[ANN] profit switching auto-exchanging pool - www.middlecoin.com

[kirk46]

It would definately reduce the reject %

I believe it wouldn't reduce the reject %.

why?

we are mining a low diff coin and the rejection Mh has gone upto 100, that 100Mh that isnt finding blocks.....

[1714214193]

1714214193

[1714214193]

1714214193

[1714214193]

1714214193

[h2odysee]

why?

we are mining a low diff coin and the rejection Mh has gone upto 100, that 100Mh that isnt finding blocks.....

The reason for the rejects isn't that the network diff is low. It's because the block rate is high. The faster blocks come in, the more likely it is that your shares will be rejected due to a block being found before you.

Edit: To add to this, if someone in the pool finds a share before you, it will not invalidate any of your shares, unless his share also happened to be a block.

[ranlo]

why?

we are mining a low diff coin and the rejection Mh has gone upto 100, that 100Mh that isnt finding blocks.....

The reason for the rejects isn't that the network diff is low. It's because the block rate is high. The faster blocks come in, the more likely it is that your shares will be rejected due to a block being found before you.

Edit: To add to this, if someone in the pool finds a share before you, it will not invalidate any of your shares, unless his share also happened to be a block.

I think the confusion is this....

You said at the current rate, a 500 KH/s system should get a share every minute or so (based on 33s on 1 MH/s). Let's assume that the blocks are at a rate of 30s.

There is a high reject rate because a lot of the time, we are working on a share that will take longer to get than it does to find the block (from a faster system).

Reducing the difficulty to 1/4 would lower the "share" estimated time to 15s, or 1/2 the time to find a block, therefore many more legitimate shares should be submitted, which would also increase the chances of finding blocks.

I think this is where we're confused; the math doesn't add up for why 512 vs 128 (for example) would be the same.

[mueslo]

When choosing a coin, do you take in mind the current hashing power of the pool? This is kinda important since we're rather large now, just as big as multipool.in

Also, have you thought about the possibility of distributing miners among different coins in such a way as to maximize total income (since each miner brings the difficulty up)? This would be rather hard to implement though, I would guess.

[kirk46]

why?

we are mining a low diff coin and the rejection Mh has gone upto 100, that 100Mh that isnt finding blocks.....

The reason for the rejects isn't that the network diff is low. It's because the block rate is high. The faster blocks come in, the more likely it is that your shares will be rejected due to a block being found before you.

Edit: To add to this, if someone in the pool finds a share before you, it will not invalidate any of your shares, unless his share also happened to be a block.

yes i know that but if the pool diff was lower we would be able to submit the shares rather than lose them....

rejection now at 124Mh

[juve4v]

I dont know of its me but I see 60% reject rate now in cgminer.

[h2odysee]

When choosing a coin, do you take in mind the current hashing power of the pool? This is kinda important since we're rather large now, just as big as multipool.in

Also, have you thought about the possibility of distributing miners among different coins in such a way as to maximize total income (since each miner brings the difficulty up)? This would be rather hard to implement though, I would guess.

In a way, I take the pool's hashrate into account when I check to see if a coin has enough buy volume that I can sell to. So we won't go creating lots of nibbles or something.

Yes, it would make sense to distribute the miners among various coins, because then they would be switching a lot. That would be very hard  to implement though.

[h2odysee]

I dont know of its me but I see 60% reject rate now in cgminer.

Ouch.

Unfortunately, I take the pool's average reject rate into consideration when choosing a coin, which is now around 25%.

So you're losing out a little.

[h2odysee]

Reducing the difficulty to 1/4 would lower the "share" estimated time to 15s, or 1/2 the time to find a block, therefore many more legitimate shares should be submitted, which would also increase the chances of finding blocks.

I think this is where we're confused; the math doesn't add up for why 512 vs 128 (for example) would be the same.

A: (low diff)
You roll a 100-sided die, and for every 1 or 2 rolled, I give you a dollar. You can roll the die as many times as you want in one minute.

B: (high diff)
You roll a 100-sided die, and for every 1 rolled, I give you 2 dollars. You can roll the die as many times as you want in one minute.


Your profit is the same.

[Damnsammit]

You should make the difficulty 2048 since it has no affect on profit

[captainfuture]

so, the greater the pool gets, the less profitbale it will be?

what if your pool gets 2 or 3 ghs. then we will mine only coins with buy supply and cant mine the smaller ones...

seems to me that the future will be to split the coins mininig, but pay out the avaerage earnings to all....

if u can get that to work thats the future.

