Historical test result
Weekday ranks on eleven indexes: only 18% of months were positive
I tested one fixed way to trade weekday ranks across eleven index markets quoted by the broker, without buying the underlying shares. That rule lost money in this sample. If you trade weekday ranks on indexes, my combined 11-index test lost 2.2% a year with only 18% of months up.
Is it true that each index has its own best weekday? I ran one fixed weekday rule as a combined portfolio on 11 indexes. An index tracks a group of large stocks as one number. My combined test on 11 indexes lost 2.2% a year and only 18% of months closed up.
What I tested
I tested one fixed weekday rule on 11 indexes through 31 December 2023. At 15:00 UTC, I ranked the eleven indexes by their average return on the same weekday over the previous twenty weeks. The rule bought the top two and sold the bottom two at 16:00 UTC, then closed them at 16:00 UTC on the next shared trading day. The run held four legs at once across those 11 indexes. A leg is one trade inside the set of four. A portfolio means all legs counted together as one line. It was not one picked index. Each leg used size 0.125. The four legs together used total size 0.50. Each leg had a stop at 1.99%. In this test, the stop set a modelled exit level. Live fills can differ from the modelled price. The four stops together gave 0.995% total starting risk for the portfolio. I included frozen costs for each index on both entry and exit. These were fixed trading-cost assumptions.
What the numbers mean
The combined portfolio lost 2.21% a year through 31 December 2023. That yearly rate describes the full run as one line. Only 2020 closed up in this run. Total winning profits divided by total losing profits was 0.584. That ratio means winners were smaller than losers in the run. Winners paid about 58 cents for each dollar lost. The largest drop from a peak to a later low was 11.47%. Traders call that drop a drawdown. The worst single day lost 0.61%. Only 18.46% of months closed up. The remaining months were not positive. The first time block lost 2.21% a year. The two later time blocks lost 1.89% and 2.54% a year. No individual index had a positive average return after costs in both later time blocks.
Why thin data coverage is the central limit
The run used only 191 shared trading days out of 1,391 expected weekdays. That is 13.73% coverage. A shared trading day means all 11 indexes had data at the same fixed times. An expected weekday means a normal Monday to Friday date in the test span. There were 764 legs on those 191 days. Thin coverage is a central limit of this test. It is an added reason to drop the rule. It does not wipe away the loss. All three time blocks lost. Leaving out any one index still left a loss in both later time blocks. Higher costs made the losing line worse. Removing the stop still left a loss. The flipped rule also lost. Only 2 of 16 early gates passed.
What the 0.995% stop risk means in dollars
This part is only simple stop math. It is not a return forecast. The tested total starting risk was 0.995%. On a purely hypothetical 25,000 dollar account, 0.995% equals 248.75 dollars. Compare that only with a clearly named 1,250 dollar loss limit in the same example. Here the example loss limit is a maximum allowed account loss. That 248.75 dollars is about one fifth of that 1,250 dollar limit. This math shows only what the starting stop distance equals on that size. It never forecasts your loss or return. It does not show intraday drawdown. Intraday means inside one day. It does not establish compliance with a broker or challenge rule. I do not turn the 2.21% yearly test rate into dollars on your account.
What I do with weekday ideas
I dropped this weekday rank rule because it lost in my first run and in both later time blocks. I keep my own work on rules I state in one plain line.
Sources
I answered your question from my combined run through 31 December 2023. See more tested rules here: https://stratcorealpha.com/tested.
Check your own report
Group your own closed index trades by weekday and compare your best weekday result after costs with your average weekday result. Your trade report is the list your broker gives you with every closed trade and its result. Pull that list for one index you trade. Group the closed trades by weekday. Add up the result for each weekday after your own costs. Compare the weekday totals after costs. A best weekday in old trades does not establish that it will stay best on new data.
These are historical test results, not a live account.
If you want to check your own rules before you use them, try the free Prop Rule Guard here: https://stratcorealpha.com/rule-guard?ref=website.