CREDITS [CRE] | Request for comments | Difficulty retarget algorithm adjustment

[Gabriel Eiger]

This is a request for comments regarding updates to the Credits difficulty retarget algorithm.

Background:
The Credit blockchain is generated from a combined proof-of-work/proof-of-deposit system. This deposit system is still coming into effect as the blockchain evolves and it will start to have larger effects on the block creation process in approximately three months time. At the moment it is unknown what the effects on the difficulty and block creation process will be when deposit starts to have a larger impact. An adjustment of the difficulty retargeting mechanism is therefore suggested, to allow for faster adaption of the difficulty in case of varying deposit and processing power inputs. Faster adaption of the difficulty will also prevent the block creation process from being stuck to long in a too high difficulty.

Suggested adjustment:
Credits has so far been using the same difficulty retargeting mechanism as Bitcoin. That means that an adjustment of the difficulty happens every 2016 block. If the block creation process has been longer than two weeks for the latest 2016 blocks, the difficulty will rise, and it will fall if the block creation process has been faster. The difficulty adjustment is capped with a factor of 4 up or down, to prevent too large swings in difficulty.

The suggested adjustment of the retargeting mechanism is to have more frequent adjustments, with a lower cap for each adjustment. The retargeting period should be lowered to 1/8 of 2016 blocks, which means a difficulty adjustment every 252 blocks. The cap would be a factor of 1.25 for each period, giving a maximum difficulty adjustment of 5.96 over a period of 2016 blocks, as compared to Bitcoins 4.

An alternative proposal would be adjustment every 504 blocks, with a cap of 1.5. This would give a maximum adjustment of 5.06 for every 2016 blocks.

The new retargeting algorithm would be scheduled to start to have effect somewhere at block 24000 to 26000, or later if deemed necessary.

Credits official thread:
https://bitcointalk.org/index.php?topic=1037244.new#new

[GreydonIselmoe]

In my opinion, the rule change is welcomed and solid at 252 blocks, but the factor of 1.25x per retarget period of 252 seems a bit low, if you wanted an adaptable coin with an adaptable difficulty, I'd go for a cap of 1.5 for each period of 252 blocks.

Just my two cents, and ultimately you're the one making the decisions based off the communities input I'm a firm believer in the shorter the retargetting period, the better.

Greydon

[Gabriel Eiger]

In my opinion, the rule change is welcomed and solid at 252 blocks, but the factor of 1.25x per retarget period of 252 seems a bit low, if you wanted an adaptable coin with an adaptable difficulty, I'd go for a cap of 1.5 for each period of 252 blocks.

Just my two cents, and ultimately you're the one making the decisions based off the communities input I'm a firm believer in the shorter the retargetting period, the better.

Greydon

Thank you for the input, most welcome.

I tend to agree with you, that it might be correct with a combination of the two, short retarget periods with a higher cap.