WO2018136089A1

WO2018136089A1 - Method and platform for matching buying and selling orders

Info

Publication number: WO2018136089A1
Application number: PCT/US2017/014463
Authority: WO
Inventors: Christopher C. FERRERI; Michael Sacks
Original assignee: Opendoor Securities, Llc
Priority date: 2017-01-20
Filing date: 2017-01-20
Publication date: 2018-07-26

Abstract

A computer-implemented method and platform enable a multi-phase, all-to-all, call auction for at least one asset. The method and platform include initiating the auction for at least one asset; conducting a pricing phase for a first period of time; calculating and publishing a calibrated mid-curve for the at least one asset; conducting a breathing phase for a second period of time, during which one or more participants can evaluate the calibrated mid-curve and at least one buying order is entered and at least one selling order is entered; initiating a matching phase and submitting the at least one buying order and the at least one selling order; conducting the matching phase over a third period of time; identifying a market gap; generating a KeyChain order; and notifying the buyer and seller of a filled order.

Description

METHOD AND PLATFORM FOR MATCHING

BUYING AND SELLING ORDERS

Field of the Invention

[0001] The present invention involves matching buying and selling orders, identifying market gaps and trading assets between or among two or more participants. In many of the embodiments, the assets, including, but not limited to, financial assets such as U.S. Treasuries, are traded anonymously. The present invention is not restricted to assets that are financial instruments. It is also applicable to other assets, such as non-financial assets, including, but not limited to, event tickets, art, vehicles and collectables.

Background of the Invention

[0002] Executing a transaction in a market typically depends on matching a buyer and seller, or a buying order and selling order, as to, for example, asset, price and amount. For a fungible asset, efficiently matching a buyer and seller or their respective orders, and realizing the asset's liquidity, is a relatively common occurrence. However, for illiquid assets, or assets of lower liquidity, efficiently matching a buyer and seller or their respective orders, and unlocking liquidity, is more challenging.

Summary of the Invention

[0003] Shortcomings of existing techniques for matching buying and selling orders are overcome by the method and platform described herein, through which orders involving at least one asset, which can be of one or more types and/or from one or more sectors and from various participants (buyers and sellers), are efficiently matched. The present method and platform efficiently facilitate trades of assets and, in some embodiments, identify, generate and fill one or more executable orders. The present method and platform enhance the liquidity of an asset by lowering transactions costs typically associated with traditional techniques for matching orders.

[0004] A computer-implemented method conducts an anonymous multiphase, call auction of at least one asset. The method comprises:

(a) initiating the auction for the at least one asset;

(b) conducting a pricing phase for a first period of time;

(c) calculating and publishing at least one calibrated mid-curve for the at least one asset;

(d) conducting a breathing phase for a second period of time, during which one or more participants can evaluate the at least one calibrated mid-curve;

(e) conducting a matching phase over a third period of time;

(f) receiving at least one buying order and at least one selling

order; and

(g) anonymously matching one or more of the at least one buying order and one or more of the at least one selling order.

[0005] In one embodiment, the foregoing method further comprises filling one or more of the at least one buying order and one or more of the at least one selling order.

[0006] In another embodiment, the method further comprises:

(h) identifying one or more market gaps;

(i) generating one or more KeyChain orders to fill one or more of the one or more market gaps; and j) completing a KeyChain of one or more buying orders, one or more selling orders and one or more KeyChain orders.

[0007] In another embodiment, the method further comprises:

(h) calculating and publishing at least one pre-calibrated mid curve for the at least one asset.

[0008] In another embodiment of the method, the at least one buying order and the at least one selling order are entered during the breathing phase and assigned a first random stamp and a second random stamp during the breathing phase.

[0009] In another embodiment of the method, the at least one buying order and the at least one selling order are submitted at substantially the same time at the outset of the matching phase.

[0010] In another embodiment of the method, the at least one asset comprises one from the group of Treasuries and other bonds.

[0011] In another embodiment of the method, the at least one asset spans multiple sectors.

[0012] In another embodiment of the method, the matching phase lasts for at least 3 minutes and ends 30 seconds after a last new order is entered.

[0013] In another embodiment, the method further comprises: (h) a last call phase.

[0014] In another embodiment, the method further comprises: (h) an odd lot cleanup phase.

