Usenet completion is the share of articles your setup actually retrieves compared with what an NZB requested. You can advertise multi-year retention and still fail downloads when Message-IDs vanished after takedowns, when indexers handed you stale pointers, or when your only provider shares a spool that already lost the segments. Completion is the metric that separates a working queue from brochure theater.
Retention describes how long a provider keeps articles in principle. Completion describes whether those articles are still there for your grab tonight. The gap between the two is why stack design, regime mix, and indexer freshness matter as much as coupon codes. For a sibling deep dive on percentages, see completion rate explained.
Definition in plain language
When SABnzbd or NZBGet finishes a job, compare completed articles against the NZB’s list. A healthy run reaches one hundred percent before PAR repair or with minor PAR work that rebuilds small gaps. Large missing counts mean completion failed upstream of unpack, no matter what marketing claimed about library depth.
PAR volumes exist because Usenet was never perfect. Parity can reconstruct limited loss if enough recovery blocks survive. PAR cannot invent articles your provider never had. Treat repair success as a subset of completion, not a replacement for backbone choice.
Why completion is not the same as retention
Long retention windows still lose individual Message-IDs when rights holders or platform policy remove posts. DMCA-heavy primaries and NTD-oriented primaries see different holes in the same release era. Two users with identical retention numbers on paper can see different completion because their takedown exposure differs.
That is why regime-diverse fill exists as an optional second server, not as a license to stack twin Highwinds brands. Fill tries alternate spools when Priority 0 misses. It does not change what your indexer discovered. Indexers discover; providers complete.
What drives good completion at home
Start with one strong primary on SSL port 563, tuned connections, and enough disk for repair temp files. Pair indexers that publish fresh NZBs for the content you actually fetch. Add optional fill on a non-overlapping backbone only when history shows regular missing articles on otherwise good posts.
Free or trial-tier providers often look fine on speed tests and fail on older or contested posts. Before annual prepay, run real NZBs from your indexers through a seven-day window and note failure modes. Our ranking lens favors evidence over banners; read how we rank Usenet providers.
How fill and policy mix interact
US-facing Highwinds primaries are common first picks for Americans. EU users often anchor on Eweka for NTD-class behavior. When misses cluster on one policy mix, Express or block fill on another regime can lift completion without a second unlimited bill on the same spool.
Fill improves odds; it does not guarantee physics. Optional Priority 1 servers should stay optional so you do not burn gigabytes when the primary already completed. Walk through DMCA primary with NTD fill for the policy framing, then avoid incomplete NZBs for step-by-step fixes.
Measuring completion without vendor hype
Vendor “100% completion” slogans rarely include your indexer mix, your content age, or your geographic path. Build a simple miss log: date, release name, missing count, which server failed, whether fill engaged. After a month you know if you need fill, better indexers, or client tuning.
Matrix comparisons on the homepage help narrow primaries; your log validates the choice. When misses are zero on a tuned primary, skip fill shopping entirely and spend the budget on indexer stability instead.
Re-test after major client upgrades or ISP changes. Completion problems that appear overnight are often DNS, SSL inspection, or a full temp disk rather than a dead backbone.
FAQ
- Is 100% completion marketing real?
- Treat perfect claims skeptically. Completion is contextual to content age, takedowns, and your indexer sources.
- Does fill guarantee completion?
- No. Fill adds a second chance on another spool when articles still exist there.
- Can indexers cause incomplete downloads?
- Yes. Stale or partial NZBs fail even on excellent providers. Isolate by retrying with a fresher NZB or another indexer hit.
- Do more Highwinds accounts improve completion?
- Not meaningfully. Twins share spool fate; use non-overlapping fill instead.