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

[jedimstr]

Hmmm... SimonG is posting here again... where's tealover?

Scamming from SimonG ("Middlecoin is a scam, so come pay me some BTC so you can join my super-secret 50x profit private pool...that does't exist")
and
Threats of Blackmail, DOXing, and Extortion from tealover.

Shit always comes in pairs...
 (seriously... look back at the entire history of this thread... tealover and SimonG always post in the same batches of time... *queue conspiracy theory*)

[zneww]

Hmmm... SimonG is posting here again... where's tealover?

Scamming from SimonG ("Middlecoin is a scam, so come pay me some BTC so you can join my super-secret 50x profit private pool...that does't exist")
and
Threats of Blackmail, DOXing, and Extortion from tealover.

Shit always comes in pairs...
 (seriously... look back at the entire history of this thread... tealover and SimonG always post in the same batches of time... *queue conspiracy theory*)

[fcmatt]

Stats frozen again?

[Tripjammer]

Unless the stats are still screwed, yesterday was still a loss. My first bad day on this pool. Hopefully h2o gets these bugs fixed. Our payouts should be double the normal value tonight!

[SimonG]

Stats frozen again?

This is just another way for h20 to steal yoru money buy temporary freezing stats to make you not notice.

[smoovie]

The high unexchanged is fairly common. People ask about it every week.

I've been on Middlecoin continuous since last summer, and beta tested them with another address earlier in 2013.  Others pools have popped up on occasion and I've tried most of them.  Only Hashcows came close to matching Middlecoin's rate of payout averaged over a long span of time, but they haven't been the same since their holiday hack.

None of the other pools have Middlecoin's primary strength: H20dyssey's ability to sell coins when no one else can at a price no one can match.  He manually trades like a hedgefund manager instead of just relying on Autotrade. High unexchanged is an artifact of this.

Biggest example was during the Doge craze during the winter holidays.  He jumped on Doge 4 days before any other pool did, then waited to sell until Cryptsy locked up with Doge exchange issues.   Hashcows and direct Doge miners couldn't sell on their favored exchange.  People were clamoring to buy but couldn't.   H20 sold and we made a killing.  That's what your 3.88% or so fee goes to.   Well worth it.  

So don't worry about these high unexchanged values.  I love it when they get high, because at the end of the day, we get a bigger payout than if we just autosold everything like all the other multi pools out there during a market downtrend.

BTW- this isn't even close to the highest unexchanged values we've seen.

Just out of curiosity, but how do you know he was mining Doge 4 days before any other pool did. How do you know any of the coins middlepool mines. Still haven't been able to figure it out myself.

p.s. Welcome back SimonG

[vm_mpn]

Jed, have you evaluated WeMineAll.com ? It's open Beta right now so it's fairly new pool but it's from folks who run WeMineLTC so it's something to watch as well.

[fcmatt]

Jed, have you evaluated WeMineAll.com ? It's open Beta right now so it's fairly new pool but it's from folks who run WeMineLTC so it's something to watch as well.

Eventually there are going to be so many alt switching pools that it will mean lower payouts for all involved and one might as well mine ltc directly as that is pretty much what everyone aims to soundly beat.

[FrankieSaysRelax]

Unless the stats are still screwed, yesterday was still a loss. My first bad day on this pool. Hopefully h2o gets these bugs fixed. Our payouts should be double the normal value tonight!

Yesterday was horrible but it looks like h20 is importing some trades right now so we'll see what happens once the stats are updated.

[g0re79]

Out of all the other multi-scrypt coin most-profit pools out there other than Middlecoin, there are only three I've seen with the possibility of competing long term (the key here is LONG TERM).

1. HashCows - if they can get their pool back up to the rates and power they had before the hack, they're in contention.  Right now they're woefully underpowered (my sentiments and assessment of their current state of affairs have been agreed to by their admins Atriz and NearMiss).
2. Wafflepool - I've evaluated them 3 times, and each time it totally underperformed MC over the course of a week.  They rely heavily on autotrading, so they'll need to get their hands dirty with manual trading if they hope to compete.  Still, they have a good infrastructure and site setup so far.
3. ScryptGuild - Eleuthuria of BTCGuild has this as a new pet project and it's currently in Alpha with partial features.  He's planning on a public beta soon.  He will eventually have auto-trade and conversion to BTC, but this is not active yet.  Right now it's more like Multi-pool, but it will eventually resemble Hashcows in capability (options to either keep coins or trade to BTC) with the added capability to keep mining your specific controlled set of coins or let it Auto-jump to most profitable (like MC and Hashcows now).  I haven't seen this one talked much about here, but it has one of the biggest BTC Pool masterminds behind it and as such will probably have the strongest backend infrastructure which will be shared with the huge and powerful BTCGuild.  One to watch.

