Topic: [ANN] profit switching auto-exchanging pool - www.middlecoin.com - page 521. (Read 829908 times)

Yeah. I don't believe it affects profits. But lots of people are saying it is. So I'm still thinking about it to see if I'm wrong.


With https://github.com/CryptoManiac/stratum-mining, I'm getting:

  File "/home/pushpool/stratum-mining-novacoin-testnet-cryptomaniac2/mining/DBInterface.py", line 33, in connectDB
    if settings.DATABASE_DRIVER == "sqlite":
exceptions.AttributeError: 'module' object has no attribute 'DATABASE_DRIVER'

I must have some module installed wrong. Did you get that?

Thank you so much for the suggestion on which branch to use. I was spinning my wheels trying to get a few others to work.

So the payout amounts will just be more volatile/random for the lower hashrate miners is what you're saying.

https://github.com/CryptoManiac/stratum-mining

Edit: My mistake, nvm.

Miner A is expected to have 4 shares, but may get less or more, in some kind of distribution around that. Miner B is expected to get 0.5 shares on average. They work out to the same thing. The work isn't in discrete blocks like that. It's a probability.

I am not mining NovaCoin right now, and it's the most profitable coin to mine.

I can't for the life of me get the NovaCoin pool to work. I'm sorry guys.

Makes Sense and makes me wonder if that is what I am seeing  myself at 1,300 MH..  I dont see where I am getting any better then just mining LTC.  Even today looks like I am on track for another 0.05 payout...

FWIW, here's why I think share difficulty matters.

When a new block is found on the network, every miner drops the share it's working on and starts working on a new share from the new block.

Sometimes, the miner doesn't get the notification fast enough and submits the share anyway (which then becomes a stale/reject).

On average, each miner should be about halfway through a share on a block boundray.

In this situation, a miner sumbitting e.g. difficulty 64 vs one submitting difficulty 512 shares (at roughly the same hashrate) would look like this:


In this scenario, miner A would have 4 difficulty-64 shares submitted before the first block boundary, and one aborted.  The miner using 512 difficulty shares has not finished one share yet and so he loses the equivalent of ~4.5 difficulty 64 shares when he restarts work.

Now, you could say that over time, this would even out (I am doubtful of that..  But..) Consider when the miner is not very fast.  Now you have the following situation:


Miner B in this scenario loses many shares because blocks are being found faster than he can find a share.  It looks like your pool is already up to around 300 MH, which means that on fast coins you are probably finding multiple blocks per minute.  At 512 share difficulty, I doubt a 300KH miner can even find a share in a under a minute on average.  It takes a 650KH miner  about 6 seconds on average to find a diff 64 share.  So a diff 512 share would average around 48 seconds.

I think people with large rigs that do 2000KH or more are going to receive a disproportionate amount of the rewards on your pool due to the high share difficulty.

h2odysee,

Do you plan to "diversify" by mining/dumping multiple coins at the same time? Say, mine the top 5 profitable instead of the top 1?

I suspect your automation might be related to an increase of sell versus buy order ratio being fill at Cryptsy.

If blindly dumping daily (even slowly) >8BTC on just a couple of coins, then you may affect the market at your disadvantage. It will work well at first, but on the long run the profit advantages will be less and less for most coins.  The buyer have little need to increase the price (except when pumping of course). The buyer just need to wait for the next automated sell order. Such bearish pressure might kill the goose.

Just food for thought.

Since i had <5% stales and my real hashrate was 720 KH/s, my effective real hashrate was 685 KH/s. Assuming the 65 shares do indeed represent 494 KH/s, that would mean I should have gotten 90 shares in that time in an hour.

Modelling block finding with a Poisson distribution, the probability of only finding 65 shares is about .2%

That's why I was a bit skeptical.

Liking the pool so far h2odysee, great idea. Looks like it's growing fast also.

I would imagine h2odysee is dealing with many parameters here. This  mining / trading combination is the first I've seen, it seems he's off to a very nice start.

I hope to see continued development with this type of pool system. My ping to this server is ~25ms.

Having mined at erundook's pools on the faster coins over the last few weeks, I have seen the changes hes made since adopting VARDIFF and how it has affected my productivity.

Obviously having properly scaled difficulty would lead to an increase in production,which would cause your pool to grow even faster and that could be a problem, depending on how you view it.

Since most pools are coin specific, is it likely that effectively applying VARDIFF to a dynamic pool like middlecoin could result in longer lag times and lower overall production when switching coins?  

If nothing else I would also be in favor of a slightly lower static difficulty if the server/s can handle it.



Thanks again for the great pool and I hope you continue to pursue this beyond the pizza factor  

Over the last six hours, with 340Kh/s on average, I've had 221 shares accepted and 31 (so ~12%) rejected. I'm pretty sure it's not the 200 ms ping, because other EU-users with similar ping report much lower rejected rates, so I think it is the difficulty. I'd like to second trying a slightly lower difficulty, to see whether people with lower hashrates (and those with higher too) get lower reject rates

Posted back on Page 13:

"As a note to everyone wondering why the hashrates on the site are different from the ones on your client, this is because the site only sees shares submitted, so bases the hashrate on those. However, the chance of getting a share is slightly random, and the 512 difficulty here makes it more random than the 256 or dynamic difficulty found other places. So your hashrate will often show slightly off; sometimes high, sometimes low. This doesn't necessarily mean anything is wrong, it's just the way the randomness of coin mining works"

Something seems to be wrong about your hash-rate estimate.

After mining for 1 hour, I had 65 accepted shares, 2 rejected (unknown job), of whichever coin we are mining at the moment (diff 108k). CGMiner claims an average of 720 KH/s, yet middlecoin.com tells me the average for the last hour was 494 KH/s. Something is very much off there, the probability that a miner running at 720 KH/s only finds shares worth a statistically average 494 KH/s miner during a whole hour is extremely small. The rejected shares work out to about 3% rejeccts, so that does not explain it.

H20..  You had stated this a few days back... 

":All I can say is that it will be more profitable than mining LTC. I may have a stat soon on the site displaying the profit per hash of the last day".

Would be cool to see the stat on the site displaying the profit per hash of the last day.   Coming soon?

Wish I was seeing that kind of increase but I am not..   

+1

I'm agree with you, there is not any hardware shortage i think, reject rate is because difficulty is so high, and when a block change in network or pool switch to other coin, the amount of work that should is garbage will be high.

I vote for decrease the difficulty a bit (e.g. 256)

Very true grasshopper haha, thanks for the idea lol.

Why not continue ot mine here and then buy ltc with btc you will have?