Tie-Breaks: Minimum Surplus And Reference Price

A call auction picks the single price that lets the most shares trade. But "most shares" doesn't always point to a unique price. If two neighboring prices both clear, say, 350 matched shares, the exchange needs a second rule to decide which one actually prints — otherwise the auction algorithm isn't well defined.

The two-step tie-break

The first tie-break is minimum surplus: among all prices tied for maximum matched volume, pick the one with the smallest leftover imbalance (buy demand or sell supply that couldn't be matched at that price). Less unmatched interest means the price is doing a slightly better job of clearing the book. If minimum surplus also ties, the exchange falls back to a reference price rule — usually the price closest to the last continuous-market trade or the prior day's close — which anchors the print to the most recently observed "fair" level rather than an arbitrary tie-breaking coin flip.

Worked example: two prices, same matched volume

Suppose the pre-open book has these cumulative orders at four price levels (a buy limit order at 102 is willing to buy at 102 or lower; a sell limit order at 99 is willing to sell at 99 or higher):

Price	Cumulative buy volume	Cumulative sell volume	Matched volume	Imbalance
99	500	100	100	400 (buy)
100	450	350	350	100 (buy)
101	350	550	350	200 (sell)
102	200	500	200	300 (sell)

Matched volume peaks at 350 shares at both 100 and 101 — a tie on the primary rule. Minimum surplus breaks it: the imbalance at 100 is only 100 shares of unmatched buy demand, versus 200 unmatched sell shares at 101. The auction prints at 100, because that price leaves the least interest stranded.

If the imbalance had also tied at, say, 100 shares on each side, the exchange would move to the reference price. With a prior close of 99.80, a price of 100.00 sits closer to that anchor than 101.00 would, so 100.00 would win on that basis instead.

What this means in practice

Tie-breaks rarely change the print by more than a tick, but they matter for anyone reverse-engineering or predicting an auction price from the imbalance feed: two candidate prices with identical matched volume can look interchangeable until you check which one minimizes leftover imbalance. Index-tracking and closing-auction algorithms that estimate the likely clearing price need to replicate this exact sequence of rules, not just "find the volume-maximizing price," or their prediction will occasionally land one tick off.