Two weeks ago an early estimate said difficulty was headed down about 2.7%. It ended at about zero. This week an early estimate says up almost 6%. Same lesson, opposite sign.
How the last epoch actually closed
The previous retarget came on September 19, a rise of about 4.16% to 132.76T. Four days later, on September 23, that epoch was about 30% done. Blocks were running a little slow and the live mempool.space projection read about -2.7%. By September 30, with about 81% of the window behind it, the projection had drifted to roughly +0.08%. By Friday afternoon, October 2, it read about -0.2%.
The retarget landed at height 969,696 just after midnight Pacific time on Saturday, October 3 (12:16 AM PT). Difficulty moved about -0.03%, to 132.72T. Essentially flat. The early -2.7% was a fair read of slow blocks at the time. It was never a forecast.
Where the new epoch sits this morning
As of Wednesday morning Pacific time, October 7, the chain tip is 970,349. That is 653 blocks into the new window, about 32% of 2,016. There are 1,363 blocks left to the next retarget at 971,712.
Blocks are averaging about 568 seconds, roughly 9.47 minutes, against the ten-minute target. That is why the live projected change reads about +5.8%. If block times hold, mempool.space puts the next retarget around the morning of Friday, October 16, about nine days out.
Hashrate reads about 984 EH/s on mempool.space's current figure. The one-day averages in its weekly series ranged from about 869 EH/s to about 1,177 EH/s across October 1 to 7. That spread is mostly measurement noise from luck in block discovery, and it is exactly why a single day should not set your expectation for the retarget.
Fees are quiet. Recommended rates are 3 sat/vB for fastest and half-hour, and 1 sat/vB for hour and economy. The mempool holds about 79,800 transactions (about 42.8 million vbytes).
Why +5.8% is not a lock either
A projection at one third of the epoch is computed from about a third of the blocks. If the next 1,363 blocks come in closer to ten minutes, the estimate shrinks. If hashrate keeps arriving, it can hold or grow. Both happened in the last two windows: September's early read overshot to the downside, and the epoch before that finished with a hike of about 4%.
The useful habit does not change with the sign:
- Watch epoch progress and blocks remaining. At 32%, there is plenty of room for the estimate to move.
- Watch average block time against the ten-minute target. Today's 9.47 minutes is what the +5.8% encodes.
- Watch smoothed hashrate over days and weeks, not one-day prints. The past week alone swung by about 300 EH/s between daily averages.
- Watch mempool depth and fees. A hike into a quiet fee market at 1 to 3 sat/vB is a different miner-math story than a hike into a busy one.
Where NodeView fits
NodeView puts that board in one place: live hash rate, difficulty, mempool, fees, active nodes, market charts, and a block explorer. The core dashboard is free. Premium adds valuation models and forecasts if you want them later.
Check back over the next nine days. Ask whether the +5.8% holds as the window fills in, or settles the way the -2.7% did.
Not financial advice. NodeView provides data and analytics for informational purposes only.