🪑 27 August 2026 — the champion seat is vacant
This page has always promoted a champion candidate. As of today it does not have one, and the reason is worth more than the title was.
- The bar moved, not the candidate. Our promotion rule is that a variant's worst threshold cell must beat buy-and-hold — the floor, never the best cell. V5 "Activity Tide" cleared it with a floor of ×845 when holding returned ×642. Refreshed to 26 Aug 2026, holding now returns ×905 and V5's floor is unchanged at ×845. It did not get worse; the benchmark caught up while V5 sat in cash. No variant we have built currently clears the floor.
- Why the number never moved. V5 sold on 15 April 2026 and has been out ever since. A strategy holding cash has a frozen equity curve — so a rising benchmark can quietly overtake a published claim with no change at all on the strategy's side. We had this page telling readers V5 beat holding for weeks after it stopped being true.
- The one candidate that did clear the bar was paying for it with churn. A trough-turn re-entry rule scored a floor of ×1,121 — comfortably above holding — with zero trading costs. At a realistic 0.3% per turn its floor collapses to ×664, because most of its grid cells trade 60–200 times and a lot of those are one-day round trips. The edge was not signal, it was frequency, and frequency is not free.
- Two ideas are now measured dead. Reading the sign backwards — treating cheap gas as bullish, which the post-2023 data appears to suggest — returns ×0. Momentum-only machines return ×12–14. Both were plausible. Both are gone.
- And a humbling control. Bolting a dumb 60-day timer onto V5's exit rule — re-enter after 60 days, no signal, no thinking — beats V5's own re-entry logic. Whatever the level-recovery rule was contributing, it was worth less than a calendar.
The homepage indicator was replaced today on the strength of this research. The changelog is there. This page keeps no champion until something clears the floor with costs included.
Every change that rewired gas
"Gas price" sounds like one continuous series, but the machine generating it was redesigned repeatedly. The big ones for the math:
Genesis: gas as convention
Blocks are nearly empty; most wallets pay a hard-coded default (~50 gwei). Price reflects software defaults, not demand. A fee "market" barely exists.
Gas repricing after DoS attacks
Opcode costs rebalanced. Changes what gas buys, not how it's priced — minor for the indicator.
The fee market wakes up
Demand exceeds blockspace for the first time (ICO gas wars, then CryptoKitties). From here, gas price is a true first-price auction — users bid, the highest bids get in. Gas finally measures crowd urgency.
Supply inches up: 6.7M → 15M
Miners vote the block gas limit upward in steps. Each raise slightly deflates gas price for the same demand.
The auction is abolished
Fee formation is replaced by an algorithmic base fee that adjusts ±12.5% per block targeting half-full blocks — and is burned. Limit doubles to 30M. Same demand now produces a differently-shaped price series: smoother baseline, sharper spikes.
Proof of stake
Block times become a fixed 12s (+~3% throughput, more regular inclusion). Fee mechanics untouched — the smallest break of the four.
Blobs: activity stops paying L1 gas
Rollups move their data into blobs with a separate fee market. The largest source of L1 gas demand walks away overnight; user activity migrates to L2s where it no longer registers in L1 gas price at all. The biggest break in the series.
Pectra + relentless limit raises
Gas limit: 30→36M (Feb) → 45M (Jul) → 60M (Nov). Pectra (May 7) doubles blob capacity 3/6→6/9. Supply grows ~2× in one year while demand sits on L2s — structural fee deflation.
PeerDAS: the deflation continues
Blob capacity steps to 10/15, then 14/21; 80–100M gas limits proposed for 2026. Any indicator using absolute fee levels is permanently broken; even relative ones must adapt fast.
Four epochs, four different machines
Grouping by how the gas price was formed: E1 auction · immature E2 auction · mature E3 EIP-1559 burn E4 blob / L2. Two raw series below — the familiar per-transfer fee, and total network fee revenue (gas used × gas price × ETH price), which is more robust when the gas limit doubles.