[0015] In another embodiment of the method, pennying by one or more participants is mitigated. [0016] In another embodiment of the method, the at least one asset is an illiquid asset.

[0017] In another embodiment of the method, the pricing phase includes:

(i) evaluating; and

(ii) accepting or rejecting a level that a participant contributes to calibrate the calibrated mid curve.

[0018] In another embodiment of the method, any orders entered during the matching phase are prioritized based on price/time priority.

[0019] In another embodiment of the method, the matching phase includes performing a credit check on a participant prior to matching any order submitted by the participant.

[0020] In another embodiment of the method, orders entered during matching phase are matched continuously in real-time.

[0021] In another embodiment of the method, each order is filled at a displayed price or filled at an updated price more favorable to the participant that placed the order.

[0022] In another embodiment, the method further comprises:

(h) determining if there is sufficient participant interest for the

auction of the at least one asset.

[0023] In another embodiment of the method, the auction is all-to-all.

Background and Terminology

[0024] During a session, which is a period of time during which a call auction can be held, a participant can enter, or stage, an order from a platform menu. A participant selects (1) a side (buy or sell), (2) an asset, including, for example, a particular issue or instrument, (3) an amount, or size, of an order and (4) a price; that is, a level or yield spread, where the level or yield spread can be below, at or above a mid. An order below (for example, behind), a mid can include, for example, ten million of the asset at the mid in order to promote liquidity.

Brief Description of the Drawings

[0025] FIG. 1 is a flow diagram illustrating an exemplary method of matching orders that traditionally would not be matched or not be matched as efficiently.

[0026] FIG. 2A is a flow diagram illustrating an exemplary Mid- Calibrating Algorithm.

[0027] FIG. 2B is a flow diagram illustrating an exemplary Pricing Phase

[0028] FIG. 3 is a flow diagram illustrating an exemplary KeyChain.

[0029] FIG. 4 is an illustrative screenshot of a Pricing Phase Menu as seen on an exemplary platform.

[0030] FIG. 5 is an illustrative screenshot of a Breathing Phase Menu as seen on an exemplary platform.

[0031] FIG. 6 is an illustrative screenshot of a Matching Phase Menu as seen on an exemplary platform.

[0032] FIG. 7 is an illustrative screenshot of an Order Book as seen on an exemplary platform.

[0033] FIG. 8 is an illustrative screenshot of a KeyChain Menu as seen on an exemplary platform. Detailed Description of Illustrative Embodiments)

[0034] A method and platform for anonymously matching buying and selling orders that are not limited to a single asset type can include an all-to-all, call auction of assets, such as, but not limited to:

(a) U.S. Treasuries, namely, bills, notes and bonds, including, but not limited to:

(i) Off-the-run (OFTR) treasuries;

(ii) On-the-run (OTR) treasuries; and

(iii) Treasury Inflation-Protected Securities (TIPS);

(b) other bonds, including, but not limited to, municipal bonds and corporate bonds;

(c) Non- financial instruments that do or may not have readily

identifiable values or a readily identifiable market, but have some relationship, for example, as to asset type, including but not limited to:

(i) event tickets;

(ii) art;

(iii) vehicles; and

(iv) collectables.

[0035] The method and platform are configured such that participants' trading strategies can be kept anonymous, meaning confidential.

[0036] In some embodiments, assets can be identified by type and/or sector. Sectors can consist of grouping assets by maturity. For example, U.S.

Treasuries can be grouped in, among others, 2-year, 3-year, 5-year, 7-year, 10- year and 30-year sectors. Similarly, TIPS can be grouped in, among others, 5- year, 10-year and 30-year sectors.

[0037] The present method and platform can unlock liquidity in illiquid or low-liquidity assets, or eliminate, or at least reduce, an asset's illiquidity, by identifying one or more market gaps, which arise when a buy order lacks a matching sell order or vice versa. The present method and platform can combine, two or more orders, which can be a combination of swap orders (swaps) and outright orders, to generate one or more live, executable and non- negotiable orders to fill one or more market gaps. An outright order is an order that involves an asset to be traded or exchanged outright on price or yield. An outright order is not a hedge.

[0038] In some embodiments, the orders are combined according to a price/time priority, which is discussed in detail below.

