**30-second read:** When-issued (WI) trading is the grey market for a Treasury that has been announced but not yet auctioned: it trades yield-only, when-as-and-if-issued, from the announcement through the settlement date. It matters because the WI yield at the 1:00pm ET bidding deadline is the exact yardstick the auction is judged against: the auction tail is the stop (high yield) minus that WI mark. Watch WI cheapen into supply to build a concession, watch where it sits at the deadline, then watch whether the auction stops through it or tails it. Helious captures the real WI mark off the squawk at the deadline, not a stale screen print, which is why our tail dataset matches the desk. This is the signature edge: read the WI, and you've read the auction before the results hit.
From the moment Treasury announces a new issue, typically a few business days before the sale, dealers begin quoting it in the when-issued market, a forward that settles on the issue date, contingent on the auction actually happening. For a new coupon issue the coupon rate isn't struck until the auction clears, so WI trades on a yield basis only; the dollar price is backed out once the coupon is set. For a reopening (same CUSIP, additional size) the coupon is already known, so WI trades against the outstanding line. Either way, WI is a real, tradable market: bids and offers in yield, made two-way by primary dealers, not a theoretical fair value.
The practical point: WI lets the entire dealer community and their clients express a view on the new supply and hedge it before a single competitive bid is submitted. If the auction is cancelled, WI contracts void. That contingency is why it's a grey market: you're trading a security that does not yet legally exist, priced off the deadline yield it will be born at.
The single number that defines whether an auction was strong or weak is the tail: the auction's high yield (the stop) minus the WI yield prevailing at the 1:00pm ET competitive deadline (11:30am for bills). A positive tail, the stop printing above WI, means bidders demanded extra yield to absorb the size, a sign of soft demand. A negative tail is a stop-through: the auction cleared below where the security was trading, the mark of strong demand. On a coupon a tail of even one or two basis points is a genuine market event; on a 2s or 3s the whole distribution is tighter, so a 0.5bp tail already reads soft.
This is why the WI mark has to be exact and time-stamped to the deadline. A WI print grabbed even 90 seconds early, while the market is still cheapening into the bid, will understate a tail or manufacture a phantom stop-through. Everything downstream (the STRONG / IN LINE / WEAK verdict, the bid-to-cover read, the bidder split) is context; the tail off the correct WI mark is the headline. Pair this guide with How to Read a Treasury Auction Result for the full results-tape workflow.
In the days before an auction, the WI issue usually cheapens relative to its neighbors on the curve: dealers and investors sell it (or the surrounding tenor) to build an auction concession, so they can bid with a cushion and not get run over post-award. You see it as the WI yield backing up versus the on-the-runs, or the relevant sector steepening/cheapening on the curve board. A healthy concession going into the deadline is bullish for the auction outcome: it means the price has already adjusted to clear the size, which is exactly the setup that produces a stop-through.
The failure mode is the opposite: WI that richens into the auction, no concession built, dealers forced to reach. That's the classic tail setup, and it shows up most on the long end where duration risk is unforgiving: watch the 30-year bond and 20-year bond WI in the final hour. Front-end names like the 2-year note rarely need much concession because the risk is small and demand is rate-driven, so read the concession relative to each tenor's own history, not in absolute bp.
WI and the auction are two windows onto the same demand. Dealers who are short WI going into the sale have a natural incentive to bid the auction to cover, which supports the stop; dealers who are long WI have already absorbed inventory and can bid less aggressively. So the WI positioning of the street partly predicts the dealer take: a heavily short street tends to produce tighter stops and stronger covers, while a long, well-supplied street more often lets it tail.
After the results, the indirect share and direct share tell you who actually took the paper down versus how it was trading in WI. A stop-through delivered on a high indirect take is the cleanest strong print: real end-investor demand showed up where WI said it would. A stop-through carried entirely by dealers (high dealer allotment, thin indirects) is a weaker signal even with a nominally 'through' tail: the concession got bought by the backstop, not by the market, and that inventory often has to be re-marketed.
The core trade is the concession: sell the WI (or the sector) as it cheapens into the deadline, then cover on a strong auction as the issue snaps back richer post-award. A stop-through typically triggers an immediate richening: the 'sell the rumor, buy the fact' reversal, so the payoff is being short the concession into 1:00pm and flat-to-long into the snapback. A tail does the reverse: the issue cheapens further after a weak stop as dealers who got stuffed try to move inventory, and that cheapening can bleed into the whole tenor for the rest of the session.
Size the trade to the tenor. A 1bp tail on the 10-year note moves the complex; the same tail on a front-end name is noise. And never trade the snapback off a screen mid that's lagging the WI: by the time a delayed feed shows you the tail, the richening is half over. The edge is in reading WI live and being positioned before the results tape, not reacting to it. The [live desk](/) arms per-symbol alerts on the move so you catch the snapback in real time.
Most auction recaps you'll see quote a WI level lifted from a screen at some fuzzy time near the deadline, which is why their tails routinely disagree with the desk by a basis point or more. Helious pulls the real when-issued yield off the squawk at the competitive deadline and computes the tail against that mark, then posts the STRONG / IN LINE / WEAK verdict seconds after Treasury releases results: the same number the trading desk is reacting to. Our auction tails dataset and the per-tenor auction hubs are built on that captured WI, not a reconstruction.
That's the signature edge, and it compounds: a correct WI mark means a correct tail, which means a trustworthy history of how each tenor has been clearing, so you can see when a 2bp tail is genuinely alarming versus in-line for a beaten-up sector. Read the WI concession into the deadline, know the exact mark the tail is struck against, and you're reading the auction before the wire does. See the methodology for exactly how the WI mark and tail are sourced and timestamped.
