Market simulation

Interactive simulation

What-if market simulation with user-set or assumed parameters. Never a measurement of any real market, platform, or seller.

A multi-seller logit-share demand model (analysis/market.py). Drag the sliders to see what a given individual-level effect size would imply for share, revenue, and consumer surplus -- every number below is INTERACTIVE SIMULATION, never a measurement of any real market, platform, or seller.

Simulation inputs

Interactive simulation

Seller count5

Bias-exposed population share0.5

Assumed effect size (log-odds)0.5

Favored seller quality gap0

Favored seller price189 USD

Every input here is a stated assumption, adjustable by you -- not fit to any real market. The default effect size mirrors the simulated-pilot H1 estimate on /results, but you can set it to anything.

Favored seller’s market share

Unexposed segment

9.4%

Exposed segment

14.7%

Overall share

12.0%

All sellers

Seller	Price	Quality	Overall share	Revenue
favored seller(favored)	$189	8.1	12.0%	$2,275,863
competitor 1	$120	8.3	53.3%	$6,399,762
competitor 2	$153	7.3	19.9%	$3,056,483
competitor 3	$187	8.6	10.5%	$1,953,291
competitor 4	$220	8.1	4.2%	$930,468

Consumer surplus (McFadden log-sum)

McFadden log-sum CS has no absolute-dollar meaning without an outside-good normalization; only the exposed-vs-unexposed delta is interpretable here.

Unexposed, per consumer

$-46.09

Exposed, per consumer

$-43.72

Exposure delta, per consumer

+$2.38

How this is wired

Each slider change calls POST /api/market, which shells out to the real analysis/market.py CLI (its stdin-JSON mode) with your chosen parameters and returns its exact output -- the same code path that produced results/market_bundle.json, so this page can never drift from the analysis pipeline’s own model.