Status as of Thursday, October 9

[Steve Sokolowski]

Quick status for today:

• poolpicker.eu added the site to their list of pools, which has drawn in more miners. There are 1.3 bitcoins left in the Elacoin buy order, so at 2.4Gh, it will take another 1.5 days to mine out the buy order, assuming the difficulty doesn't rise in the meantime. Those interested can see our sell history at https://www.cryptsy.com/markets/view/12; look at the green bars of steady sells as our hashrate has been increasing, and you can see how the system is selling coins every two or three minutes in the trade history.
• Today, I'll be putting out a request for help with merge mining on this forum. Someone said he would help, so I'll give him first crack at it, and then I'll offer $50 to the first person to tell me what's wrong with the block I post if he doesn't.
• Those spammers are sending huge numbers of 1-satoshi transactions to our wallet, but for now they aren't confirming. If they start confirming, the payout algorithm is going to fail one morning because it will detect that the transaction fees going out are unreasonable. If that happens, don't be surprised; Chris will manually send out the payouts later that day. I'm going to need to add a feature request into the system to ignore inputs to our wallet that are neither mined nor exchange-transferred coins, so we don't end up wasting people's money in transaction fees once the "ignore payout threshold" option becomes available in the next release.

[yan7181]

yes i am one of the miner that was drawn through poolpicker, just wished they added ur site earlier enough.

[bakedshibe]

It seems like we are getting a bunch of orphans with this high hashrate. I was also getting a big amount of rejected hashrate for some reason so i pulled that miner off.

[cryptorific]

Such wow, many miners! Seriously though, I expected alot of people from poolpicker but I'm still surprised. The highest priority should be to lower the auto-orphans because thats what appears to be happening. Can i suggest that rather than have everyone pile into the most profitable coin we split people into several groups that each go after the 1st, 2nd and 3rd most profitable coins?

[Steve Sokolowski]

Welcome, yan! Feel free to offer any comments you might have.

That's a feature that's in development, to go to multiple coins at the same time. But we can't release it yet because it's so intertwined with merge mining, which has cost me 45 hours so far. If I can't figure out merge mining, we may just go right to testing and allow all the merge mined blocks to be listed as "lost," which won't affect payouts but will allow us to release all the other features.

But the orphans are not affecting the profitability as much as you might think. I looked at the algorithm we are going to use to proportion miners across coins, and it would still go 100% to Elacoin. If the orphanage rate were high enough to make it a bad choice to mine the current coin, the pool would switch, because orphan rate is taken into account. Elacoins are so profitable that the algorithm has computed it's better to accept a high orphanage rate and mine out that buy order as soon as possible. If orphans got so high that they were causing huge losses, then the pool would switch to a coin with a higher difficulty, but it's just ridiculous how this guy has put in this buy order at such a high price. The orphanage rate would need to be 75% before a switch would happen. I still don't understand it.

As a side note, the algorithm is displaying emergent behavior that we hadn't anticipated before. For example, Chris has found that there is a "resonance" for the pool depending on hashrate: the algorithm doesn't just switch back and forth between networks all the time as we thought it would. Instead, it causes networks to stabilize at around our hashrate and we end up mining long strings of blocks until another pool figures out what's up, at which time we get out.

[Chris Sokolowski]

Unfortunately, the Elacoin party seems to have ended. Looks like that order was pulled around 10:30AM and the market is slowly declining in value as we sell our remaining coins.

[Steve Sokolowski]

I talked with Chris about the issue of orphaned blocks at lunch. We determined that the best thing to do for now is to raise the minimum difficulty configuration value from 0.01 to 0.5 or 1.0. We'll also reduce the orphan lookback period that computes the ratio of orphans from 2.5 weeks to 2.5 days. These things should reduce orphanage rates for now.

[Chris Sokolowski]

The config values have been changed. Let's see what happens.

[Steve Sokolowski]

You might need to increase the minimum difficulty even higher, perhaps to as high as 12.

A 20Mh Zeus miner will raise the difficulty of a coin to 1.2. Therefore, if we have coins that have 1 min blocks, and we have a hashrate of 2Gh, then the coins would be accelerated by a factor of 10 to 6 seconds, which is reasonable. I would set the minimum difficulty to just below whatever the Elacoin difficulty is.

The other option is to set the maximum orphan rate to something like 0.7, instead of the current 0.9. Then, the pool can teach itself which coins are too easy or which have other pools joining in.

But after I looked at this now, it seems like the pool has adjusted well, so I think now that you shouldn't make any changes at all for the time being.

[bakedshibe]

I'm still having weird issues with connectivity. workers getting banned and rejecting 25-30% of my hash. i have tried it on different rigs even rented one from miningrigrentals to check and still the same issues. no big deal but figured you guys would want to know.