Pay On Target: New High variance payout System Offered by Ozcoin

[Graet]

Pay On Target

In an effort to bring some fun back to mining
and offer a high variance reward system to those that like a gamble.
You get more for shares of higher difficulty, meaning you get a big reward if you're the block finder.

We are offering a unique NEW payout system:
 Pay On Target
where higher value shares are paid at a higher value following this formula

Code:

[(1-a)/(1-a*wd^(1-a)*X^(a-1))]*(wd*B/D)*(min(X,sd)/wd)^a

where
B = Block reward - Pool fee
D = current difficulty
X = cap, currently 5*DNEW
work difficulty = difficulty pool sends to miner (see variable difficulty)
share difficulty = difficulty or target of share submitted to pool after hashing
a = a variable currently set to 0.8 but share value will be capped at 5*D

Pay On Target Pool fee will be set at 2% NEW

As always over time everything should even out to a 98% PPS rate BUT there isn't much time till GPUs become unviable, it is possible to come out ahead if you are lucky.
So I will ask "do you feel lucky? well do ya?"

IMPORTANT: POT will only be available on stratum servers.
You WILL make a loss if attempting to mine on getwork/ecoinpool servers
Stratum mining node addresses listed on "Worker" page on site.
This is a Per Account setting - if you want to split miners over payout methods feel free to make a second account
Set Pay on Target radio button on User Account -> Pool Options

Who to blame
ckolivas/conman for the idea
organofcorti and Meni Rosenfeld for help and advice on the more technical aspects and formula.
Blaksmith for coding it
Thanks for the help guys, much appreciated
Me for thinking it could be fun, give our stratum servers a good loadtest and attract some miners to Ozcoin.

Please consider this an initial trial period, we have done many simulations but reserve the right to change parameters. Please check this thread for updates regularly. Once the
trial period is over no further adjustments should be needed

Best wishes and best of luck to all.
Graet


POT calculator, thanks to kano
http://tradebtc.net/potcalc.php




Note: No Merged Mining available on Stratum servers

[-ck]

A lot of this idea came out of the fact that people feel cheated if they're the block finder and get no more than any other share for submitting it, and that now that cgminer scrolls difficulty, people constantly ask "so are the higher shares worth more?". It also acts as a disincentive to people acting maliciously performing what's known as a withholding attack by not submitting the block solve. It also adds an incentive to people mining on the fringe of profitability that they might actually make a profit if their hardware is the one that is lucky at that time. Ultimately, it's just gambling and over the very long term leads to the same rewards as PPS, but you never know your luck...

Here's an example:
[2012-12-18 22:52:10] Accepted 0000aa69 Diff 98.5K/1 GPU 1 pool 0
That share was worth 0.01435750 BTC where the current PPS rate is 0.00000712, meaning that one share was worth more than 2000 shares. On the other hand, diff 1 shares are worth 1/5th of what the going PPS rate is.

Note that in cgminer, up to and including version 2.10.2, the "best share" value may be wrong whereas the per-share statistic is accurate. This should be fixed from 2.10.3 onwards.

[kano]

I've been mining on it for 50 minutes and got a 51.7K/8 share in the first 6 minutes
Edit: Correction not 51.7K/8 - it was 2.25K/8 ... but still good

[Wave]

I'm in with my 1.21 Jiggawatts for a spin, what the heck ;-)


-Wave

[John (John K.)]

Awesome! 

[Meni Rosenfeld]

high variance reward system

That is an understatement. With a=0.8 (or any a>0.5) the variance is infinite. Greater than the variance of solo mining. The distribution will be quite different though, you'll have a high chance to get moderate rewards, and a small chance to get very high rewards.

The variance quickly goes down when decreasing a past 0.5, though. To get the variance to be less than solo variance, you need a < D / (2D + work_difficulty). Edit: should be a < 1/2 - wd/(8D) (approximately).

A lot of this idea came out of the fact that people feel cheated if they're the block finder and get no more than any other share for submitting it, ... It also acts as a disincentive to people acting maliciously performing what's known as a withholding attack by not submitting the block solve.

The goals should be clarified because this is a very ineffective way to solve these problems. You could do the standard method of block finding fee, but it greatly increases the variance.

[John (John K.)]

high variance reward system

That is an understatement. With a=0.8 (or any a>0.5) the variance is infinite. Greater than the variance of solo mining. The distribution will be quite different though, you'll have a high chance to get moderate rewards, and a small chance to get very high rewards.

The variance quickly goes down when decreasing a past 0.5, though. To get the variance to be less than solo variance, you need a < D / (2D + work_difficulty).

A lot of this idea came out of the fact that people feel cheated if they're the block finder and get no more than any other share for submitting it, ... It also acts as a disincentive to people acting maliciously performing what's known as a withholding attack by not submitting the block solve.

The goals should be clarified because this is a very ineffective way to solve this problem. You could do the standard method of block finding fee, but it greatly increases the variance.

Correct me if I'm wrong (which I think I am ), but won't the payouts be substantially more for previous PPS users on the pool? Previously, we get paid a fixed fee per share no matter how high the difficulty is, but now we get paid the fixed fee + extra stipend based on the difficulty. I do agree on the high variance though.

