[ANN][MULTI] HashBros Beta [OPTIONAL-PS][VD][1-1.5%][PPLNS]

[Sovietaced]

Thanks to our new algorithm tweaks, earnings have truly increased quite a bit. Once we stopped mining slightly higher diff coins we got much more consistent payouts. Also having the engine look at the profit history of coins instead of the instant profit value has given us more reliable switches. On top of that we have been recording the purity of a coin, aka how many orphans/blocks mined. We get a truly better understanding of the real value in a coin.

Give us some more time and things will be even better. Daily estimated earnings will be up soon as well.

[1714964547]

1714964547

[1714964547]

1714964547

[Sovietaced]

In other news payouts are now redeemable. We raked in 1.5 BTC today, good job bros.

[chilin_dude]

Going to have pool/user wide profitability per MH on each round/day shortly. We're collecting worker, round, and pool data as we speak. Going to be displaying it soon!

sounds great!

pool metrics are not public but are working, we also have profitability estimates for each rounds graphed too. I just wanted to say that the pool profitability has increased drastically since we stopped trying to mine coins with higher diff. We cutoff coins that have an estimated block time of more than 5 mins and we are no longer wasting time on LOT, NET, PHGS, ETC and the earnings show very well.

Sounds very good that profitability has increased drastically and that you are logging pool metrics, any ETA when they will go public?
And is there any plan on when you would reimplement these higher diff coins when the pool gets to X amount of mhs?

[Sovietaced]

Going to have pool/user wide profitability per MH on each round/day shortly. We're collecting worker, round, and pool data as we speak. Going to be displaying it soon!

sounds great!

pool metrics are not public but are working, we also have profitability estimates for each rounds graphed too. I just wanted to say that the pool profitability has increased drastically since we stopped trying to mine coins with higher diff. We cutoff coins that have an estimated block time of more than 5 mins and we are no longer wasting time on LOT, NET, PHGS, ETC and the earnings show very well.

Sounds very good that profitability has increased drastically and that you are logging pool metrics, any ETA when they will go public?
And is there any plan on when you would reimplement these higher diff coins when the pool gets to X amount of mhs?

Higher diff coins will be considered automatically. We just set the threshold lower so the estimated block time has to be less than 5 mins. Once the hash rate goes up higher diff coins will have lower estimated block finding times. I suspect the metrics will go public tomorrow, HashBrosB is working on it and styling it up. We're just both in school so somedays we have less time that others.

In other news rounds aren't being reported properly for some coins. I will look into it when I get home from work.

[Sovietaced]

Had some really awful issues yesterday where worker credits weren't being updated. This happened due to a timeout when calling the REST API of the coin nodes. The REST API call runs some rather expensive queries and nodes were returning the query in ~ 100 seconds. Rewrote that so that it's much less DB heavy and now response times are about ~3 seconds. So that's fixed that issue.

We have some other issues where worker stats are taking a while to process but I have a pretty good idea of how to multiprocess a ton of stuff. Will keep everyone updated tonight. Metrics are looking good for tonight.

[Sovietaced]

METRICS ARE UP! http://hashbros.co.in/metrics

[spherix]

about a week since I joined.
I must say I'm very pleased with both the profits and the website.
I hope it stays for a long time.

kind regards

[Sovietaced]

about a week since I joined.
I must say I'm very pleased with both the profits and the website.
I hope it stays for a long time.

kind regards

It's only going to get better
Thanks for the support <3

[Sovietaced]

Over .01 BTC /MH!
 
Also SSL!

https://hashbros.co.in/metrics

[Sovietaced]

Greater visibility into Historic User and worker hash rates, access them by clicking the user icon and going to workers or by clicking on your username!

In other news leaderboards disabled because doing the calculations makes our web server timeout. We are going to be calculated the leaderboard stats in a scheduled job so that they will display very very fast. Also removing current hash rates from the leaderboards as that is a very expensive thing to calculate accurately.

[chilin_dude]

Keep up the improvements, the 24hour metrics are good, how about some longer term ones Ie 24 hour avg, 3day avg, 7 day avg?

[Sovietaced]

Keep up the improvements, the 24hour metrics are good, how about some longer term ones Ie 24 hour avg, 3day avg, 7 day avg?

We can do that as well.

I think i may have discovered a pretty big flaw in percentage of work calculation. I'm pushing a fix right now. We'll see how this influences earnings.

[chilin_dude]

Nice, I think you've got a brilliant GUI on the website atm, and 0.008 per mh atm is a pretty good base to start at - middlecoin which you're undoubtedly going to be compared to is only averaging ~0.01 at the minute, so we're only slightly trailing currently and this is still early stages of the pool.
Hopefully the potential flaw you spotted makes it even more profitable

[Sovietaced]

Nice, I think you've got a brilliant GUI on the website atm, and 0.008 per mh atm is a pretty good base to start at - middlecoin which you're undoubtedly going to be compared to is only averaging ~0.01 at the minute, so we're only slightly trailing currently and this is still early stages of the pool.
Hopefully the potential flaw you spotted makes it even more profitable

Well I just fixed the bug and you can see the share rate has exploded since we are now accounting for diff. What it will actually do is propagate earnings more properly. Our top hasher just lost percentage from 22% to 14%. So that means more earnings for me and you.

