Volatility Emergence Lab: Reader Guide

This app exists to make one market idea visible:

Volatility is often the temperature of price discovery.

When everyone knows the same thing at the same time, price should stay close to fair value. When an asset is new, strange, hard to value, or still being understood, investors disagree. They trade against each other. Price moves. Volatility appears.

That does not mean all volatility is good. Some volatility is information. Some is fragility. The point of the simulation is to help you see the difference.

Start Here

Open index.html.

You will see three assets, all starting at 100.

Near-perfect info: the control case. Think of this like a T-bill-style asset. Low uncertainty, no meaningful disagreement, no herding.
Emerging asset: the main lesson. A new asset with high uncertainty, learning, social feedback, and real upside potential.
Selected asset: the asset you control with the sliders.

The app is trying to answer a simple question:

Why do newly emerged assets often have high volatility, and why can that volatility coexist with real upside?

The answer is not “because people are irrational.”

The better answer is:

The market is still learning what the asset is.

How To Use This Guide

Read the first half if you want the plain-English version.

Use the Here be Dragons sections if you want the model mechanics, formulas, and risk details.

The app itself is meant to be understood visually first. The math comes later.

The Core Idea

A mature asset with clear cash flows, deep liquidity, and shared information should have lower volatility. There is less to argue about.

A newly emerged asset is different.

People disagree on basic questions:

What is it?
How big can it get?
Who will own it?
What is the right comparison set?
Is the current price information, noise, or both?
Is volatility a sign of opportunity, or a sign that liquidity is breaking?

That disagreement becomes price movement.

Volatility is not added to the model from the outside as decoration. It emerges from the market trying to clear different beliefs.

What You Are Looking At

Asset Comparison

The main chart compares the three assets.

The near-perfect-information asset should stay calm. It is the baseline. If that line is cold and stable, the model is doing its job.

The emerging asset should run hotter. It has uncertainty, learning, and upside. The market does not agree on what it is worth yet.

The selected asset lets you test the same logic yourself.

The shaded one-sigma ranges show volatility intuition for each asset. They are not forecasts. They show how much room the current realized volatility implies around each path.

Trailing 90-Day Vol

This chart shows market temperature through time.

If an asset is in real price discovery, its realized volatility should usually be higher than the near-perfect-information benchmark.

That does not automatically mean the asset is broken. It means the market is still processing uncertainty.

Trailing 90-Day Return

This chart sits next to volatility because volatility without payoff context is incomplete.

High volatility can mean pain. It can also mean the payoff range is still wide. For emerged assets, that wide range is the whole point.

Rolling Return Correlation

This compares the daily return correlation between the near-perfect-information asset and the emerging asset.

The windows are:

30d
90d
180d
365d

The short window should move around more. The long window should be steadier.

This is a risk view. Return correlation matters more than price-level correlation because risk managers care about what moves together day to day.

Belief Cloud

Each dot is an agent.

The dot’s vertical position shows whether that agent thinks the selected asset is cheap or expensive relative to price.

Higher means bullish valuation.
Lower means bearish valuation.
Green means bullish.
Red means bearish.
White means neutral.

The cloud is the emotional center of the app. It shows disagreement as a living distribution, not a single number.

Learning Compression

This shows information noise falling over time when learning is active.

The idea is simple: early in a new asset’s life, everyone is guessing with weak information. Over time, information spreads. Disagreement can compress. Volatility can fall.

That is the new-tech story. Early uncertainty, then learning.

Risk Comparison

The risk chart views compare the three assets across return, volatility, drawdown, Sharpe, tail loss, and path outcomes.

Use it as a sanity check.

If the emerging asset has much higher volatility than the control asset, that is expected. If it also has much higher upside in the simulated path, that supports the core lesson: volatility is the cost of entering the market before certainty arrives.

The path odds, like chance of +100% or chance of -50%, are realized frequencies inside the simulation. They are not forecasts.

Why We Built This

Most financial dashboards show volatility as a risk statistic.

That is useful, but incomplete.

For emerged assets, volatility is also information about the market’s state of knowledge.

Low volatility can mean:

The asset is well understood.
The upside range is narrow.
The market agrees.
Liquidity is stable.

High volatility can mean:

The asset is still being understood.
Investors disagree about fair value.
Price itself is becoming information.
The upside and downside range is still wide.

But high volatility can also mean:

Leverage is forcing selling.
Liquidity is disappearing.
The market is breaking, not learning.

The app is designed to keep those two ideas separate.

Information volatility is price discovery.

Liquidity and forced-selling volatility is fragility.

The Simulation Loop

Each simulated day does roughly this:

Fair value moves.
Agents receive private signals about fair value.
Agents update their social belief.
Agents form valuations.
Agents become bullish, bearish, or neutral.
Their demand pushes market price.
Liquidity can absorb or amplify the move.
If drawdowns are large and leverage is active, forced selling can appear.
Metrics and charts update.

