Keith Kaplan has invested $17 million into his own AI research and tools. He's
built a platform that 180,000 people worldwide use in the stock market. Now he
says there's a handful of stocks you need to buy before...



<[link removed]>

September 02


 



Major Buy Alert Issued for September 30th

Here is the Stock →
<[link removed]>

 







<[link removed]>
Keith Kaplan has invested $17 million into his own AI research and tools.
He's built a platform that180,000 people worldwide use in the stock market. Now
he says there's a handful of stocks you need to buy before September 30, to set
yourself up for1,000% potential returns in the near future.

Here are the stocks you need to buy for the chance to profit.
<[link removed]>


This ad is sent on behalf of TradeSmith at 1125 N. Charles Street, Baltimore,
Maryland 21201. If you’re not interested in this opportunity, pleaseclick here
<[link removed]>
.
 

 





This Week’s Top Market Stories (Ad)
Hot
#1 Stock to Own as Trump Launches Historic Mission Backing “Medical AI”
<[link removed]>
Stansberry Research


Three Stocks to Sell Before You Buy Anything
<[link removed]>
Behind the Markets


A Better Retirement Stock Than Nvidia
<[link removed]>
Chaikin Analytics








THE HILL REPORT

A Signal Shared Is a Signal Consumed

Connor Hill · InsightfulWord · September 2, 2026

There is a property of market signals that has no equivalent in most other
kinds of useful information, and it determines almost everything about what a
distributed signal is worth.

A recipe does not stop working because a million people have it. A map does
not become less accurate as it is copied. A trading edge does. It is consumed
by being acted upon, and the consumption is proportional to how many people act.

The mechanism is not mysterious. An edge exists because an asset is mispriced
relative to something. Buying it moves the price toward correctness. Enough
buying eliminates the gap, and the gap was the edge.

This is not a theory about whether markets are efficient. It is arithmetic
about what happens when purchase orders arrive. The price at which a hundred
people can buy and the price at which a hundred thousand people can buy are
different prices, and the difference is a cost borne by whoever is later in the
queue.

The effect has been measured directly and at scale, in a setting where the
timing of disclosure is precisely known: the academic finance literature, where
predictive patterns are discovered, then published, and where returns before
and after publication can be compared.

The result is one of the more robust findings in empirical finance, and it is
unambiguous in direction. Published predictors decay, and the decay is larger
than what mere sample variation accounts for.

A commercial signal service is the same phenomenon with a shorter timeline and
a larger simultaneous audience. Publication in a journal reaches a specialist
readership over months. Publication to a subscriber list reaches everyone
within hours.

What follows is what happens to an edge after it is published, the volume
signature that discovery leaves in the data, why capacity is the constraint
nobody advertises, why the least tradable names decay slowest, and what a large
simultaneous audience does to execution.

What Happens to a Published Edge

The definitive study on this took the question directly and answered it with
an unusually clean design.

Two researchers assembled ninety-seven cross-sectional return predictors from
eighty peer-reviewed studies — the accumulated set of published patterns
claiming to forecast relative stock returns. For each one, they identified the
sample period the original authors used and the date of publication.

That produces three distinct windows. The original in-sample period. The
interval between the end of that period and publication, when the pattern
existed in the world but was not yet widely known. And the period after
publication.

Returns fell in each successive window. Out-of-sample but pre-publication
returns were about twenty-six percent lower than in-sample. Post-publication
returns were about fifty-eight percent lower than in-sample.

The gap between those two figures is the informative part. If the decline were
purely statistical — the result of the original researchers having found
patterns that were partly chance — the out-of-sample decline would capture it,
and publication itself should change nothing. Instead publication produced a
further large decline, which the authors attribute to investors learning about
the pattern and trading it.

The estimate is that roughly a third of the total decline is attributable to
publication specifically. The remainder is the correction of statistical
over-fitting, which is a separate problem and was covered in this space before.

Two-thirds of an edge, then, tends to survive publication in a journal read by
specialists. That is the optimistic reading. The pessimistic one is that
fifty-eight percent of the original figure was never available to anyone who
learned about it from the paper.


The Volume Signature of Discovery

The same study looked at what happened in the market around publication, and
the trading data corroborate the return data in a way that is difficult to
explain otherwise.


📊 Market Snapshot

58 percent

The average decline in returns to published cross-sectional stock return
predictors after publication, relative to the original in-sample period, across
97 predictors drawn from 80 peer-reviewed studies. Out-of-sample but
pre-publication returns declined about 26 percent, implying that roughly a
third of the total decline is attributable to publication itself rather than to
statistical over-fitting. Trading volume in the affected portfolios rose after
publication. Source: McLean and Pontiff, Journal of Finance.


Support or oppose: should paid signal services be required to disclose
subscriber counts?

Supporters argue that the number of people receiving a recommendation is the
single most relevant fact for judging what it is worth, that the provider knows
it precisely, and that withholding it while advertising past performance omits
the variable that most determines whether the past can repeat. Opponents answer
that subscriber counts are commercially sensitive, that they would be gamed the
moment they were mandated, and that the relevant capacity constraint depends on
the securities involved rather than on the headcount alone. Which is better?Hit
reply — one line is enough.

Trading volume in the affected portfolios rose measurably after publication,
on the order of a fifth by share volume. The difference in short interest
between the long and short sides of the portfolios widened by a multiple.

Those are the fingerprints of people acting on the information. Volume rises
because trades are being placed; short interest diverges because the strategy
involves shorting one side; and both changes appear after publication rather
than before.

The cross-section of the decay is equally telling. Patterns concentrated in
liquid, easily traded, low-idiosyncratic-risk stocks decayed more. Patterns
living in names that are expensive or awkward to arbitrage decayed less.