[0039] Generating an order forms a KeyChain of at least one buy order and at least one sell order. The KeyChain is a completable chain or at least

completable portion of a chain. Each KeyChain order that is executed and completed fills at least one market gap and unlocks liquidity, or eliminates, or at least reduces, illiquidity.

[0040] In some preferred embodiments, the orders are anonymous, meaning that participants cannot view any other participant's orders or identity or the total volume of any asset traded, even if the participant is involved in a completed trade for that asset. Anonymity protects both the identity of the participants and the confidentiality of their trading strategies. In addition, during a call auction, a participant cannot see any state of any order book and, at the end of the auction, a participant cannot see the size of any order imbalance. In some embodiments, this trade information can remain confidential even after an auction has ended. [0041] In some embodiments, the method and platform includes determining whether to hold a call auction. In some embodiments, such a determination is made by assessing whether there is sufficient interest in one or more assets to be auctioned. In some embodiments, if there is sufficient interest in at least one asset, then a call auction will be held; if there is not sufficient interest, then a call auction will not be held.

[0042] FIG. 1 shows a schematic diagram of an embodiment of the present method for matching orders and unlocking liquidity in illiquid or low-liquidity assets. Method 100 is an all-to-all call auction of assets. In some embodiments, method 100 is multi-phase and can involve, but is not limited to, one or more of a Pre-Calibrating Phase 110, a Pricing Phase 120, a Breathing Phase 130, a Matching Phase 140, a Last Call Phase 150, an Odd Lot Cleanup Phase 160, and a Post Trade Phase 170.

Pricing Phase

[0043] In the embodiment illustrated in FIG. 1, method 100 starts at Pricing Phase 120. Pricing Phase 120 can be employed to establish a mid to be employed during the call auction. A call auction is an auction in which two or more participants enter one or more orders to buy and/or sell a certain amount of an asset at a certain level, that is, a certain price or spread, followed by a matching of orders.

[0044] In some embodiments, at the end of Pricing Phase 120, a mid is calibrated through Mid-Calibrating Algorithm 210. A mid is a price

representative of an average bid and/or average offer spread.

[0045] In some embodiments, Mid-Calibrating Algorithm 210 involves a Tabulating Step 212, an Averaging Step 214 and a Calibrating Step 216. [0046] In some embodiments, Tabulating Step 212 involves tabulating any buy and/or sell prices contributed by any sponsors; any buy and/or sell prices contributed by any participants; and any buy and/or sell prices contributed by any third parties. Typically, Tabulating Step 212 will involve at least tabulating prices from sponsors and third parties.

[0047] In some embodiments, Averaging Step 214 involves calculating an arithmetic, or Olympic, average by eliminating a percentage of high and low participant prices and then averaging all remaining prices. In some

embodiments, Averaging Step 214 involves taking the mean.

[0048] In some embodiments, participants include institutional investors, traders and dealers. In some of these same embodiments, sponsors can also be participants. In some embodiments, sponsor membership enables end-user access to the platform.

[0049] In some embodiments, Calibrating Step 116 involves calibrating the averages arrived at via Averaging Step 114 based on previous call auction data.

[0050] Mid-Calibrating Algorithm 110 is designed to produce a robust mid, even in instances where no participant contributes a price during Pricing Phase 120. That is, a participant can, but is not obligated to, contribute one or more prices during Pricing Phase 120. A robust mid is more likely to result in a mutually acceptable price to a buyer and seller during Matching Phase 140.

[0051] In some embodiments, the mid calibrated by Mid-Calibrating

Algorithm 110 is not displayed to any participants. In some embodiments, the mid can be shared with at least one sponsor in order to address information asymmetry that currently exists or may exist on certain third-party platforms. [0052] In some preferred embodiments, Pricing Phase 120 is confidential such that participants are not shown any pre-calibrated mid curves or any levels contributed by other participants.

[0053] In some embodiments, during Pricing Phase 120, prices are updated after a given time period. In some preferred embodiments, the time period is an hour. In some more preferred embodiments, the time period is a minute. In some preferred embodiments, the time period is a second. In some

embodiments, the time period is designed to enhance actual or potential liquidity of market assets.

[0054] In some embodiments, participants submit a price during Pricing Phase 120 via electronic feed or directly into a call auction. In most, if not all, embodiments, a participant's submission of one or more prices has no bearing on any order or potential to be matched during Matching Phase 140.

