Brian L. Roberts

Founder, CMCSA

In the role: Jan 1990 – present
Birth date: 28 June 1959 No source recorded here yet
Birth time: Unknown. A reading is issued only when it holds whichever hour Brian was born.

2026 reading

Weak

0 signals, 1 drag

Read for the calendar year from the birth date, one of five levels. Signals lift it, drags lower it. The rules stay in-house.

Brian has not had a Weak year leading CMCSA before. Across every leader on record, Weak years rose 50% of the time.

This month: October is a quiet month. Next Active month: November. This month’s lists
Low tide: December to May: the half-year before Brian’s birthday, where Weak years run weakest.
2027: Published 1 December 2026, sealed before the year starts.

2026 month by month

Active months are the six in which a year’s reading tends to show in the stock. A qualifying reading enters the Rebound list in an Active month that follows a month the stock did not rise. The outlined cell is this month; the move is the month’s price change.

Years in role by level

Strong: none
Firm: 2 of 3 rose
Even: 13 of 21 rose
Soft: 4 of 7 rose
Weak: none

All leaders, 1995–2025: Strong 76% of 256 · Firm 71% of 709 · Even 67% of 2,016 · Soft 59% of 310 · Weak 50% of 191.

Year by year

Every calendar year Brian led CMCSA with a reading on record, and what the stock did that year. Past years are read with the rules now fixed; the return is the stock's from January to December.

CMCSA rose in 19 of 31 years.

Brian L. Roberts at CMCSA, by year
Year	Reading	Stacked	CMCSA	Result
2025	Firm	1 signal, 0 drags	−17.3%	Fell
2024	Even	0 signals, 0 drags	−11.8%	Fell
2023	Even	0 signals, 0 drags	+29.1%	Rose
2022	Even	0 signals, 1 drag	−28.7%	Fell
2021	Even	0 signals, 1 drag	−2.2%	Fell
2020	Soft	0 signals, 1 drag	+19.1%	Rose
2019	Even	0 signals, 0 drags	+34.8%	Rose
2018	Soft	0 signals, 1 drag	−13.2%	Fell
2017	Even	0 signals, 0 drags	+17.9%	Rose
2016	Firm	1 signal, 0 drags	+24.5%	Rose
2015	Even	1 signal, 1 drag	−1.1%	Fell
2014	Even	1 signal, 1 drag	+13.5%	Rose
2013	Even	0 signals, 0 drags	+41.7%	Rose
2012	Even	0 signals, 0 drags	+60.8%	Rose
2011	Even	0 signals, 0 drags	+10.0%	Rose
2010	Soft	0 signals, 1 drag	+33.1%	Rose
2009	Even	1 signal, 1 drag	+1.7%	Rose
2008	Soft	0 signals, 1 drag	−6.6%	Fell
2007	Even	0 signals, 0 drags	−35.3%	Fell
2006	Soft	0 signals, 1 drag	+63.3%	Rose
2005	Even	0 signals, 0 drags	−22.1%	Fell
2004	Firm	1 signal, 0 drags	+1.5%	Rose
2003	Even	0 signals, 1 drag	+39.1%	Rose
2002	Soft	0 signals, 1 drag	−34.5%	Fell
2001	Even	0 signals, 0 drags	−12.9%	Fell
2000	Even	0 signals, 0 drags	−13.7%	Fell
1999	Even	0 signals, 0 drags	+66.8%	Rose
1998	Even	0 signals, 1 drag	+80.7%	Rose
1997	Even	0 signals, 1 drag	+81.7%	Rose
1996	Soft	0 signals, 1 drag	+0.5%	Rose
1995	Even	0 signals, 0 drags	+15.1%	Rose

Readings: One of five levels a year, with how many signals and drags stacked, computed from the birth date over the thirteen possible birth hours and issued only where at least ten agree. Built 2026-10-16 from research export f0c9e40. The rules stay in-house.
Stock data: Calendar-year and monthly % change of adjusted close. Volatility % = annualised stdev of the 12 monthly returns. No prices are included. Adjusted close is split-adjusted, so a stock split never appears as a move in anything plotted here, and stock splits are not recorded as events.
Birth dates: 206 of the 319 people in this dataset now have at least one source recorded here, across 207 of the 325 leader-company records — six people ran two companies and are stored once for each. Of those sources — all 207 of them — 204 are a stated age read out of an SEC proxy statement, which bounds a birth year and says nothing about the day; one is an age stated in a press report; two state the day itself, one from an official gazette read by a person and one from an encyclopedia. The day is what every reading on this site is calculated from, so for all but those two birth dates no source here backs the part that is used. The 204 records the pass produced were not reviewed by a person: each source names who read it. Six findings from the same pass are deliberately not entered, because in each one the date held here falls outside the bound the filing implies, and telling an ordinary record-date lag from a real disagreement needs a person rather than arithmetic. An absent entry still means no source is recorded here yet.
Research export: Version 2, generated 2026-10-16 from research commit 7d490ef with uncommitted changes

Research and exploration only. Not investment advice. A reading describes a leader's year from their birth date, not the stock.