[Passed] LiquidBits term 4 grant - Automatic Liquidity Pool (ALP)

Cybnate 2015-09-27 18:43:18 UTC #49

Another server.py crash. Rolled back to orignal. Nud database again corrupted. Started server.py, but not providing liquidity due to nud issues. Will look into this later today due to other commitments.

My sincere apologies.

zoro 2015-09-27 18:52:47 UTC #50

perhaps you should setup a new second server only for new exchanges in order to left intact the old ones

Cybnate 2015-09-28 08:23:41 UTC #51

Definitely a consideration, the cost and time have stopped me from doing that. But more important I shouldn’t try testing code myself and leave that to others

Cybnate 2015-09-28 09:21:51 UTC #52

Just restarted the server after getting the Nu daemon synchronised. The server should now be paying out automatically. Will wait until that has happened in about 24 hours from now. When successful I will dive into the logs to verify what backpay can be made based on the available logs. Given time constraints the backpay might take till the weekend as it requires a bit of work this time.

Thanks for your patience.

Cybnate 2015-09-29 09:36:52 UTC #53

Automatic payout for LiquidBits just occurred automatically and with that I expect the stability returns. As said will look into backpay over the weekend.

masterOfDisaster 2015-09-29 11:56:12 UTC #54

I can confirm that a payment arrived less than 2 hours ago.
The payment gap on 27th Sep and tiny payout on 28th Sep is due to the crashed server (and may lead to a manual payout if you have time to dig in the logs over the weekend)?

Cybnate 2015-09-30 08:00:43 UTC #55

Yes, that is correct. After that the server crashed another time and I had issues with the Nu daemon. So there will be a nice number of logs to work through for me. Can’t make it any better and my time is only limited.

masterOfDisaster 2015-09-30 08:03:52 UTC #56

Thanks for the feedback!

zoro 2015-09-30 08:06:04 UTC #57

perhaps you should consider to quit your daily job (or make it part time) and give NU the time it deserves
or is it too early for that?

Cybnate 2015-09-30 08:09:06 UTC #58

I would love to, but Nu can’t cover my basic cost of living yet. Will need to run another 10 ALPs concurrently

masterOfDisaster 2015-09-30 08:17:25 UTC #59

I will compete with NuBots for the compensation once we have a fixed cost model. Hah!
…you’d need to get back to work again

Cybnate 2015-09-30 08:23:33 UTC #60

Indeed, in the fixed cost model ALP you will likely get less than what you get now. The deal is good for shareholders but not for LPs. The Dutch auction model was much fairer and more dynamic (with bidding) in my opinion, however we should have limited and aligned the targets of the ALPs which was difficult given the decentralised model we have.

masterOfDisaster 2015-09-30 08:32:23 UTC #61

I tend to disagree.

In the Dutch auction model you get a compensation rate related to your share of the total liquidity on that pair.
With fixed costs it’s still the same, but easier to understand.

While the Dutch auction model was not good to determine the compensation LPs demand for doing their job, the fixed cost model is: just track the liquidity over time and calculate the rate!

I think that much is true. In difference to the current model I expect liquidity to hover around the amount that meets the expectation of LPs for average compensation rate for fixed cost compensation models.
Nu will have an easier time to steer liquidity in this case!
Nu will know how expensive (the compensation rate in percent) liquidity providing at a particular exchange is!

Cybnate 2015-09-30 08:36:09 UTC #62

You may

It is lacking the bidding process, everyone gets the same price whether you want or not. One could set a lower threshold in the fixed price model, but it is still different as the pool wouldn’t take your stake into consideration. It is a shame that creon isn’t updating his software any longer, so change is a given as the code is complex indeed.

And that is why I’m supporting it, it is not that great for the LPs though.

masterOfDisaster 2015-09-30 08:37:48 UTC #63

?
With fixed cost model everyone gets a share of the total payout (per minute) that is equal to the share of the total liquidity provided (in that minute).
There’s bidding. If I want a bigger share, I need to put in more money, which reduces the compensation rate per NBT, which might cause other LPs to drop off, because their minimum compensation rate is undercut.

Cybnate 2015-09-30 08:40:06 UTC #64

I meant to say the same share of the fixed price which is the same price/NuBit/minute. Not being specific enough…

Nagalim 2015-09-30 11:27:58 UTC #65

As fixed cost reduced rollover funds, it should be very clear that fixed cost tends to pay out more than fixed reward given the same size Nu grant. I don’t think it comes down to good or bad, I think it comes down to fair or not, and in my opinion fixed cost is more fair. Should someone doing their best to balance and maintain a good peg really be paid the same as someone who doesn’t care and just plops their funds in the bot? I think we should pay more to those doing the heavy lifting, which fixed cost achieves.

cryptog 2015-09-30 12:07:51 UTC #66

Could you clarify why it is the case?

zoro 2015-09-30 12:36:40 UTC #67

the exchanges with “smaller walls” will give the biggest payments to the participants. thus a user would want to always search for the exchange that will give him the max payment. the result would be that in any time, exchanges would be balanced by users trying to fill the small walls. This “trying” is a lot of work by the users

Cybnate 2015-10-04 04:15:24 UTC #68

Puzzle solved. Just transferred the remaining backpay for the 28-Sep. This day the server was running and the liquidity was on the exchange, except the liquidity was not submitted to the Nu client due to a corrupt blockchain which had to be reloaded and therefore automatic payment failed.

Thanks for your patience.