Fee per transfer & total fee revenue, by epoch
The structural break, in numbers
| Epoch | Period | Mean fee | Structural drift | Time in high tide | Time in low tide |
|---|
Verdict: one formula does NOT fit all four epochs
Fees drifted +372%/yr in E1, +84%/yr in E2, then −52%/yr in E3 and −83%/yr in E4. The v1 z-score assumes a stable baseline, so in the blob era it sits in "low tide" 33% of the time and reaches "high tide" barely 1% — the indicator is being dragged by protocol engineering, not market mood. The 365-day rolling window self-heals within ~a year of each break (that's why v1 still worked), but every break costs it accuracy right when it matters. The math must adapt.
Six candidate formulas enter
Each variant was scored against the 13 ground-truth tops & bottoms (hits within ±120 days, median lag, false signals) and on strategy equity with next-day execution. Same hysteresis machinery as v1, thresholds grid-searched per variant.
Results — all candidates, unedited
| Candidate | Idea | Events hit | Med. lag | False sig. | Equity vs hold ×585 |
|---|---|---|---|---|---|
| V0 — Gas Tide v1 | per-transfer fee z + momentum | 9/13 | 67d | 10 | ×941 (1.61×) |
| V1 — epoch-aware z | z-score reset at each epoch boundary | 7/13 | 64d | 12 | ×608 (1.04×) |
| V2 — detrended z | z of residual vs rolling 1-yr trend | 9/13 | 50d | 12 | ×9 (0.02×) |
| V3 — Revenue Tide ★ | same math, fuel = total fee revenue | 9/13 | 66d | 10 | ×1,430 (2.44×) |
| V4 — detrended revenue | V2 + V3 combined | 10/13 | 51d | 11 | ×13 (0.02×) |
| V5 — revenue + tx momentum | fee level + transaction-count derivative | 9/13 | 64d | 9 | ×1,272 (2.17×) |
| V6/V7 — hybrids | blends of V3 + V4 | 8/13 | ~63d | 13 | ×18–29 (0.05×) |
Two clean findings — and one trap
Finding 1: switching the raw series from per-transfer fee to total fee revenue (V3) beats v1 on equity by 52% with identical math — revenue is naturally robust to gas-limit raises, because when supply doubles, price halves but volume holds. Finding 2: detrending (V4) is the best turning-point caller — 10/13 events, lowest lag. The trap: detrending also deletes the trend that a hold-strategy must ride; as a strategy it returns ×13 vs ×585 hold. Mixing the two (V6/V7) poisons both. So: two tools, two jobs — no single formula wins both contests, and that's a real result, not a failure.
Revenue Tide — Gas Tide v2 ★
Identical formula to the main page — z-score level + 90-day momentum, ±0.2 hysteresis band — but reading total daily fee revenue instead of per-transfer fees.
(next-day execution)
the COVID 2020 bottom
(v1 exited at $4,307)
19 signals · ~2/yr
ETH price · Revenue Tide calls
v2 vs v1 vs holding
Every v2 trade
| # | Bought | at | Sold | at | Return |
|---|
The Turn Confirmer (V4)
Detrended revenue z-score + momentum. It is not a strategy — it's a dated stamp that a major turn has happened. It confirmed 10 of 13 major tops & bottoms, typically ~6–7 weeks after the extreme — and nailed the COVID bottom same-day. Use case: "is this dip a cycle turn, or noise?"
Turn confirmations on the price history
Scorecard against all 13 ground-truth turns
| Actual extreme | ETH | Confirmed on | Delay |
|---|
⚖️ Lab rules — why this page isn't the homepage (yet)
- 13 events, ~550 backtests, grid-searched thresholds. That is an enormous amount of searching over very little ground truth. Four rounds of research have now spent roughly 550 machine-cells against 13 real turning points, and any "best" label was chosen after looking at all of them. Treat every ranking on this page as a hypothesis, not a finding.
- The re-entry rule is the unsolved problem. Every variant we have built exits well and re-enters badly. Since 1 July 2026 ETH is up 56% and every machine on this page — champion, challengers and the retired homepage indicator alike — has been sitting in cash for all of it. Until that is fixed, none of this is ready to carry a headline.
- The epoch problem is real but not fully solved. Revenue reduces the blob-era bias (low-tide time 33%→29%) but doesn't cure it; detrending cures it but can't hold a trend. An epoch-proof single formula remains open.
- E4 keeps mutating. Fusaka BPOs and 80–100M gas proposals mean the deflation isn't done. Every supply jump is a small fresh break.
- Same caveats as v1: daily averages, no fees/slippage in equity curves, not financial advice.