FuturesAI Signals: How to Tally a Ledger by Basis Tag
Tag a signal with “MACD bearish” and it looks like a filter. Count the calls behind it, and the filter turns out to be something else.
Short answer: FuturesAI’s public scoreboard lists the basis tags behind its recent calls. I tallied 93 of the 100 ledger records read on November 16, 2026: 44 hits, 47.3%. Calls tagged MACD bearish hit 13 of 18, but 10 of those were DOWN calls, and DOWN calls as a group hit 23 of 37. Before trusting a tag, split it by direction and check the sample size.
Educational content, not investment advice. Leveraged futures can lose the full principal, and FuturesAI makes no profit promise. These are my own counts from a small public sample, not a forecast and not a trading rule.
See the payback rates on FuturesAI
Jump to a Section
- What does the scoreboard ledger actually list for each call?
- Which basis tags had the best hit rates?
- Is the tag doing the work, or is it the direction?
- How big is the sample behind each number?
- A Four-Column Sheet You Can Rebuild in Ten Minutes
What does the scoreboard ledger actually list for each call?
FuturesAI’s scoreboard ledger lists, for each frozen call, the ticker, the direction, a signal basis made of short tags, the frozen and resolved prices, and whether the direction matched.
The ledger header says it holds 100 records, 47 hits and 53 misses, as read on November 16, 2026. The records run from early September to mid November 2026 and mix Korean stocks, US stocks and crypto.
When I checked the individual rows, I found that 93 of the 100 show a signal basis and 7 do not, so those 7 are left out here. They hold 3 hits and 4 misses. The remaining 93 hold 44 hits.
The basis comes in two kinds. Calls labeled signal carry tags such as RSI low, MACD bearish or Below 20-day MA. Calls labeled focus carry tags such as Big move, Value surge, Search spike or AI strong signal.
The page does not describe how these tags are weighted, so treat them as labels, not scores.

Which basis tags had the best hit rates?
Among the 93 tagged calls on FuturesAI’s ledger, MACD bearish (13 of 18) and RSI low (16 of 24) had the highest hit rates, while MACD bullish (4 of 14) had the lowest.
| Group (ledger read Nov 16, 2026) | Hits / calls | Hit rate | Rough 95% range |
|---|---|---|---|
| All 93 tagged calls | 44 / 93 | 47.3% | 37% to 57% |
| Labeled signal | 31 / 67 | 46.3% | 35% to 58% |
| Labeled focus | 13 / 26 | 50.0% | 32% to 68% |
| Tag: MACD bearish | 13 / 18 | 72.2% | 49% to 88% |
| Tag: RSI low | 16 / 24 | 66.7% | 47% to 82% |
| Tag: MACD bullish | 4 / 14 | 28.6% | 12% to 55% |
Own tally from the public ledger. A call can carry several tags, so tag rows overlap. The range is a Wilson interval, a standard way to show how much a small sample can move.
Now look at the last column. The ranges overlap almost everywhere, and the 72.2% row still starts at 49%, which is a coin flip.
A big number on a small sample is the easiest thing to fall for.
Is the tag doing the work, or is it the direction?
In FuturesAI’s 93 tagged ledger calls, DOWN calls hit 23 of 37 (62.2%) and UP calls hit 21 of 56 (37.5%), so a tag that only appears on one side may just be measuring direction.
Split the tags by direction and the picture changes. MACD bearish appears on 11 DOWN calls (10 hits) and 7 UP calls (3 hits). RSI low appears on 14 DOWN calls (11 hits) and 10 UP calls (5 hits).
Two other tags never mix at all. In this ledger, Underperforming index shows up only on DOWN calls (16 of 27) and Outperforming index only on UP calls (8 of 19).
| Tag | On DOWN calls | On UP calls |
|---|---|---|
| MACD bearish | 10 / 11 | 3 / 7 |
| RSI low | 11 / 14 | 5 / 10 |
| MACD bullish | none | 4 / 14 |
| Underperforming index | 16 / 27 | none |
| Outperforming index | none | 8 / 19 |
So what looked like a strong MACD bearish tag is mostly a strong stretch for DOWN calls. Between September and November 2026 several stocks in this ledger fell hard, so DOWN calls had an easy time.
That is a market-window effect, not a rule. A different quarter could flip it.
For a different check on the same record, see FuturesAI Signals: A 3-Gate Filter Before You Trust One.
How big is the sample behind each number?
A hit rate on fewer than about 30 calls can swing by 20 percentage points or more either way, so the 18 calls behind MACD bearish are too few to name a winning filter.
Here is the quick test I use. Take the hits, divide by the calls, then ask how many calls it would take to be sure. With 18 calls the rough range is 49% to 88%. With 93 calls it narrows to 37% to 57%, and even that still brackets 50%.
I noticed that the page itself leans on this idea. Its own ticker table only lists names with 5 or more scored calls, which is a built-in admission that smaller counts are noise.
Fresh data will also change every figure above. The ledger shows the latest 100, so tomorrow’s version will drop the oldest rows. For a related test on how many reasons sit behind a call, see FuturesAI Signal Stacking: Does More Reasons Mean Better?.
A Four-Column Sheet You Can Rebuild in Ten Minutes
You can rebuild this tally in a spreadsheet with four columns: tag, direction, hit or miss, and date.
- Copy the ledger rows you can see into a sheet. One row per call.
- Put each tag in its own column, with a 1 where it applies.
- Add a direction column. Always filter by direction before you read a tag.
- For each tag and direction, write hits over calls. Mark anything under 30 calls as unproven.
- Re-run it when the ledger updates. A tag that survives two different months is worth another look.
Tips
- Read every hit rate together with its call count.
- Split by direction first, then by tag.
- Use your own tally from the date you start, not just the one shown here.
- Test with a small size or the demo account before risking real money.
Warnings
- Seven of the 100 ledger rows show no basis, so this tally covers 93 of them.
- One quarter of data cannot show that a tag will keep working.
- Leveraged futures can lose the full principal. Nothing here is advice.
- This counts direction hits only. Fees and slippage are not in the ledger.
FAQ
Does a MACD bearish tag mean a signal is more reliable?
Not on this evidence. In FuturesAI’s public ledger, calls tagged MACD bearish hit 13 of 18, but 10 of those were DOWN calls, and DOWN calls as a group hit 23 of 37. The tag and the direction are mixed together.
How many calls do I need before trusting a hit rate?
There is no fixed number, but under about 30 calls a hit rate can move by 20 points or more. At 18 calls, a 72% hit rate has a rough range of 49% to 88%.
Where can I see the ledger I used?
The public scoreboard on futuresai.io lists the latest 100 scored calls with their basis tags, and I read it on November 16, 2026. It updates, so your counts will differ from mine.
Sources
- FuturesAI: AI Prediction Scoreboard, scoring ledger (read November 16, 2026)
- FuturesAI: Exchange Fee Payback
Next step: build the four-column sheet today and write down the date you started, so your own sample begins now.
Counts are the author’s own tally of 93 tagged records on FuturesAI’s public ledger as read on November 16, 2026. Educational content, not investment advice.
๐ Related guides
