Page 1 of 1

Misleading profitability metrics for previous two days

Posted: Thu Nov 22, 2018 9:55 am
by Steve Sokolowski
Some customers have submitted support tickets about the "1MH/s expected payouts" chart over the past two days.

The "automatic" line is calculated by dividing the total amount earned by all miners by 24 hours. Since the system was offline due to the scheduled database maintenance during part of each of those days, the "Automatic" line is lower than what would have actually been earned during any particular part of the day.

The anchor coin lines, like "LTC" or "Dash," are estimates of what other pools could earn by mining those coins alone, so they are not affected by whether the system goes offline.

This is a known issue that would be complex to resolve, and therefore we decided not to devote effort to resolving it because it only causes issues on days when there is system downtime like this.

Re: Misleading profitability metrics for previous two days

Posted: Thu Nov 22, 2018 10:55 am
by CSZiggy
Steve Sokolowski wrote:The "automatic" line is calculated by dividing the total amount earned by all miners by 24 hours. Since the system was offline due to the scheduled database maintenance during part of each of those days, the "Automatic" line is lower than what would have actually been earned during any particular part of the day.
Can you elaborate the ALL MINERS part. Just to clearify, when we look at the scrypt 1MH/s expected:

1. Is that ONLY those members on the AUTOPOOL(i.e. the pool auto switches them based on its algorythm)?
2. Solo mining is NOT included as they are not autopool.
3. Static miners (not auto switched from coin to coin by the pool) are NOT a part of those readings?

Feel like I've asked this question before and it keeps getting blown off.
Just to make sure those numbers represent the numbers they say they are.

Re: Misleading profitability metrics for previous two days

Posted: Mon Nov 26, 2018 1:38 pm
by CSZiggy
Bump.
Still looking for confirmation please.

Re: Misleading profitability metrics for previous two days

Posted: Mon Nov 26, 2018 2:34 pm
by Steve Sokolowski
CSZiggy wrote:
Steve Sokolowski wrote:The "automatic" line is calculated by dividing the total amount earned by all miners by 24 hours. Since the system was offline due to the scheduled database maintenance during part of each of those days, the "Automatic" line is lower than what would have actually been earned during any particular part of the day.
Can you elaborate the ALL MINERS part. Just to clearify, when we look at the scrypt 1MH/s expected:

1. Is that ONLY those members on the AUTOPOOL(i.e. the pool auto switches them based on its algorythm)?
2. Solo mining is NOT included as they are not autopool.
3. Static miners (not auto switched from coin to coin by the pool) are NOT a part of those readings?

Feel like I've asked this question before and it keeps getting blown off.
Just to make sure those numbers represent the numbers they say they are.
All of those answers are correct.

Re: Misleading profitability metrics for previous two days

Posted: Mon Nov 26, 2018 5:23 pm
by CSZiggy
OK question. Since the bot-miner/constant coin miners are still on the system, I am guessing they are still making enough of a profit to bother doing it. When you buy/sell coins generated by those miners, are they being bought-sold at the same prices the automatic pool peoples coins are? Or do they get their own price points?

I obviously don't have access to the data the pool admins do, I have asked for the data to be reviewed/looked at/shown for transparency sake.
This is how I see it, please let me know where I assumed something wrong:

(MADE UP EXAMPLE WITH FAKE NUMBERS TO ILLUSTRATE)
I mine with my 30 machines, I make $100 in the autopool mining digibyte and all the other coins the pool assigns me. Those coins get sold, the pool gets $100, I select LTC as my payout and those coins cost $50, so I get 2 of them. The next day nothing changes, same numbers, except Johnny comes along and puts his 30 machines up to mine and decides he can try to get more profits by picking the top coins each time they flop. Now due to stales and orphans on the coins he is mining he only gets half as many shares, but each share is worth 4X as much at least. 100x4=400/2 = 200. Johnny also wishes to get paid in LTC, so you go out to the exchange and sell his coins...which were being paid 4X more then they were worth(should look at what you pay the bot-mined top end coins on the blocks found readout page vs how much they are worth on exchanges I've seen like 10X-100X more) so when you sell his $200 worth of coins you only get $50 in cash to buy LTCs with? The issue I see is the pool is not taking $150 out of its pocket to pay this difference each and every time it happens. Which means when you sell all the coins you end up with $150 from both miners. at $50 each you get 3 LTCs for the payout instead of 6. Now instead of paying $50 for each coin don't you have to charge us all $100 for each coin? so now instead of my 2 coins I only earn 1? instead of Johnny's 4 coins he only gets 2. Which means 50% less payout than someplace else, or even same place if we were all mining the same amount for the same price?