I'll be keeping an eye on all of these, and other smaller pools mentioned here, but for now Middlecoin is still the most profitable I've seen.

There is also clevermining.com

Yeah I haven't evaluated them and they're brand spanking new, so didn't want to render an unqualified opinion on them yet.  We'll see how they do long term.

And also trademybit.com - I split my hashpower between MC and trademybit for last two weeks and reslts are quite similar in long term (to be honest most multipools mine DOGE most of the time, so no surprise). But above all, TMB have a fancy web, bit confusing stats and nice graphs 

[jedimstr]

The high unexchanged is fairly common. People ask about it every week.

I've been on Middlecoin continuous since last summer, and beta tested them with another address earlier in 2013.  Others pools have popped up on occasion and I've tried most of them.  Only Hashcows came close to matching Middlecoin's rate of payout averaged over a long span of time, but they haven't been the same since their holiday hack.

None of the other pools have Middlecoin's primary strength: H20dyssey's ability to sell coins when no one else can at a price no one can match.  He manually trades like a hedgefund manager instead of just relying on Autotrade. High unexchanged is an artifact of this.

Biggest example was during the Doge craze during the winter holidays.  He jumped on Doge 4 days before any other pool did, then waited to sell until Cryptsy locked up with Doge exchange issues.   Hashcows and direct Doge miners couldn't sell on their favored exchange.  People were clamoring to buy but couldn't.   H20 sold and we made a killing.  That's what your 3.88% or so fee goes to.   Well worth it. 

So don't worry about these high unexchanged values.  I love it when they get high, because at the end of the day, we get a bigger payout than if we just autosold everything like all the other multi pools out there during a market downtrend.

BTW- this isn't even close to the highest unexchanged values we've seen.

Just out of curiosity, but how do you know he was mining Doge 4 days before any other pool did. How do you know any of the coins middlepool mines. Still haven't been able to figure it out myself.

p.s. Welcome back SimonG

Because I split mined between both Hashcows and Middlecoin at the time.   Second day in with Doge on MC and I saw no movement from Hashcows even though I started seeing the numbers move on up so I moved all my hashpower to 100% MC.  Two days after that, and hype caught up on Reddit and BitcoinTalk and Hashcows finally jumped on board the Doge train.  Then Cryptsy started having trade problems, and we made outstanding BTC profits here while everyone else were beholden to Cryptsy.

We knew what coins Middlecoin mined based on the network diff. People here track that sort of thing.

So how do I know? Experience. 

[jedimstr]

Out of all the other multi-scrypt coin most-profit pools out there other than Middlecoin, there are only three I've seen with the possibility of competing long term (the key here is LONG TERM).

1. HashCows - if they can get their pool back up to the rates and power they had before the hack, they're in contention.  Right now they're woefully underpowered (my sentiments and assessment of their current state of affairs have been agreed to by their admins Atriz and NearMiss).
2. Wafflepool - I've evaluated them 3 times, and each time it totally underperformed MC over the course of a week.  They rely heavily on autotrading, so they'll need to get their hands dirty with manual trading if they hope to compete.  Still, they have a good infrastructure and site setup so far.
3. ScryptGuild - Eleuthuria of BTCGuild has this as a new pet project and it's currently in Alpha with partial features.  He's planning on a public beta soon.  He will eventually have auto-trade and conversion to BTC, but this is not active yet.  Right now it's more like Multi-pool, but it will eventually resemble Hashcows in capability (options to either keep coins or trade to BTC) with the added capability to keep mining your specific controlled set of coins or let it Auto-jump to most profitable (like MC and Hashcows now).  I haven't seen this one talked much about here, but it has one of the biggest BTC Pool masterminds behind it and as such will probably have the strongest backend infrastructure which will be shared with the huge and powerful BTCGuild.  One to watch.

I'll be keeping an eye on all of these, and other smaller pools mentioned here, but for now Middlecoin is still the most profitable I've seen.

There is also clevermining.com

Yeah I haven't evaluated them and they're brand spanking new, so didn't want to render an unqualified opinion on them yet.  We'll see how they do long term.

And also trademybit.com - I split my hashpower between MC and trademybit for last two weeks and reslts are quite similar in long term (to be honest most multipools mine DOGE most of the time, so no surprise). But above all, TMB have a fancy web, bit confusing stats and nice graphs 

I've heard a lot about both clevermining and trademybit.  We'll see how well they do.  I won't be able to evaluate for a few weeks (been focused on RL stuff and Bitcoin p2pool mining lately), but interested in how you guys do after a few weeks. 