[STT]

You should make the difficulty 2048 since it has no affect on profit

Ironically that might help level the playing field     Set it to diff 10k


I got 3 pools on rotate, fast coin with diff 16 has raised accepted shares by 100 just since I started typing this.
LTC is on diff 32 switched to 128  and in a few mins its done +13 shares.  
Middlecoin is 512 diff of course and in 5 minutes, it raises accepted shares by 1 share.  That's it, near to a complete failure


[set cgwatcher to restart miner if theres no accepted shares, which will also take it to another pool till some time later]

[Eugen3]

There are more important things to work on. I will at some point do a better low diff test. But some things need to come first.

If you want me to make that a priority, then you can explain logically why having a lower diff would increase profits. So far, most explanations seem to fail by misunderstanding what a share exactly is.

Also, there is a coin that has a 42% rejection rate, but it's still sometimes the most profitable, even with that included. So it's ok when you see that high of a rejection rate.


It's not just a matter of upgrading servers. It would take a significant amount of time for me to rework the code to include diffs other than 512. And, I would need to do some more database optimizations, because that's where the slowdown will occur.

Thank you very much for this pool!  I appreciate the fact that the current coin being mined is not public information and that you are steadfast in defending the idea of not broadcasting it.  All those who want it to be public should first think about what will come of this type of operation (this is more than just a mining pool) if the market can expect X whatevercoins to be added to the sell side, since middlecoin.com had 500 M/Hash on it for 3 hours.  It can only hurt the effectiveness of the bot, which is contrary to why we are even here in the first place.

[TXSteve]

I like this pool just the  way it is, nice & simple, no bells & whistles -- maintain & manage the best you see fit,
you are doing a great job,

Steve

[Kuroth]

I like this pool just the  way it is, nice & simple, no bells & whistles -- maintain & manage the best you see fit,
you are doing a great job,

Steve

+1

[dddbtc]

What reward system does middlecoin use?

PPLNS?

Proportional?

[h2odysee]

What reward system does middlecoin use?

PPLNS?

Proportional?

Proportional.

[c1010010]

So, I love the concept of this site... just makes sense to me.  So, first, thanks for putting it together.

The concern I  have is that I (think I ) see a lot of wasted work.

Here is my thought process... please poke holes in it, or provide me with any thoughts as to how to improve.
If I look at the available stats, this is what I see as far as reject rate (or what I see as my measure of wasted work).

Total :
  Rejected : 9.9452
  Accepted :  562.2977
  Reject Rate : 1.76%

#1 spot:
  Rejected : 0.5779
  Accepted : 83.2802
  Reject Rate : 0.69%

#2 spot:   
  Rejected : 0.4847
  Accepted : 28.0646
  Reject Rate : 0.17%

#3 spot:
  Rejected : 0.7550
  Accepted : 24.1685
  Reject Rate : 3.1%

Me, specifically :
  Rejected : 0.2983
  Accepted :  8.7335   
  Reject Rate:  3.4%

5x the reject rate of spot #1
20x the reject rate of spot #2
About the same of spot #3
and 2x the total reject rate.

I feel like I'm doing something wrong.  I'm not the heaviest hitter here by any stretch, but I am in the Top-10 it appears. 
Why such a high reject rate?

First thing I do with other pools is check my latency and distance to the pool servers.  Things seem fine in that area.
Next thing I would do is to evaluate difficulty - but I don't have that ability here.

So, I welcome any comments/feedback/advice.

My goal would be 0% reject rate - but I know that is not possible.  Clearly <1% is... so what can I do to achieve that?

Thanks.

[h2odysee]

I just refreshed the page and ran the numbers myself, before the switch just now to a higher block rate coin.

Total :
  Rejected : 16.9450
  Accepted :  547.2448
  Reject Rate : 3.00%

#1 spot:
  Rejected : 1.0160
  Accepted : 78.9461
  Reject Rate : 1.27%

#2 spot:   
  Rejected : 0.7643
  Accepted : 27.7383
  Reject Rate : 2.68%

#3 spot:
  Rejected : 1.1744
  Accepted : 22.4815
  Reject Rate : 4.96%

So I guess those guys just had an hour of pretty good luck before you looked at their stats. Also, they may be using the --no-submit-stale option, which would skew their results a little.

We were mining feathercoin. Do you know the standard reject percent, for feathercoin pools? Is my pool much worse than the best? There might be ways to shave off a half a percent or so, if I improve my server's performance, but I'm not sure.

I'm not a big miner myself. I only have two 7970s. So I don't know all the tricks to getting good performance. Maybe there's some --expiry magic or something.

[h2odysee]

Just to note: In the hours before the stats I posted, we were mining FTC, CAP, and another coin with similar stats to FTC.

Anyone who has low rejects, it would be cool if you could chime in and say how you do it.