[0055] In some embodiments, a pre-calibrated mid is calculated prior to Pricing Phase 120 in Pre-Calibrating Phase 110. The pre-calibrated mid can enable more efficient discarding of out-of-band prices that participants may submit during Pricing Phase 120. For example, in some embodiments, if the difference between a participant's submitted price and the pre-calibrated mid exceeds a configurable or predetermined tolerance, then the participant's price is discarded when the mid is calibrated in Pricing Phase 120. In some embodiments, the participant is not notified that a submitted price has been discarded. In other embodiments, the participant is notified that a submitted price has been discarded.

[0056] In some embodiments, the algorithm employed to determine the pre- calibrated mid is Mid-Calibrating Algorithm 110. In other embodiments, the pre-calibrated mid is determined in a different way, including, but not limited to, algorithms other than Mid-Calibrating Algorithm 110. In some embodiments, if no participant contributes a level during Pricing Phase 120, then the pre- calibrated mid and calibrated mids are identical. In some embodiments, participants are able to view calibrated mids.

[0057] In some embodiments, resulting mids are shared with participants, while maintaining the confidentiality of the sponsor and participant

contributions.

[0058] Pricing Phase 120 can include calculation, calibration and publication of one or more robust mid-market level, or mid, curves for assets in the call auction.

Breathing Phase

[0059] In some embodiments, Breathing Phase 130 follows after Pricing Phase 120. Breathing Phase 130 can include, among other things, an

opportunity for a participant to pause and evaluate one or more published mids.

[0060] In some embodiments, Breathing Phase 130 can involve two or more participants entering one or more orders prior to Matching Phase 140. During Breathing Phase 130, a participant can enter one or more swaps, one or more outright buy or sell orders or a combination of swaps and outright buy or sell orders. A swap is an order in which an asset is proposed to be traded, switched or exchanged for another asset.

[0061] In some embodiments, a participant can originate (1) an outright, OFTR order on price, in 32nds, or yield and not tied to a benchmark, (2) a swap involving an OFTR, which can be entered in a yield spread tied to (a) the sector benchmark or OTR (for example, buy the OFTR and sell a duration-weighted amount of the OTR), (b) an alternative sector benchmark or OTR or (c) another OFTR (for example, buy one OFTR and sell a duration-weighted amount of another OFTR). [0062] A duration- weighted amount is the amount of an asset necessary to balance a hedged transaction. A yield spread is the difference between the quoted rates of return on two different investments, usually of different credit quality. A yield spread is often indication of a risk premium for investing in one asset over another. A yield spread is also known as a credit spread.

[0063] In some embodiments, orders entered during Breathing Phase 130 can be restricted by a minimum fill constraint. In some embodiments, an anchor leg of a swap can have a minimum fill constraint. In other or the same

embodiments, no leg of a swap can have a minimum increment constraint.

[0064] In some swaps, an asset for which the participant has selected an amount is the anchor. The participant selects a second asset, or leg, in the swap by entering one or more criteria for the second asset, such as a particular issue. In some embodiments, the participant is presented with assets having the one or more criteria. Often this information is presented on a screen. The participant can then select the second asset, determine its direction, that is, buy or sell, and the platform can capture a corresponding mid. In some embodiments, the participant then selects and confirms the level or spread, that is, the difference in value for the second, or hedged, asset, and observes a calculated duration- weighted amount for the second asset. The participant then submits and confirms its order.

[0065] In some embodiments, method 100 allows various buyers and sellers to originate orders during Breathing Phase 130, Matching Phase 140 or both Phases. A buyer or seller may originate the same types of orders described above during either or both of those Phases.

[0066] In some embodiments, two-legged swaps can be executed, as can butterfly or three-legged swaps. [0067] In some embodiments, orders entered during Breathing Phase 130 receive unique time-stamps prior to Matching Phase 140. In some embodiments, the time-stamps are randomized to remove any advantage of being "first in" with an order during Breathing Phase 130. In some embodiments, once an order is entered, a participant's order book is updated with various order details.

Matching Phase

[0068] Matching Phase 140 involves matching orders. At the outset of Matching Phase 140, all orders previously entered, such as those entered during Breathing Phase 130, are submitted. In some embodiments, most if not all previously entered orders are submitted substantially simultaneously.