[ckoeber]

Eventually there are going to be so many alt switching pools that it will mean lower payouts for all involved and one might as well mine ltc directly as that is pretty much what everyone aims to soundly beat.

Hmmm, then it will come down to the switching algorithms themselves; how fast do you switch from coin to coin and how you execute your trades.

[fcmatt]

Eventually there are going to be so many alt switching pools that it will mean lower payouts for all involved and one might as well mine ltc directly as that is pretty much what everyone aims to soundly beat.

Hmmm, then it will come down to the switching algorithms themselves; how fast do you switch from coin to coin and how you execute your trades.

I do not think it will matter so much about switching from coin to coin. When you have a dozen plus switching mining pools they will all be mining the top 1-5 coins. So switching from one coin that is being mined by 4 pools to a different coin being mined by 2 pools to a different coin being mined by 3 pools.. you get my point. The algorithms are not exactly rocket science. Jump in when diff drops || price rises || speculative mining (which can back fire) || bid depth || knowing when the next block to be found is more valuable then the last || etc... In the end the top alt coins are being heavily mined by hashing power that comes and goes.

So would a pool admin try the highly speculative route in their algorithm trying to predict the future price of an alt coin? I doubt it. Might try it once in a while after they are established but in the end they will all fall into place when their are 1-5 obvious winners to mine.

So that leaves trading. But not trading how most people think of it. More like timing your selling. I tend to think the best seller could probably beat an automatic sell process but for most pools they will probably do much better just to sell every day and get an average over time. No one can predict the future reliably and in the end just selling every day will more then likely beat the majority who micro manages it.

So that leaves implementation of all the pieces together. Who has the best uptime and gets fewer denial of service attacks while gaining more hash rate. Who can scale their servers up to handle load to reach the optimal size of pool for the best payouts. Who can minimize stale/rejects for users making things more efficient and probably having a dozen servers world wide running stratum.

Dunno tho. I think about it sometimes. I have run a pool before but not a switching one.

[MikkaMK]

Having very high reject rates with new rig running 2x r9 290 tri-x runni sgminer and 13.11 beta drivers. Bat file as follows

sgminer -o stratum+tcp://middlecoin.com:3333 -xxxx-p x -I 23 -w 256 --thread-concurrency 32765 --gpu-engine 1020 --gpu-memclock 1500 --gpu-fan 85 --temp-target 88 --temp-overheat 92 --temp-cutoff 98 --gpu-powertune 20

Can't get these cards running with cgminer at all.

Any suggestions?

"-I" is too high. Start with 18/19/20. Check the hahsrate/rejects/WU and settle for the best values (depends also on what server you're hashing).

GPU and memclock might be too high as well. safer settings are 947/1250 or 975/1350 or 987/1425. Let it run a while (day or two), see if anything goes sick/dead/off. Have a good PSU?

I have 4xr9 290 cards running those setting stable and no issues, strange these cards don't like it. Tried intensity lower keep getting display adapter failed message it runs perfectly for 10 mins thn get the error. Psi is good antec 850w.

I am no expert but I would think to run 4xr9 290 you need at least 1200w, preferably 1300w.

Sorry, i am running cgminer on one rig with 4x R9 290 cards and an antec 1300w psu with the following settings

--lookup-gap 2 -I 20 -g 1 -w 512 --thread-concurrency 28456 --gpu-engine 1020 --gpu-memclock 1500 --gpu-fan 85 --temp-target 88 --temp-overheat 92 --temp-cutoff 98 --gpu-powertune 20

on middlecoin currently getting rejects of around 15%

second rig has 2 x R9 290 Tri-x couldnt get it running with cgminer so read a few threads saying sgminer is better for those cards. litecoin mining hardware comparison page recommended settings had intensity of 23 thats why i ran it at that. Currenly running that rig with the following settings:

-I 20 -w 256 --thread-concurrency 32765 --gpu-engine 975 --gpu-memclock 1350 --gpu-fan 85 --temp-target 88 --temp-overheat 92 --temp-cutoff 98 --gpu-powertune 20

on middlecoin i was getting around 40% rejects. I switched both machines over to altcoinplex just to try it out, rejects went to around 3%. With that being said no to confident with that pool, my total hashrate is 5.2mh/s and their stats page was only showing up as 3.3. Payout looks really low as well.

