to repeat the test.
Yes, the plan is to add support for the "NetHash" way of doing the calculations as well, where Block Time is one of the factors. The idea is also have this configurable per coin so if you run into a coin where the standard way of doing the calculations doesn't give the expected results, you should be able to switch to this other mode.
Last detail, the benchmark. When I do a 2 or 5 hour test, what value does it give?
The last value he had at the end of the test
The highest peak of hash
The average of the hash of those 5 hours (This would be the correct one)
@patrike should average the benchmar hash, if not, the benchmark is totally useless for AL-GOS that change subalways, it just is not worth it.
@PAtrike is there no way to average the hash value of 2, 5 or 10 hours, during a benchmark?
It should be the average but excluding a few samples in the beginning as those first samples are typically too low. However, I did notice an issue here with long running benchmarks so I will make a correction for that case. Thanks for finding it.
Update - coin properties:
I've just finished the implementation for Network Hashrate and Block Time in the Coin Properties, together with a setting where you can select one of the following calculation methods:
- Difficulty, Block Reward and Exponential factor
- Network Hashrate, Block Reward and Block Time