What is an execution gap for a bitcoin treasury company?

The difference between the average price a company discloses paying for bitcoin and what the order book would have allowed over the same days, executed in sized clips split across venues. It is an execution measurement, not a timing one: both sides cover the identical window.

How much did Strategy's execution gap cost across its largest buying weeks?

Across the 8 largest buying weeks between August 2025 and August 2026, the disclosed average sat 92.3 basis points above a clipped, venue-split execution over the same days — about $67.6M on $7.33B deployed, or roughly 704 BTC. On one of the eight weeks, Strategy's disclosed average beat the benchmark by 1.6bp.

Can a treasury company check this themselves?

Yes. Every input is public: the BTC Update table in each 8-K gives the date range, BTC acquired and average purchase price, and historical spot order books are commercially available. EnsoTrade's /v1/plan and /v1/route endpoints price the same comparison on live books.

EnsoTrade execution research

Order-book replay · Aug 2025 – Sep 2026

What bitcoin treasuries actually pay to execute

Name: Bitcoin Treasury Execution Gap — Strategy Inc (MSTR)
Creator: EnsoTrade

Every public treasury company discloses the average price it paid for bitcoin. Almost nobody checks that figure against the order book it was filled on.

We did, for Strategy Inc (MSTR) — the largest and most transparent of them. Across its 8 largest buying weeks, the disclosed average sat 92 basis points above what a clipped, venue-split execution would have paid over the same days. On $7.33B deployed, that is $67.6M, or about 704 BTC.

Every input is public. Date ranges, BTC acquired and average purchase price come from the BTC Update table in each 8-K. Prices come from historical spot order books for the same dates. Institutional taker fees are applied to our side; Strategy's disclosed figures are already inclusive of fees and expenses, per their own filing footnote.

92.3 bp

average gap, 8 largest weeks

$67.6M

on $7.33B deployed

704 BTC

at each week's own average

The eight weeks

Week	Disclosed avg	Order book allowed	Gap	USD
Jan 12 – Jan 19, 2026	$95,284	$94,470	86.2 bp	$18.16M
Jan 5 – Jan 11, 2026	$91,519	$91,534	−1.6 bp	−$0.20M
Dec 8 – Dec 14, 2025	$92,098	$91,107	108.8 bp	$10.55M
Dec 1 – Dec 7, 2025	$90,615	$90,163	50.1 bp	$4.80M
Nov 10 – Nov 16, 2025	$102,171	$99,893	228.0 bp	$18.63M
Aug 26 – Sep 1, 2025	$110,981	$110,025	86.9 bp	$3.87M
Aug 24 – Aug 30, 2026	$80,318	$78,646	212.6 bp	$7.70M
Aug 18 – Aug 24, 2025	$115,829	$114,499	116.2 bp	$4.10M
Total			92.3 bp	$67.60M

On one of the eight weeks, Strategy beat the benchmark. Jan 5–11, by 1.6bp. Same method, same code, same data. We are publishing it because a study in which the author's product wins every single time is not a study.

What this is not

It is not a timing claim. Both sides of the comparison deploy the same money across the same days. If Strategy had bought at different moments the number would move — so we removed that by matching their cadence rather than choosing our own.

It is not a claim that they execute badly. Strategy's 10-K states they execute through trade execution partners affiliated with their custodians, close in time to each capital raise, to limit drift between raising and deploying. That is a deliberate and defensible choice. A gap of 92bp against an idealised benchmark is what a competent desk without a dedicated execution router looks like — not a failure.

It is not the largest number available. Priced against a single order sent all at once, the same eight weeks return roughly $161M. We do not use that figure, because nobody sends $137M as one market order and a baseline nobody would choose is not a baseline.

Method

Filings. All 8-K filings from CIK 0001050446 since Aug 2025. The BTC Update table gives period start, period end, BTC acquired, aggregate price and average price. Rows are kept only where BTC × average = aggregate to within 2%, which rejects filings where the parser picked up holdings or share-count columns instead.
Books. Historical spot L2 order books for Binance, Kraken and OKX on the same dates, replayed from snapshot plus incremental updates.
Execution model. The week's total spread evenly across its days, sent in $5M clips, each clip filled against all three venues merged and ordered by fee-adjusted price.
Fees. Published institutional top-tier spot taker rates — Binance 4bp, Kraken 10bp, OKX 2bp — applied to our side. Strategy's disclosed averages already include fees and expenses.

Known limits. Intraday timing is not controlled — we sample two hours a day, so a desk that consistently picks better moments within a day would close part of the gap. Coinbase is not in the venue set, and Strategy custodies roughly 40% of its holdings there. Fee tiers are the published institutional rates; Strategy's negotiated rates are likely better, which would narrow the gap further.

Check it yourself

Every figure above is reproducible from public sources. The same comparison runs on live books through /v1/plan and /v1/route — pass a notional and a venue set and you get the one-shot cost, the clipped cost and the split. A free key is issued instantly at ensotrade.tech, no signup.

If you run a treasury and want your own filings measured this way, the workings for any single week are available on request.