The important point: price is not just a line. It is the result of many agents disagreeing, learning, and reacting to each other.

The Three Assets

Near-Perfect Info

This is the control case.

It has:

Very low fundamental volatility.
No information asymmetry.
No herding.
No friction.
Deep liquidity.

It should stay cold.

If this asset were volatile, the model would be confusing noise with price discovery.

Emerging Asset

This is the fixed demonstration asset.

It has:

High early uncertainty.
Learning over time.
Social feedback.
Positive fair-value drift.
Real upside potential in the deterministic scenario.

It should run hotter than the benchmark.

That is the main demonstration.

Selected Asset

This is your sandbox.

You can change the assumptions and watch the system respond.

If you raise information asymmetry, disagreement should rise.

If you raise herding, beliefs should become more synchronized.

If you lower liquidity, the same disagreement should move price more.

If you turn on forced selling, drawdowns can become more dangerous.

Controls

Fundamental Drift / Trend

This is the asset’s underlying growth or decay.

Higher trend means fair value has more room to rise over time. That creates a better environment for upside discovery.

Fundamental Vol

This is real uncertainty in fair value itself.

A business, protocol, or asset can have real underlying uncertainty. Not all price movement is just market psychology.

Information Asymmetry

This controls how different the agents’ private signals are.

Higher asymmetry means agents see the asset differently. That should increase disagreement.

Discovery / Learning

This controls how quickly information noise compresses over time.

When learning is high, early uncertainty should fade faster.

Animal Spirits / Herding

This controls how much agents let price action and social belief influence valuation.

High herding does not simply mean “more volatility forever.”

It means beliefs synchronize more. Sentiment can become more persistent. Tails can cluster.

Participants

This changes the number of visible agents.

More participants alone should not create volatility. More participants with noisy information create a richer belief distribution.

Market Liquidity Depth

This controls how much flow price can absorb.

Deep liquidity dampens disagreement. Thin liquidity turns the same disagreement into larger price moves.

Execution Friction

This adds imperfect clearing and market microstructure noise.

Small frictions can matter when liquidity is weak.

Leverage / Forced Selling

This adds drawdown-triggered selling pressure.

This is where volatility can stop being healthy price discovery and start becoming fragility.

Target Speed

This controls the target playback speed. The run starts slower while the market is unstable, then accelerates toward the target as information compresses and the path becomes easier to read.

One model step is one trading day.

How To Read A Good Run

A good run should show:

The near-perfect-information asset staying cold.
The emerging asset running hotter.
The emerging asset having wider payoff outcomes.
The belief cloud moving and compressing as the market learns.
The risk chart views showing more drawdown and more upside for the emerging asset.

This matters because investors often treat volatility as a single bad thing.

It is not.

Volatility can be the price of entering an asset before the market agrees on what it is.

What The App Is Not Saying

The app is not saying:

High volatility is always good.
New assets always go up.
Disagreement guarantees upside.
Volatility should be ignored.
This is a trading model.
The path frequencies are forecasts.

The app is saying:

When an asset is newly emerged, high volatility can be a natural result of price discovery under uncertainty.

That volatility can be worth underwriting if you believe you understand the asset better than the market.

But if volatility comes from leverage, forced selling, or liquidity withdrawal, it is a different thing. That is fragility.

Key Assumptions

The asset has a fair value.

The simulation assumes there is a real underlying value process.

In the real world, fair value is not observable. That is part of the problem.

Agents do not see fair value perfectly.

They receive noisy signals.

That creates disagreement.

Price can become information.

Agents partly learn from price action.

This is realistic. Investors often treat price as evidence, especially when the asset is new and fundamentals are hard to anchor.

Liquidity is not constant.

Liquidity gets worse when recent volatility rises and herding is high.

That is important. Markets often look liquid until they are stressed.

Learning can compress uncertainty.

As information spreads, signal noise can fall.

This is why some new assets become less volatile over time.

Forced selling is different from discovery.

When leverage is active, drawdowns can trigger selling that has little to do with new information.

That is not healthy volatility. It is mechanical pressure.

Here be Dragons: Fair Value

Fair value follows a geometric random walk with optional drift.

V(t+1) = V(t) * exp(drift / 252 - 0.5 * sigma_F^2 + sigma_F * z)

Where:

V(t) is fair value today.
drift is annualized trend.
sigma_F is daily fundamental volatility.
z is a seeded normal shock.
252 is the number of trading days in a year.

The -0.5 * sigma_F^2 term is the standard lognormal adjustment.

The app uses a seeded random number generator. The browser model and Python model can be compared for the same seed.

Here be Dragons: Private Signals

Agents do not observe fair value cleanly.