[ckoeber]

I do not think it will matter so much about switching from coin to coin. When you have a dozen plus switching mining pools they will all be mining the top 1-5 coins. So switching from one coin that is being mined by 4 pools to a different coin being mined by 2 pools to a different coin being mined by 3 pools.. you get my point. The algorithms are not exactly rocket science. Jump in when diff drops || price rises || speculative mining (which can back fire) || bid depth || knowing when the next block to be found is more valuable then the last || etc... In the end the top alt coins are being heavily mined by hashing power that comes and goes.

So would a pool admin try the highly speculative route in their algorithm trying to predict the future price of an alt coin? I doubt it. Might try it once in a while after they are established but in the end they will all fall into place when their are 1-5 obvious winners to mine.

So that leaves trading. But not trading how most people think of it. More like timing your selling. I tend to think the best seller could probably beat an automatic sell process but for most pools they will probably do much better just to sell every day and get an average over time. No one can predict the future reliably and in the end just selling every day will more then likely beat the majority who micro manages it.

So that leaves implementation of all the pieces together. Who has the best uptime and gets fewer denial of service attacks while gaining more hash rate. Who can scale their servers up to handle load to reach the optimal size of pool for the best payouts. Who can minimize stale/rejects for users making things more efficient and probably having a dozen servers world wide running stratum.

Dunno tho. I think about it sometimes. I have run a pool before but not a switching one.

I like your analysis; in the end you are showing that it is extremely cost prohibitive for tons of multi-pools to spring up.

Just think; after having about 300-400 MH/s your network connection to the internet is heavily taxed; you need a 50 Mbps line minimum that is dedicated. Anything less and high rejects.

You also need a system(s) that can handle High I/O; this generally means at least two boxes with one being a dedicated MySQL box.

After 1 GH/s you need multiple servers to do off-loading/failover/etc.

After 2 GH/s or more you need to start thinking about multiple network lines (if you haven't done so already) and multiple geographic locations.

Add all of that up and I highly doubt a whole lot of newcomers will start showing up once they see all of the work.

[fighter]

When are the payouts made?

[FrankieSaysRelax]

Having very high reject rates with new rig running 2x r9 290 tri-x runni sgminer and 13.11 beta drivers. Bat file as follows

sgminer -o stratum+tcp://middlecoin.com:3333 -xxxx-p x -I 23 -w 256 --thread-concurrency 32765 --gpu-engine 1020 --gpu-memclock 1500 --gpu-fan 85 --temp-target 88 --temp-overheat 92 --temp-cutoff 98 --gpu-powertune 20

Can't get these cards running with cgminer at all.

Any suggestions?

"-I" is too high. Start with 18/19/20. Check the hahsrate/rejects/WU and settle for the best values (depends also on what server you're hashing).

GPU and memclock might be too high as well. safer settings are 947/1250 or 975/1350 or 987/1425. Let it run a while (day or two), see if anything goes sick/dead/off. Have a good PSU?

I have 4xr9 290 cards running those setting stable and no issues, strange these cards don't like it. Tried intensity lower keep getting display adapter failed message it runs perfectly for 10 mins thn get the error. Psi is good antec 850w.

I am no expert but I would think to run 4xr9 290 you need at least 1200w, preferably 1300w.

Sorry, i am running cgminer on one rig with 4x R9 290 cards and an antec 1300w psu with the following settings

--lookup-gap 2 -I 20 -g 1 -w 512 --thread-concurrency 28456 --gpu-engine 1020 --gpu-memclock 1500 --gpu-fan 85 --temp-target 88 --temp-overheat 92 --temp-cutoff 98 --gpu-powertune 20

on middlecoin currently getting rejects of around 15%

second rig has 2 x R9 290 Tri-x couldnt get it running with cgminer so read a few threads saying sgminer is better for those cards. litecoin mining hardware comparison page recommended settings had intensity of 23 thats why i ran it at that. Currenly running that rig with the following settings:

-I 20 -w 256 --thread-concurrency 32765 --gpu-engine 975 --gpu-memclock 1350 --gpu-fan 85 --temp-target 88 --temp-overheat 92 --temp-cutoff 98 --gpu-powertune 20

on middlecoin i was getting around 40% rejects. I switched both machines over to altcoinplex just to try it out, rejects went to around 3%. With that being said no to confident with that pool, my total hashrate is 5.2mh/s and their stats page was only showing up as 3.3. Payout looks really low as well.

turn the intensity down to 18 and reduce the thread concurrency down by 512 increments till you find the most stable for your system. Factors like ram and other things effect TC

[fcmatt]

I do not think it will matter so much about switching from coin to coin. When you have a dozen plus switching mining pools they will all be mining the top 1-5 coins. So switching from one coin that is being mined by 4 pools to a different coin being mined by 2 pools to a different coin being mined by 3 pools.. you get my point. The algorithms are not exactly rocket science. Jump in when diff drops || price rises || speculative mining (which can back fire) || bid depth || knowing when the next block to be found is more valuable then the last || etc... In the end the top alt coins are being heavily mined by hashing power that comes and goes.