[Wave]

Feedback - On the POT stats page we need per share price.  I'm going to track every hour for a bit, but that would be nice.  I have been tracking more frequently for now.  Way over PPS, way under, and now trending slightly under.  Figure I'll run it 24 hours and see what the luck shows.


-Wave

[Meni Rosenfeld]

high variance reward system

That is an understatement. With a=0.8 (or any a>0.5) the variance is infinite. Greater than the variance of solo mining. The distribution will be quite different though, you'll have a high chance to get moderate rewards, and a small chance to get very high rewards.

The variance quickly goes down when decreasing a past 0.5, though. To get the variance to be less than solo variance, you need a < D / (2D + work_difficulty).

A lot of this idea came out of the fact that people feel cheated if they're the block finder and get no more than any other share for submitting it, ... It also acts as a disincentive to people acting maliciously performing what's known as a withholding attack by not submitting the block solve.

The goals should be clarified because this is a very ineffective way to solve this problem. You could do the standard method of block finding fee, but it greatly increases the variance.

Correct me if I'm wrong (which I think I am ), but won't the payouts be substantially more for previous PPS users on the pool? Previously, we get paid a fixed fee per share no matter how high the difficulty is, but now we get paid the fixed fee + extra stipend based on the difficulty. I do agree on the high variance though.

It's not plus, it's instead. Note the (1-a) term in the payout formula; it means that for a share that is just barely valid (share_difficulty = work_difficulty) you would get just 20% of the normal payout (with a=0.8 ), and you'd need the share to be much more difficult than that to get near the normal payment. On the flip side, if your share happens to be very difficult, you could get for it more than the entire block reward.

This poses high variance for the pool, not just the miners. I don't think 3% fee will cut it. Maybe a=0.4 will be better.

[John (John K.)]

high variance reward system

That is an understatement. With a=0.8 (or any a>0.5) the variance is infinite. Greater than the variance of solo mining. The distribution will be quite different though, you'll have a high chance to get moderate rewards, and a small chance to get very high rewards.

The variance quickly goes down when decreasing a past 0.5, though. To get the variance to be less than solo variance, you need a < D / (2D + work_difficulty).

A lot of this idea came out of the fact that people feel cheated if they're the block finder and get no more than any other share for submitting it, ... It also acts as a disincentive to people acting maliciously performing what's known as a withholding attack by not submitting the block solve.

The goals should be clarified because this is a very ineffective way to solve this problem. You could do the standard method of block finding fee, but it greatly increases the variance.

Correct me if I'm wrong (which I think I am ), but won't the payouts be substantially more for previous PPS users on the pool? Previously, we get paid a fixed fee per share no matter how high the difficulty is, but now we get paid the fixed fee + extra stipend based on the difficulty. I do agree on the high variance though.

It's note plus, it's instead. Note the (1-a) term in the payout formula; it means that for a share that is just barely valid (share_difficulty = work_difficulty) you would get just 20% of the normal payout (with a=0.8 ), and you'd need the share to be much more difficult than that to get near the normal payment. On the flip side, if your share happens to be very difficult, you could get for it more than the entire block reward.

This poses high variance for the pool, not just the miners. I don't think 3% fee will cut it. Maybe a=0.4 will be better.

Oops, didn't notice the (1-a). Now this would be like a lower variance solo mining, where you can end up getting ~20% of your regular PPS income to a lot more. Seems to be better than solo when you could end up with 0 though.

[John (John K.)]

Feedback - On the POT stats page we need per share price.  I'm going to track every hour for a bit, but that would be nice.  I have been tracking more frequently for now.  Way over PPS, way under, and now trending slightly under.  Figure I'll run it 24 hours and see what the luck shows.


-Wave

That'll be the 3% fee PPS rate, but depending on the share difficulty you submit. Man I'm hoping that I get more 300k shares like the one I got a couple days ago. 

[K1773R]

well this seems ot be the fairest payout system so far (no its not gambling, ur getting payed for ur work and not proportional/similiar), gl with it

[John (John K.)]

I guess I'm down under if the average share is below 5. Is that correct? 

[Wave]

Yup, understood.  But curious about trending.  I hear you on the share!  My last block solve was a 40M diff share!  Let me figure the payoff on that one...


-Wave

[Graet]

Please consider this an initial trial period, we have done many simulations but reserve the right to change parameters. Please check this thread for updates regularly. Once the
trial period is over no further adjustments should be needed

Sounds like we need to invoke this and reduce "a"
edited 1st post with some a = 0.5 samples

[DutchBrat]

Can I use Hashpower.com on Stratum servers yet ?

That would be like gambling to the power of 10 !!!

I like it  

[Wave]

I guess I'm down under if the average share is below 5. Is that correct? 

Does not appear that way:

I'm at 5021 shares, avg/ Dif 10.938 and payout is 0.03173515, or per share of 0.00000632.


-Wave

[JWU42]

Jump off PoT and then hit a 143K share 

[DutchBrat]

Jump off PoT and then hit a 143K share 

The system is rigged I tell ya... Rigged !!! 

[Graet]

Moving to a = 0.5 while i look at more figures
and 4% fee on advice