That gradient is exactly what a mechanical explanation predicts and exactly
what a purely statistical explanation does not. Over-fitting has no reason to
be worse in liquid stocks. Trading pressure has every reason to be.

The implication for reading any performance claim is direct: the question is
not only whether the pattern was real, but whether it lives somewhere that
arriving money can reach without moving the price.


Capacity Is the Constraint Nobody Advertises

Every strategy has a maximum size beyond which its own trading destroys its
returns, and the size is rarely stated because stating it caps the addressable
market.


Context — what a performance figure does and does not establish

A stated past return describes what a strategy produced in a period, under the
assumptions used to compute it, at whatever size it was run. It does not
establish what it would produce at a different size, in a different period, or
after the costs of executing it in size. Nor does it establish that any
particular participant achieved it, since a published series and a funded
account are different objects. Nothing here is a comment on any specific
service, company, sector or security, and none of it is a recommendation.

The constraint arises from market impact: buying pushes the price up, and the
push grows with the quantity relative to what the market normally trades. The
empirical regularity is that impact grows roughly with the square root of order
size relative to daily volume, which means doubling the size increases the cost
per share by about forty percent rather than doubling it.

That relationship is forgiving at small size and punishing at large size, and
the point at which it starts to bite depends entirely on the liquidity of the
names involved. A strategy in the largest listed companies has enormous
capacity. The same strategy in companies trading a few hundred thousand dollars
a day has almost none.

Institutional investors treat capacity as a first-order question, model it
explicitly, and close funds to new money when they judge it reached. That
practice is the clearest available evidence that the constraint is real, since
closing a fund forgoes fee revenue.

The arithmetic that matters for a distributed signal is simple to set up. Take
the number of recipients, multiply by a plausible position size, and compare
the total to the daily dollar volume of the security. If the product is a
meaningful fraction of a day's trading, the recipients are collectively the
market for that day, and the price they receive is a price they created.

Nothing about that requires anyone to behave badly. It follows from arithmetic
alone.


Why the Smallest Names Decay Slowest

There is a genuine counter-argument to all of this, and it deserves to be
stated properly rather than dismissed.

If an edge lives in securities that are hard to trade — small, illiquid,
expensive to borrow, or subject to constraints that keep institutions out —
then the arbitrage that would eliminate it is costly, and the edge can persist
for a long time.

The published evidence supports this. The predictors that decayed least after
publication were exactly those in the hardest-to-arbitrage names, and this is
the standard explanation for why some anomalies have survived decades of
documentation.

But the argument cuts both ways, and the second edge is sharper. The same
illiquidity that protects the edge from institutional arbitrage guarantees that
a distributed retail audience cannot capture it either.

A pattern that survives because it lives in a stock trading two hundred
thousand dollars a day is a pattern that cannot absorb a thousand buyers, let
alone a hundred thousand. The protection and the impossibility are the same
property.

So the two cases collapse into one answer. Where a signal is easily traded, it
decays quickly and the early recipients get most of what there is. Where it is
not easily traded, it persists — and cannot be acted on at scale. There is no
configuration in which a large simultaneous audience captures a durable edge in
a thin security.


What a Subscriber Base Does to Execution

The final mechanism is the one closest to the actual experience of anyone
acting on a distributed recommendation.

Recommendations are released at a moment. Everyone receives them at
approximately that moment. The orders therefore arrive together, in the same
direction, in a compressed window.

Correlated order flow in one direction is precisely the condition that
produces the largest price impact per share. A dispersed set of buyers across a
week costs far less in aggregate than the same buyers arriving within an hour.

Order type compounds it. A market order in a thin security during a burst of
correlated demand executes against whatever is resting on the offer, and the
resting quantity in such a name is frequently small.

The pattern is observable from outside. A thinly traded security that trades
many multiples of its normal volume on a single day, with a sharp move and a
subsequent partial reversal, has a well-known shape, and it can be checked
after the fact from public price and volume data.

The general observation is that a signal's value depends on how many people
have it and how easily the underlying can absorb them, and that both variables
are knowable. Neither appears in a performance figure. A past return is
computed as though the trades were free and the trader was alone, and a
distributed recommendation is neither.


The bill, not the debate

Published return predictors decline about 58 percent after publication, with
roughly a third of that attributable to publication itself, and trading volume
in the affected names rises when it happens. An edge is consumed by being acted
upon, and the more thinly traded the security, the less of it a crowd can
capture. Before acting on a distributed recommendation, do you know the daily
volume of what it names?Connor Hill reads every reply.


Sources checked: McLean and Pontiff — Does Academic Research Destroy Stock
Return Predictability?, Journal of Finance
<[link removed]> · Working paper
version with full tables
<[link removed]> · U.S.
Securities and Exchange Commission — investor bulletin on performance claims
and advertising <[link removed]> · Financial
Industry Regulatory Authority — communications with the public, performance and
projections
<[link removed]> ·
U.S. Securities and Exchange Commission — market structure data and trading
volume statistics <[link removed]> · Financial Industry
Regulatory Authority — short interest reporting
<[link removed]>


Connor Hill · InsightfulWord





You’re receiving this email from Insightful Word (IW), a brand of TerraTrance
Technologies, LLC.

Our mailing address: 200 Broadway Blvd NEAlbuquerque, NM 87102



Have a question or need assistance? Reply directly to this email or contact us
[email protected] <mailto:[email protected]>



The content of this email may not be copied, reproduced, forwarded, shared, or
distributed without the prior written consent of TerraTrance Technologies, LLC.

Privacy Policy <[link removed]>
Terms & Conditions <[link removed]>
Unsubscribe
<[link removed]>



© 2026 Insightful Word (IW). All Rights Reserved.