So would a pool admin try the highly speculative route in their algorithm trying to predict the future price of an alt coin? I doubt it. Might try it once in a while after they are established but in the end they will all fall into place when their are 1-5 obvious winners to mine.

So that leaves trading. But not trading how most people think of it. More like timing your selling. I tend to think the best seller could probably beat an automatic sell process but for most pools they will probably do much better just to sell every day and get an average over time. No one can predict the future reliably and in the end just selling every day will more then likely beat the majority who micro manages it.

So that leaves implementation of all the pieces together. Who has the best uptime and gets fewer denial of service attacks while gaining more hash rate. Who can scale their servers up to handle load to reach the optimal size of pool for the best payouts. Who can minimize stale/rejects for users making things more efficient and probably having a dozen servers world wide running stratum.

Dunno tho. I think about it sometimes. I have run a pool before but not a switching one.

I like your analysis; in the end you are showing that it is extremely cost prohibitive for tons of multi-pools to spring up.

Just think; after having about 300-400 MH/s your network connection to the internet is heavily taxed; you need a 50 Mbps line minimum that is dedicated. Anything less and high rejects.

You also need a system(s) that can handle High I/O; this generally means at least two boxes with one being a dedicated MySQL box.

After 1 GH/s you need multiple servers to do off-loading/failover/etc.

After 2 GH/s or more you need to start thinking about multiple network lines (if you haven't done so already) and multiple geographic locations.

Add all of that up and I highly doubt a whole lot of newcomers will start showing up once they see all of the work.

The most expensive part of running a pool for those who are well connected in the ISP world is DDOS protection. Hands down. And even paying for protection is no guarantee of being protected. No DDOS protection provider will stop a 100 gigabit per second DDOS that takes place for 4-12 hous for a couple hundred measily bucks a month. They will null route you. So you start paying more and perhaps you can be protected from only the most vicious of attacks. Plus pools are not just websites. Protecting port 80 and 443 is well refined by them. Not port 3333.

Surprisingly a pool does not use a lot of bandwidth. For every 1000 mhash/sec you can basically say you need 10 mb/s. And that is with a low diff of 64 for every user. Honestly bandwidth is hardly a concern. Low latency is much more appreciated by being well connected to the backbones. So if you use VARDIFF or just hardcode a diff of 1024 you can cut your bandwidth usage down quite a bit.

Which leads to I/O. The ultimate challenge for a pool. Without SSD drives (and not cheap ones) you will crumble and die. Mysql can be highly tweaked and I always wondered if pool admins have learned how to do this effectively. I had ~1000 mh/sec on one samsung pro 840 SSD at diff 64. I was doing about ~1000 insert/updates/reads a second on mysql. So if I turned diff to 1024 you can imagine how much more I could have scaled it up. Plus I had all sorts of stats running and a few UNIONS left to remove from my mysql queries which are horrible in mysql (always create a tmp table in mysql, fyi) before I folded. So I could only dream what a RAID array of 5 of those SSD drives could do!

So that leads to your point of multiple servers. I never got to that point myself so I can only guess how I would do it.
I would probably want a friend to spin up several VMs in his racks at major data centers world wide. These would run stratum and the local alt coin daemon(s). All shares would be sent to a hidden mysql server so it could not be DDOSd.
Crap knows what I would use for DDOS protection for them. Cloudflare crumbles under 60-100 gigabit per second DDOS and black holes you after a few hours of it. I doubt I would use a load balancer type hardware/software. Just let people pick a stratum server close to them. Dunno. One crosses that bridge when you get there. DNS can be used to switch to other backup servers, etc.. but if someone wants to knock you offline they can. Simple as that.

As for multiple internet connections that is assumed by being in a good data center. They always have more then one up to dozens. 2x10 Gig to the top of rack switches or other crazy amounts of bandwidth. gigE to the servers themselves.

So in the end.. will many of these alt coin pools pop up? YES. But what I am questioning is that once we have a half dozen successful ones all mining the half dozen best coins.. will it start to reach the point of driving down prices where beating litecoin mining itself is pretty tough? Look at ltcrabbit. That PPS pool. It added multi coin mining for a few coins and were paying litecoin pps + bonus on top of that. They are back to mining LTC again. Not even bothering to mine the 3 others they added in. Just aint worth the extra work I imagine to earn an extra 1%.

[KleinsBottle]

Anyone have any experience with Poolwarz.com?