Topic: [1050 TH] BitMinter.com [1% PPLNS,Pays TxFees +MergedMining,Stratum,GBT,vardiff] - page 228. (Read 837101 times)

Half hour downtime...

Sorry, a new version of the mining server was not playing nice with all miner versions in use. I quickly reverted back to the old one.

I logged in to see my hashrate at 0 and cgminer was getting authentication errors. I deleted my workers and re-created them and now things are working for me.

Did even more downtime just happen?  Pool down to 1 TH/s

All the data is available but it would take a few days work.  I did  this back in the fall and bitminter was fine for a 90 day period of time.  I will tell you what i will go back and do it again. I will go back to aug 2012.

Below is from aug 1 2012.  I used this calculator https://bitclockers.com/calc   to get the normal number.

keep in mind 50 coins a block if you try to double-check my work  until late nov.

hash power is the hash for each block made added and divided by the blocks made .  so it is a good but not perfect number.  I am up to aug 10th 2012 we made 61 blocks and normal would be 53.16


  61 blocks made vs 53.16  is huge .  1.147 to 1   or 14.7 percent over normal then toss in the namecoins.  you also need to understand that determining the true daily hash rate  is not possible I don't have enough samples of hash rate my samples are when a block is found.


DATE------------------Blocks----------------Difficulty--------------Hashpower-------------- normal blocks for this

 Aug  1 2012------------7-------------- 2,036,671------------------470.85Gh/s--------------------- 4.64
 Aug  2 2012------------3---------------2,036,671------------------534.00Gh/s----------------------5.26
 Aug  3 2012------------5---------------2,036,671------------------531.99Gh/s----------------------5.25
 Aug  4 2012------------6---------------2,036,671------------------513.04Gh/s----------------------5.07
 Aug  5 2012------------4---------------2,036,671------------------518.61Gh/s----------------------5.11
 Aug  6 2012------------3---------------2,036,671------------------552.93Gh/s----------------------5.46
 Aug  7 2012------------5---------------2,036,671------------------536.80Gh/s----------------------5.30
 Aug  8 2012------------8---------------2,036,671------------------540.74Gh/s----------------------5.34
 Aug  9 2012------------9---------------2,036,671------------------557.76Gh/s----------------------5.51  
 Aug 10 2012-----------11--------------2,036,671------------------633.70Gh/s----------------------6.26



This method has a flaw  the hash rate of bitminter is found by using the hashpower when a block was made.  so if it was a bad day and 3 blocks were made only 3 samples of hash rate were used to find  the daily hashrate.  So the hash rate could be higher and the luck is worse but more likely the average hash rate is lower .  reason is if bitminter goes down and hash rate is 0  for 1 hour no block will be made.  So my hash rate average is missing all down time.   BTW I believe that is true when you look at bitminter's luck charts.  In order to be really accurate 4 samples of hashrate every hour would be far better to know the true daily hashrate. A total of 96 samples a day.

    I think the doctors luck chart is wrong  and always reads low for the same reason. He uses the current hashrate when the block is found. Thus the luck is calculated with all down time missing.  So in a 24 hour day if the system is down 1 hour  luck will be based on 24 hours up time.  The only way to really figure out luck is for some one to write a script that gives us 4 samples(even 10 per hour)  of bitminter every hour of every day for the next month. I would be certain that the average hashrate with those numbers will be 3 to 5% less then the hash rate used by counting it only when a block is minted.

Yup, my point was that you can't expect 24 hrs based "results" after just a few hours of higher hashrate.

Either way, it's a statistical expected average.
Give it a week or two and you still can't guarantee it to be very near the average ...

I think you need to give things more time.

Somebody correct me if I'm wrong, ASICMINER started mining on Bitminter from around shift 14893 2013-05-04 08:24 UTC (an educated guesstimate).

Now, 22 hours later, at ~ 6:15 UTC we have mined 15 blocks. Looks like we are right on track.

I thought so, but nope, we took over BTCGuild on 234580 thanks to 234581 right behind, and theirs was orphaned.

2 Stale blocks in a row!?! 

On the long run, if the attacker submits enough 'rotten' shares (i.e. discarding winning ones), wouldn't it easily be detected by the lower than expected (i.e. zero, if the attacker discards all winning shares) found blocks / submitted proofs of work?

Waiting on that run of good luck...I mean the other guy did the math. How many days will it take to get 16 blocks (average) in a day???

Sorry, there was a few minutes downtime as the server was overloaded. Running OK again now.

And We're up =)

The shares are still accepted, or am I wrong? (might be the backup server I can't tell from here)

And We're Down.....

DDoS or server screwed up?

A clearer description of the x-axis would probably be helpful. I.e the actual dates.

I said that actual rewards have trended below expected for several months.  That is true.  And it continues to be true.

If everything was kosher, you would expect to see that trend chart having equal areas above, and below the expected reward line.  I have never looked at the chart and seen it ahead at the end of a 500 block period.

It's not FUD to give you fair warning that you are likely being exploited in some fashion.

Graet was robbed for months by miners hopping payout methods and didn't notice until they broke the bank.  You need to be more vigilant.

Edit:  The 500 block shortfall is also remarkably consistent at slightly greater than 10%.  I mined several thousand coins with Bitminter before deciding the pattern was too suspicious to tolerate.  That's the price of a nice car that is missing, and I was only 3% of your pool at my peak.

Do you mean we have not had one lucky day for a full year? You know that's not true.

The actual is above the expected for part of the graph when I look at it now: http://bitminter.com/stats/rewards
And it has not been 1 year since this could last be seen.

The rewards graph is only the last 500 shifts. Currently, as I write this, it goes back to shift 14399 which was 2013-04-18 20:38 UTC. That's only 16 days. I'll agree the last 5 days or so were bad, but let's not go overboard.

I'll get some more data to show how the luck has been. In the meantime please give the FUD some rest.