EIP-3554: Ice Age delay targeting December 2021

The delay is 500,000 extra blocks from the previous delay.

The adjustment is less than would be expected because the difficulty is so much higher today than previous adjustments.

There is a script in there as well to play around with cr: @vbuterin

It will be helpful to re-run the numbers from time to time as the difficulty changes to make sure we are on target. A severe drop In difficulty could lead to increased block times seemingly out of nowhere.

The block prediction in the formula isn’t very accurate. I am rerunning the simulation using a targeted specific block number. https://etherscan.io/block/countdown/13766666

which is currently Sat Dec 04 2021

@MadeofTin could we perhaps add 200,000 blocks to the delay to get one extra month?

As per your comment on Github we would start to see the bomb go off in November:

Reference block 12382958 (May-06-2021 08:38:33 PM +UTC)
Nov - 5 months - 1.9% + .2 seconds
Dec - 6 months - 7.6% + 1 second
Jan - 7 months - 30% + 4 seconds
Feb - 8 months - 121% + 16 seconds
March - 9 months - 487% + 64 seconds

I personally would feel more confident if that first delay started in December vs. November.

Roughly, it seems there are ~200 000 blocks in a month (2.628e+6 secs ./ ~13 sec = 202153 ~= 200000). Would this change be as simple as swapping 9,500,000 for 9,700,000 ?

1 Like

This commit addresses my concern above.

2 Likes

I backtested for all of the times the difficulty bomb was first visible.

#First bomb sight May 1st 2017  (Coeff = 0.10)
#diff_adjust_coeff = predict_diff_bomb_effect(3630000,336173000000000,0,0)

#Second Bomb Sighting Dec 20th 2018 (Coeff=0.12)
#diff_adjust_coeff = predict_diff_bomb_effect(6920000,2263365000000000,3000000,0)

#Third Bomb Sighting Nov 10th 2019 (Coeff=0.11)
#diff_adjust_coeff = predict_diff_bomb_effect(8910000,2532534000000000,5000000,0)

This confirms the bomb should be seen when the Coeff reaches 0.1

The first time the diff_adjustment_coeff is greater than 0.1 is at block Height 13800000 https://etherscan.io/block/countdown/13800000

This is because the difficulty bomb steps up only ever 10,000 blocks.

13700000 = 0.0761
13800000 = 0.1522

1 Like

This is great, thanks!

Small ask: could we change the name of the EIP to " EIP-3554: Difficulty Bomb Delay to December 1st 2021" so that it just mentions the month and not specific day?

I did some further analysis and modeled what the difficulty bomb would look like under half the current difficulty and under double the current difficulty.

You can see if under the bomb will show 100_000 blocks early.
If double, the bomb will show 100_000 blocks later

More accurate graph

An overview of Difficulty bomb and EIP-3554 by @MadeofTin

As far as I understand – miners will start feeling this bomb someday in June 2022 and then it’ll be more and more noticeable.
But they still will be able to mine till ~June 2023 with a lot less rewards / lot more effort?
Or is this a race between difficulty bomb and the PoS transition?

Sorry if those are dumb questions, just spend a week trying to analyse miner’s perspectives and I’m still kind of lost. Warm hugs to @MadeofTin for his interview, it got me some insights on the process.

2 Likes

I would love this answered. Is your statement accurate?

Would love to hear response for @zemek if mining will still be possible with a lot fewer rewards after POS (and for some time it will be a hybrid model) or once POS is there it will be not a single block to mine?