This is a totally exaggerated example but shows why I do not think the autopool 1MH expected profit payout on this pool is as high as it could/should be. Even if Johnny only made 1% more than the people on the autopool, doesnt the fact he got more per coin than they sold for change the payout price everyone else has to pay for that same coin as a payout coin? The difference between here and the best pool being 10%, is there 1/10 the hash from top-coin miners vs the autopool miners that are selecting that SAME payout coin?

If the algos were not allowed to cross algo coins for payout this might be limited. If this was only happening on scrypt for scrypt payouts it would be limited. But if any algo(x-11/equihash/sha) can do this same-mining the top coin get paid in LTC doesn't that compound the issue further for everyone else especially those in the autopool selecting that same payout coin?

If they were only allowed to solo mine the top coins, then they would be responsible for trying to sell and get their 3-4X payout themselves. If they aren't actually making a profit, then isn't it in the pool's best interest to force them into autopool mode by disabling static mining that isnt in solo mode to force their profits to rise and collect more of the 5% pool fees from them? The fact they continue and are allowed to mine the top coins says to me they are making a higher profit at everyone else's expense. They are robbing the Peters to pay the Pauls.

So can you run some reports/grab some data for us/at least do an internal review with the real numbers?

1.Of the all the people on static PPS across all the different algos which coins are they getting paid out in the most?

2. How much hash, say just on scrypt is from static PPS mining vs autopool PPS mining(im guessing they arent popping just a single miner at a time on to do this, more like 100 or a meaty nicehash contract worth of hash) that are selecting the same coin, say...LTC as a payout?

3.How much more if any is the percent of the profit of the static PPS miners vs the autopool PPS miners when you look at their 1/MH expected?(again if its lower then shouldnt it be disabled and force higher profits so the pool makes more off their 5% fee)

4.Are you selling the static PPS coins for what you are crediting those miners for them or is there a loss the rest of the pool selecting their same payout coin is making up for?
If you re-ran the data, removing the static PPS miners from the equation does that affect the profitability of the 1/MH expected chart?

5.Any chance the 1MH expected could be changed from $ to satoshi or show satoshi along with the $. The price of coins goes up or down but the number we get paid each day stays the same, so showing a higher payout $ doesn't really mean you are earning anything more than the day or week or month before. If the payout coin percents increase or decrease that might offer a more representative data number as to increased/decreased profitability.

Re: Misleading profitability metrics for previous two days

Posted: Tue Nov 27, 2018 12:52 am
by CSZiggy
???

Re: Misleading profitability metrics for previous two days

Posted: Fri Nov 30, 2018 4:05 am
by CSZiggy
Would like to understand what I am guessing wrong on.
Any comments?

Re: Misleading profitability metrics for previous two days

Posted: Fri Nov 30, 2018 5:54 am
by CSZiggy
??

Re: Misleading profitability metrics for previous two days

Posted: Fri Nov 30, 2018 7:06 am
by Steve Sokolowski
I apologize for not getting to this issue right away. This is very complex, so it will take some time to understand.

There are a lot of other issues and tickets at the moment and I plan to get to this as soon as possible.

Re: Misleading profitability metrics for previous two days

Posted: Fri Dec 07, 2018 1:39 pm
by CSZiggy
BUMP - It's been another week.
I know you were busy fixing the WAMP and getting E3 mining going for the past week

Now that those jobs have been completed, you said you would be going back to review this.
When will that be reviewed? If Constance's job with the miner groupings has been completed, maybe she has a spare moment to review this
and give you a quick summary/rundown with the data numbers she will have access to.