Five months of calls, scored against the market

2026-10-31 · Natkal Research

Through July, Max beat a no-skill guess by 11 to 15 points. Since August, no tier has.

Scored 31 Oct 2026 through the 30 Oct close · calls issued 10 May to 29 Oct · same rule as our 1 August benchmark.

Did the calls beat a no-skill guess?

Hollow dot: the best fixed guess (call everything up, or everything down). Solid dot: how often the council was right. Green means it beat the guess.

Calls issued before 1 Aug20%30%40%50%60%70%80%90%Lite30 days-6.360 days-12.490 days+2.8Pro30 days+0.060 days-0.590 days+1.4Max30 days+11.360 days+13.190 days+15.4Ultra-Max30 days-17.360 days+6.090 days+7.8edgeCalls issued from 1 Aug20%30%40%50%60%70%80%90%Lite30 daysno live calls60 daysno live calls90 daysno live callsPro30 days-18.460 days-15.290 daysnot matured yetMax30 days-9.160 days-30.090 daysnot matured yetUltra-Max30 days-10.360 days-10.290 daysnot matured yetedge
best fixed guesscouncil, beat the guesscouncil, lost to the guess

Where the fixed guess comes from

What the called stocks actually did. The bigger of green or red is the fixed guess. Before August the split was even, so a guess was weak. Since August most of the stocks we called fell, so “always down” was hard to beat.

Calls issued before 1 AugLite30 days26%31%44%always down60 days27%23%50%always down90 days36%22%42%always downPro30 days44%30%26%always up60 days53%21%26%always up90 days51%19%29%always upMax30 days38%26%36%always up60 days42%21%36%always up90 days40%22%38%always upUltra-Max30 days53%26%21%always up60 days34%29%37%always down90 days31%27%41%always downCalls issued from 1 AugLite30 daysno live calls60 daysno live calls90 daysno live callsPro30 days20%40%40%always down60 days27%12%61%always down90 daysnot matured yetMax30 days24%35%42%always down60 days18%27%55%always down90 daysnot matured yetUltra-Max30 days15%35%51%always down60 days24%28%49%always down90 daysnot matured yet
rose more than 5%moved less than 5%fell more than 5%

The market backdrop

SPY, adjusted for dividends, since the first call. SPY kept rising after 1 August while most of the stocks we called fell, so the August problem is which stocks were called and how, not the market turning.

0%+5%1 Aug05-0806-0907-1008-1009-0910-08+3.0%

Every call

One dot per call. Left to right: how strongly the council leaned Downside or Upside. Up and down: what the stock did. Dots in the shaded corners were right.

Calls shown–
Council right–
Best fixed guess–
Edge–
rightwrongmoved less than 5%Returns beyond ±60% are pinned to the edge. Calls with no recorded conviction sit at the halfway mark.
What each number means
Council right
Of the calls where the stock moved more than 5%, the share that moved the way the council called.
Best fixed guess
Call every one of those stocks the same way, all up or all down, whichever scores better on the same stocks and days. No research needed; it only reflects the market’s drift.
Edge
Council right minus best fixed guess, in points. Above zero means the calls knew something the drift did not.
Horizon
Days after the call when the price was checked. We buy at the next open and sell at the first close on or after 30, 60 or 90 days, on split- and dividend-adjusted prices.
What we are changing
  • Splitting the August drop by market, by what we shipped, and by runs with part of the council missing.
  • Asking for performance relative to the market, so a call is judged on beating SPY, not on going up.
  • Letting the council make no call when conviction is weak, and leaning on the Downside list, which held up in both periods.
  • Weighting council members by their track record, counting repeated calls on one stock once, and publishing this every month.
Read with care
  • Repeated calls on one stock are not independent. This is one market period, and 90-day results only exist for calls made before about 1 August.
  • Research calls, not investment advice. Past results do not predict future ones.