Most sector write-ups stop at the leaderboard. This case uses one published read, week 33 of 2026, and quotes its figures exactly as they came out. The read ranks 33 ETF groups every Friday by their one-week and four-week returns and turns them into five paragraphs. That week it ran under the headline "A wider advance led by energy, with crypto pulling the other way": 27 of 33 groups finished higher, Memory / DRAM led at +13.28% and Energy, the XLE fund, came second at +7.67%.
1. The figure that mattered was the second one
A 7.67% week is loud, but the read looked past it to the column next to it: +7.33% over four weeks. When the weekly and the four-week returns are that close, the three weeks before added next to nothing, so, in the read's words, "almost the whole monthly gain arrived in five sessions". Crude, the USO fund, was starker still at +7.31% on the week against +2.13% over four. The verdict took one sentence: "That is acceleration off a flat base, not a mature trend." Same place on the leaderboard as a sector that has climbed all month, very different animal.
2. What the leaders of the month did that week
The other half of a rotation is what gets left behind. The groups that had led the month kept rising, only slower: software rose 1.35% on the week against 12.16% over four, nuclear 0.24% against 13.34%, aerospace 0.73% against 11.12%. The read's phrase for them was "still positive, but running far below their monthly pace". Utilities did the reverse, up 1.61% on the week yet down 1.90% over four, which the read filed with transportation and real estate as "bounces inside monthly downtrends, not turns". Returns alone do not prove that money left software for oil. They show leadership changing hands, and the read was careful to say the combination "resists a clean label".
3. The test it set
The most useful paragraph of any read is the last one, the one that says what would make it wrong. For energy it was explicit: "energy and crude need their four-week figures to catch up to their weekly ones; if they do not, this was an event rather than a trend." Anyone can check that the following Friday, and it turns a headline into something that can fail. A burst off a flat base only becomes a trend when the weeks after it keep adding to the monthly number instead of handing it back.
4. What the next reads recorded
Week 34 put energy at +2.79% on the week and +6.74% over four, and the read listed it among the groups that "all extend an existing trend". Crude did not pass: up 6.35% on the week but down 1.50% over four, "a bounce inside a lower path rather than a turn". Week 35 gave some back, energy down 1.51% on the week and still up 5.26% over four. By week 36 energy stood at +2.20% and +11.41%, crude at +9.45% and +20.33%, and the read's word for moves pointing the same way on both clocks was plain: "That is continuation." None of this was a forecast coming true. It was the same two columns, read again every Friday.
5. Running the same check on your chart
The read is written in classic terms, a rising stack of moving averages, the upper edge of the volatility bands, momentum in oversold territory, and the page lists the 31 indicators behind it. Its readings are taken on weekly candles, and three of those indicators answer the burst-or-trend question on a weekly XLE chart. Edo EMA Core Cross plots the 9, 20, 50, 100 and 200 EMAs: tangled averages are a flat base, ordered and separated ones a trend, and a fresh 20 ↑ 50 label marks the turn in structure. Edo Bands Fusion names the state in its panel: Expansion Up when a close breaks above the Donchian channel, Trend Up when price holds above a rising basis. Edo RSI Dual shows whether RSI 14 has only just reached Overbought or has been living above 50 for weeks.
