Strategy research

Previously studied history · no trading authorization

What happens when stocks are traded less often?

Keeping slightly lower-ranked holdings usually reduced trading. Higher returns were less common, and model predictions remain unstable.

Missing results stay blank. Every cost point is a separate saved account calculation. The table retains unavailable reasons.

At this horizon, cost and timing, 6 comparisons traded less and 2 returned more, out of 6 comparable pairs. Another 0 pairs were unavailable. These pairs share historical windows.

The buffer can retain an existing holding ranked 51–70, then fills vacancies by score. Equal weight always uses its unbuffered pool.

Rebalance every 5 or 20 sessions and hold actual units between trades. The model funds proportional costs from cash and includes the final sale. First and last usable regular-session trades are price proxies; executable fills remain unverified.

Each independent window has 40 signal dates and 41 valuation sessions. Original timing ends with one cash session; delayed timing starts with one. The windows start separate NAV 1 USD accounts.

Scores assume end-of-session signal formation. Actual source availability, complete security types and corporate actions remain unverified.

How much cost could the earlier intraday signals bear?

Of 24 requested zero-cost intraday accounts, 15 completed and 6 made a gross profit. Their exact break-even one-way costs ranged from 5.01 to 11.31 bps. The other completed accounts already lost money before costs. These older twenty-date windows differ from the new holdings experiment.

What has been checked in the stock inputs?

Candidate-days with unresolved identity

657,731

Issuer announcements / filings reviewed

7

Raw-price boundaries matching split seeds

4

Large adjacent observed raw-price jumps

243

The input retains 658,472 candidate-days. Seven known action boundaries are quarantined without changing prices. Four split boundaries match the sparse seeds. The 243 large adjacent observed jumps include gaps and other effects; they are not a count of confirmed corporate actions. Complete type/action history is still missing.

Earlier observation preparation

The earlier free daily-archive extension stopped at its fixed 300-second limit after 345 attempts and 55,656 recorded response bytes. The receipt has uncertain final transfer spend. At that snapshot, warmup was incomplete, no new registry was frozen, and no future observations were accepted. There was no automatic retry or schedule.

Previously studied history · no trading authorization

Where did the crypto account returns come from?

Market moves account for a large part of these weekly trend accounts' returns. Selection contributions vary by window and do not establish a stable edge.

Missing results stay blank. Every cost point is a separate saved account calculation. The table retains unavailable reasons.

This reuses 288 saved USDT accounts with no new replay. Daily PnL is split into a frictionless daily equal-weight market reference, an exposure adjustment, selection within invested capital and negative costs. Summing these NAV 1 contributions gives net return. This is descriptive accounting, not a causal explanation or an investable benchmark.

Each window has 90 scored UTC days and its own initial NAV 1 USDT. Selection and exposure contributions depend on the account path. The four accounts have different risk. Final holdings are marked without a terminal sale. Crypto delay shifts targets and trade dates by one UTC day, with cash before the first delayed trade.

Earlier observation preparation

The earlier free daily-archive extension stopped at its fixed 300-second limit after 345 attempts and 55,656 recorded response bytes. The receipt has uncertain final transfer spend. At that snapshot, warmup was incomplete, no new registry was frozen, and no future observations were accepted. There was no automatic retry or schedule.

Previously studied history

Holding coins and trading the trend

Weekly trend is sensitive to execution timing. It has not shown a stable advantage over the matched-target-exposure control.

The counts below cover every requested window at the selected cost and timing. Weekly trend exceeded buy and hold in 8 of 12 comparable windows.

Compare holding the eight coins with a fixed allocation and weekly trend rules. Change the cost or execution delay to inspect the same historical window.

20 bps means 0.2% of traded value. These are assumed total costs. Crypto decisions use a daily close proxy. Delaying shifts both the target and trade calendar by one UTC day, keeping cash before the initial trade.

Each 90-day window starts at normalized NAV1, denominated in USDT. Buy and hold and the fixed 60% allocation carry different risk from trend. The matched control matches target exposure at scheduled trades. Realized weights can drift afterwards. Final holdings are marked without a terminal sale.