[Sovietaced]

Nice, I think you've got a brilliant GUI on the website atm, and 0.008 per mh atm is a pretty good base to start at - middlecoin which you're undoubtedly going to be compared to is only averaging ~0.01 at the minute, so we're only slightly trailing currently and this is still early stages of the pool.
Hopefully the potential flaw you spotted makes it even more profitable

Also, we lost hashes for about 45 mins last night so thats will kill our BTC/MH pretty hard.

[chilin_dude]

I've just pointed another couple of mh at you anyway, good feedback so far.

I was wondering if you had any input on the way you calculate profitability of certain coins, such as dogecoin. On your profitability index it is assumed the block reward is 500000.00 which is the same as dustcoin etc list, however I have read that this is the average block reward, and that each block reward is actually independent, and that the block reward is known BEFORE it is mined. Therefore by knowing if the current block reward is over 500000.00 you know it is more important to mine as more profitable, where as if the current block were only 100000.00 (5x lower) you would know it is 5x less profitable.
Another mulitmining pool on the up, wafflepool has written about this a fair few times, this being in one of his posts:

"This is the way most pools do it, and you're absolutely right that statistical averages work out to 500k/block (using doge for the rest of this example).  However, the block reward for the block you're _actively_ mining for is known when you're mining it.  It is based on the last known block hash.

For example (and its a really ELI5 style example - the actual code is much more complex).  But say the next block reward is worth whatever the last digit of the previous block hash was (assume 0-15 for last digit values).  Statistically you'd earn 8 coins for the next block, however, if we know for a fact that the last block hash ended in "F" (15), we know that the next block reward is definitively going to be worth 15 coins.  Which makes for a _much_ different profitability calculation for if the next block is worth mining.  Since the function for generating block rewards is known (in the client), and all the inputs are known (previous block hash, etc), we can use the _actual_ block reward, rather than the statistical average in our calculations.  Long term, this should give us better performance in earnings. "

Thanks.

[phzi]

Are you guys setup to predict the value of the next block for the various variable reward coins, yet?  (e.g. dogecoin, lottocoin)

Waiting for that before I give HashBros a try again, because calculating profitability based on average dogecoin reward make absolutely no sense at all.

[Sovietaced]

I've just pointed another couple of mh at you anyway, good feedback so far.

I was wondering if you had any input on the way you calculate profitability of certain coins, such as dogecoin. On your profitability index it is assumed the block reward is 500000.00 which is the same as dustcoin etc list, however I have read that this is the average block reward, and that each block reward is actually independent, and that the block reward is known BEFORE it is mined. Therefore by knowing if the current block reward is over 500000.00 you know it is more important to mine as more profitable, where as if the current block were only 100000.00 (5x lower) you would know it is 5x less profitable.
Another mulitmining pool on the up, wafflepool has written about this a fair few times, this being in one of his posts:

"This is the way most pools do it, and you're absolutely right that statistical averages work out to 500k/block (using doge for the rest of this example).  However, the block reward for the block you're _actively_ mining for is known when you're mining it.  It is based on the last known block hash.

For example (and its a really ELI5 style example - the actual code is much more complex).  But say the next block reward is worth whatever the last digit of the previous block hash was (assume 0-15 for last digit values).  Statistically you'd earn 8 coins for the next block, however, if we know for a fact that the last block hash ended in "F" (15), we know that the next block reward is definitively going to be worth 15 coins.  Which makes for a _much_ different profitability calculation for if the next block is worth mining.  Since the function for generating block rewards is known (in the client), and all the inputs are known (previous block hash, etc), we can use the _actual_ block reward, rather than the statistical average in our calculations.  Long term, this should give us better performance in earnings. "

Thanks.

Right so we don't calculate this internally and this is certainly something we wouldn't display on the public API for free. Since we've never had the ability to mine something like DOGE it's never really mattered to us because the DIFF is just too high. But that is definitely something to look into for lottocoin.

[Sovietaced]

Are you guys setup to predict the value of the next block for the various variable reward coins, yet?  (e.g. dogecoin, lottocoin)

Waiting for that before I give HashBros a try again, because calculating profitability based on average dogecoin reward make absolutely no sense at all.

We don't but we will in the future when it matters. Right now we don't mine doge or lotto because of diff. The estimated time to block time has to be less than 5 mins for us to even look at it.

[phzi]

Are you guys setup to predict the value of the next block for the various variable reward coins, yet?  (e.g. dogecoin, lottocoin)

Waiting for that before I give HashBros a try again, because calculating profitability based on average dogecoin reward make absolutely no sense at all.

We don't but we will in the future when it matters. Right now we don't mine doge or lotto because of diff. The estimated time to block time has to be less than 5 mins for us to even look at it.

I don't understand this logic.  Are you just trying to decrease variability?  The most profitable thing to mine is the most profitable thing to mine.  high diff increase your variance, but not your profits over time.