Each agent receives a private signal:

private_signal_i = V * (1 + private_noise + stale_bias)

The noise level depends on effective information noise:

effective_info = information_asymmetry / (1 + learning * t / learning_timescale)

So when learning is high, effective information noise falls over time.

This is what the learning compression chart shows.

Each agent carries a social_belief.

It updates from:

The agent’s prior belief.
Recent price momentum.
The crowd’s mean belief.
Herding strength.
The agent’s own reflexivity.

The simplified idea:

social_belief_i(t+1)
  = prior_social_belief
  + herding_speed * (social_target - prior_social_belief)
  + small_seeded_noise

This matters because the belief cloud is not random decoration.

It is a visual projection of real model state.

Here be Dragons: Agent Valuation

An agent’s valuation blends private information and price/social information.

valuation_i =
  (1 - alpha_i) * private_signal_i
  + alpha_i * blended_price_signal_i

Where alpha_i rises with herding and reflexivity.

The price signal includes momentum. The social signal includes the agent’s social belief.

When herding is low, private signals dominate.

When herding is high, price and social belief matter more.

That is how price becomes information.

Here be Dragons: Price Formation

The model price return combines several forces:

price_return =
  consensus_return
  + order_imbalance_impact
  + disagreement_noise
  + execution_noise
  + social_pressure
  + fair_value_anchor
  + forced_selling

This is the heart of the simulation.

Consensus pulls price toward the agents’ average valuation.

Order imbalance moves price when demand is one-sided.

Disagreement noise rises when private signals are dispersed.

Execution noise reflects friction.

Social pressure reflects herding.

The fair-value anchor keeps the model from becoming an untethered cartoon.

Forced selling appears when drawdowns are large and leverage is active.

Here be Dragons: Liquidity

Liquidity is dynamic.

effective_liquidity =
  base_liquidity / (1 + k_vol * recent_vol + k_herd * herding^2)

The idea is simple:

When volatility rises and the crowd is synchronized, liquidity gets worse.

That makes the same disagreement move price more.

This is why liquidity matters so much in new assets.

Here be Dragons: Forced Selling

Forced selling activates after drawdown crosses a threshold.

forced_intensity =
  leverage
  * max(0, drawdown - drawdown_trigger)
  * average_forced_selling_exposure

Forced return is capped so the simulation stays bounded.

This is deliberate. The goal is not to create a crash machine. The goal is to show why leverage-driven volatility is different from information-driven volatility.

Here be Dragons: Perfect Information Control

The model has one hard control case:

if information_asymmetry == 0
and herding == 0
and friction == 0:
    price = fair_value

This is important.

It proves the app is not creating volatility just because there are many agents or because time is passing.

If information is perfect and friction is zero, price tracks fair value.

Here be Dragons: Core Metrics

Active Vol

Active return is price return minus fair-value return.

active_return = price_return - fair_value_return
active_vol = stdev(active_return) * sqrt(252)

This is the cleanest measure of market-generated volatility.

Price / Fair-Value Vol Ratio

vol_ratio = price_vol / fair_value_vol

This is not the same thing as active volatility.

Active volatility is a volatility number.

The ratio is a multiple.

Return Correlation

The app uses return correlation, not price-level correlation.

corr(price_returns, fair_value_returns)

For the rolling correlation chart, it compares daily returns between the near-perfect-information asset and the emerging asset.

VaR And Expected Shortfall

Daily VaR is shown as a realized percentile of simulated daily returns.

Expected shortfall is the average of the worst tail observations.

These are path statistics from the simulation, not predictions.

Chance Of +100% Or -50%

These are rolling 1-year frequencies inside the simulated path.

They answer:

In this run, how often did a 1-year window double?
In this run, how often did a 1-year window lose half?

They do not answer:

What is the real-world probability this asset doubles?

Here be Dragons: Research Backbone

The app is not trying to be a full academic model.

It borrows the shape of several serious ideas:

Grossman and Stiglitz: information is costly, so prices are not perfectly informative.
Shiller: prices can move more than fundamentals.
Hong and Stein: markets can underreact, trend, and overreact.
Kyle and Glosten-Milgrom: information and liquidity are connected.
Pastor and Veronesi: uncertainty can fall as markets learn.
Margin spiral literature: leverage and liquidity can amplify crashes.

The app compresses those ideas into a visual explainer.

That means it should be judged as a teaching model, not as a trading engine.

The Investor Takeaway

When you see high volatility in a newly emerged asset, do not stop at “this is risky.”

Ask better questions:

Is volatility coming from information discovery?
Is the market learning?
Are beliefs dispersed or synchronized?
Is upside still genuinely wide?
Is liquidity deep enough to survive disagreement?
Is leverage turning uncertainty into forced selling?

Volatility is a warning light.

It is also a signal.

The work is figuring out which one you are looking at.

Volatility Emergence Lab