Methodology

LaunchTape is a promotional market simulation. Prices are fictional indicators produced by a published model, not valuations, and no part of this product is a security or an investment. This page describes exactly how every number on the terminal is produced.

What is real and what is modelled

The distinction the product refuses to blur: attention is real and prices are modelled.

• Views, sentiment votes, website clicks, shares, watchers, momentum purchases and claims are counts of events that actually happened. None of them is ever generated, seeded or simulated.
• Prices, price history, indices and index levels are produced by the platform market model described below. They are labelled as model data everywhere they appear.
• A metric with no meaningful value is hidden rather than shown as zero. Hiding is a presentation rule; the underlying number is never altered.

How a price is produced

Three layers combine into the number on screen.

1. Baseline strength

Every listing receives a deterministic baseline score between 5 and 25, derived from a hash of its ticker. It is stable for the life of the listing, is not purchasable, and does not depend on the company's real size, revenue or funding.

2. Paid momentum

Each successful momentum order adds units that decay exponentially with a 72 hour half life. Active momentum is recomputed from the complete order ledger on every read, never incremented, so two purchases landing at the same moment both count.

activeUnits = purchasedUnits × 0.5 ^ (ageHours / 72)
marketStrength = baselineScore + Σ activeUnits
fundamentalPrice = 1 + 2.5 × √marketStrength

The square root is deliberate: the first euro spent on a quiet listing moves it further than the hundredth.

3. Ambient market motion

Displayed prices fluctuate continuously so the terminal is alive with no traffic at all. This motion is deterministic - derived from the ticker, the sector and the current UTC time bucket - so every visitor sees the same market at the same instant, and the server can reproduce any past value. It never writes back into the momentum ledger.

The motion has two timescales: a slow correlated drift that gives the 1H, 24H and 7D windows something to measure, and fast bounded jitter clamped to ±0.8% that makes prices tick. Both are scaled by a per-listing volatility coefficient between 0.35 and 1.8. Ambient motion is kept small enough that even the smallest package produces a visibly larger move than the weather.

Price history

Charts are built from persisted snapshots taken every five minutes, plus the live computed price as the final point. A five second visual tick is never persisted as though it were a trade.

Listings seeded at launch have model-generated history so their charts are not blank. That history contains no paid momentum and no visitor events, because neither existed. It is model output, labelled as such, and it is the only history any listing starts with.

Rolling windows

The market never closes, so there is no open, close or daily bell. Every change is measured against the price at now − window over 1H, 24H, 7D, 30D and ALL. A window younger than the listing cannot be computed honestly, so it is omitted and replaced by SINCE LISTING rather than printed as 0.00%.

Heat

Heat measures genuine attention and is entirely separate from paid momentum. Each real event contributes a weight that decays with a 6 hour half life, and the decayed sum is compressed for display:

heatRaw = Σ weight × 0.5 ^ (ageHours / 6)
heatScore = min(100, round(20 × log₁₀(1 + heatRaw)))

Current weights: qualified view 1, sentiment vote 3, website click 4, share 6, watch 8. Heat influences discovery surfaces and can briefly raise a listing's ambient volatility. It never becomes paid momentum.

A page view counts once it has been genuinely visible for 3 seconds or the visitor interacted, which is what makes heat resistant to refreshing.

Sentiment

One effective vote per anonymous browser identifier per listing per 24 hours, enforced on the server. The percentage is published only once at least 5 votes exist.

Indices

LT100 and the sector indices are equal-weighted averages of member price movement. They are platform model data for ambience and discovery, and are not investable, tradeable or representative of anything outside this website.

Market events

The live tape carries payments, claims, visitor activity and model transitions such as a new all-time high or a sector leader change. Events fire on real state transitions only, with cooldowns so a price oscillating across a threshold does not repeat the same headline. No event is ever inserted to make the feed look busier.

Display thresholds

Counts are published only above a floor, so the terminal never shows "1 online" or "€0 volume". Current floors: online 10, viewers per listing 3, paid volume 50 EUR, orders 5, sentiment votes 5, watchers 3. These are presentation rules; the stored data is unchanged.

Parameters

Every coefficient above is operator-configurable and may change. Defaults ship as 72 hour momentum half life and 6 hour heat half life; this page always renders the values currently in force.

Audio credits

The optional trading-floor ambience is built from the "Pit trading" recordings by fugacious (freesound.org), licensed under CC BY 4.0, edited into a continuous loop. The market bell is synthesised.