BitShares Forum

Main => General Discussion => Topic started by: sudo on August 02, 2014, 02:43:54 pm

Title: PTS difficulty adjusting algorithm is so poor
Post by: sudo on August 02, 2014, 02:43:54 pm: difficulty adjusting algorithm of PTS is so poor
snapshot announced →price increase →POW rush into PTS →difficulty become high in a short time
→snapshot passed →price fall →POW run away →but difficulty still high for several months to adjust next time

1 PTS pow →dpos
2 PTS pow update the difficulty adjusting algorithm
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: toast on August 02, 2014, 03:36:40 pm: Testz is making a difficulty patch for short term, then we can upgrade to dpos long term

Sent from my SCH-I535 using Tapatalk
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: santaclause102 on August 02, 2014, 03:46:36 pm: why did the difficulty in the first place not adjust every two weeks like with BTC??
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: sudo on August 02, 2014, 04:40:30 pm: Quote from: toast on August 02, 2014, 03:36:40 pm
Testz is making a difficulty patch for short term, then we can upgrade to dpos long term

Sent from my SCH-I535 using Tapatalk

+5%
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: toast on August 02, 2014, 04:46:49 pm: Quote from: delulo on August 02, 2014, 03:46:36 pm
why did the difficulty in the first place not adjust every two weeks like with BTC??

It did... read the OP. Retarget happens at "two weeks' worth of blocks" and not actually "two weeks".
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: santaclause102 on August 02, 2014, 06:24:13 pm: Quote from: toast on August 02, 2014, 04:46:49 pm
Quote from: delulo on August 02, 2014, 03:46:36 pm
why did the difficulty in the first place not adjust every two weeks like with BTC??

It did... read the OP. Retarget happens at "two weeks' worth of blocks" and not actually "two weeks".
....which made me think... what is testz then patching? Making the default to be less than 2 weeks?
BTW: (I didnt use PTS lately) Is it relatively fast (~10 minutes)?
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: toast on August 02, 2014, 06:25:31 pm: changing it to be a better retargeting algorithm (KGW or similar)
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: santaclause102 on August 02, 2014, 06:30:13 pm: fancy POW http://bitcoin.stackexchange.com/questions/21730/how-does-the-kimoto-gravity-well-regulate-difficulty
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: gamey on August 03, 2014, 03:41:12 am: I believe the problem has more to do with large GPU farms hammering it when the difficulty becomes low. They then leave when they blow through it. It isn't hit by normal multipools because it doesn't use scrypt or some standard algorithm and the hobbyists guys aren't on top of it that much.

Glad to see testz patching it. +5%
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: jsidhu on August 03, 2014, 05:28:31 am: Quote from: toast on August 02, 2014, 06:25:31 pm
changing it to be a better retargeting algorithm (KGW or similar)

I did this already but was abondoned, when the ideas was to do dpos and then that was abondoned too? I still have the code on my github.
Title: Re: PTS difficulty adjusting algorithm is so poor
Post by: gamey on August 03, 2014, 04:26:04 pm: Quote from: jsidhu on August 03, 2014, 05:28:31 am
Quote from: toast on August 02, 2014, 06:25:31 pm
changing it to be a better retargeting algorithm (KGW or similar)

I did this already but was abondoned, when the ideas was to do dpos and then that was abondoned too? I still have the code on my github.

+5% maybe testz can give it a thorough review ?