[0069] During Matching Phase 140, the present method anonymously identifies one or more market gaps, which arise where a buying order lacks a matching selling order or vice versa, and combines, according to price/time priority, two or more orders, which are a combination of swaps and outright orders, to generate one or more live, executable and non-negotiable orders to fill at least one of the market gap(s).

[0070] In at least some embodiments, generating the one or more live, executable and non-negotiable orders forms a KeyChain of at least one buying order and at least one selling order. The KeyChain is a completable chain, or completable portion of a chain, of anonymous orders. In some embodiments, the orders are prioritized according to price and then time as discussed in detail below.

[0071] Orders handled via a KeyChain include orders for one or more different asset types, such as, but not limited to, OFTRs, Treasury Inflation- Protected Securities and OTRs. [0072] In some embodiments, a KeyChain order does not have a minimum fill or minimum increment constraint. Typically, a round lot is the minimum amount of an asset that is allowed to be traded on a given platform. However, in an effort to complete KeyChains, amounts other than a round lot can be part of an executable order.

[0073] In some embodiments, a KeyChain order for an odd lot (for example, an amount of an asset that is other than a round lot to the nearest 1,000 units) is calculated to the nearest 1,000 units and a partial order for less, but not more, than the published bid or offer amount is allowed. In other words, when a participant hits (sells) or lifts (buys) a KeyChain bid or offer, the participant will be notified, for example, via a pop-up message, to specify the amount of an order, but not a price. A participant can specify an amount up to the full amount published, and the order can be filled up to the amount entered.

[0074] An odd lot of an OTR is adjusted to produce a duration-weighted round lot of an OFTR in order to further enhance liquidity of the OFTR. A KeyChain order that is a spread order is duration-weighted to the corresponding benchmark.

[0075] Each KeyChain order is non-negotiable as to price. A KeyChain order is not a resting order entered by a participant. Rather, for a KeyChain order, a participant can either sell a bid or buy an offer, but cannot negotiate or counter a price of the KeyChain order. In some embodiments, the KeyChain is not and does not provide an order book. A KeyChain order does not have a minimum fill or minimum increment constraint.

[0076] Each KeyChain order that is executed fills at least one market gap and unlocks liquidity, or eliminates, or at least reduces, illiquidity. [0077] In some embodiments, during Matching Phase 140, at least one published price is updated periodically. In some embodiments, the period is one second. As such, an updated price can be different from a price at which a participant sells a bid or buys an offer. In some embodiments, a participant's order is filled at the published price or at a better price. If an updated price is worse than the price a participant hit or lifted, then the participant will be notified by, for example, a pop-up message informing the participant that the price has adversely changed and asking if the participant still wants to execute the trade at the updated price. The execution of the trade depends on the participant's answer. If an updated price has changed in favor of a participant, then the participant's order is simply filled at the more favorable price.

[0078] During Matching Phase 140, a participant can originate any of the same orders, including swaps and outright orders, that a participant can originate during Breathing Phase 130. In some embodiments, a participant can specify any of the same constraints available during Breathing Phase 130.

[0079] Orders entered during Breathing Phase 130 have temporal priority over orders entered during Matching Phase 140. Orders entered during

Matching Phase 140 are matched according to price/time prioritization against other orders entered during Matching Phase 140.

[0080] In some embodiments, orders are matched such that both sides of a matched order are substantially simultaneously filled according to price/time priority. In some embodiments, orders are matched in real time and/or continuously.

[0081] In some embodiments, price takes priority over time. In some of these embodiments, if multiple orders have the same price, time takes priority between or among them. In some embodiments, the amount, or size, of an order has no priority. [0082] In other or the same embodiments, orders are credit- verified. In some embodiments, what accounts for a credit-verified order is determined by a counter party credit or credit limit.

[0083] A participant can increase its odds of being matched if, during

Matching Phase 140, it improves its level by, for example, buying at a higher price, lower yield or more aggressive spread or selling at a lower price, higher yield or more aggressive spread.

[0084] During Matching Phase 140, a participant can choose settings that enable it to view indications of participant interest in an asset, as well as trade activity in an asset. In some embodiments, no participant can view direction, that is, buy or sell, price or amount of an issue in which it has interest or that it has traded.

[0085] During Matching Phase 140, a participant can view the progress of a filled order in real time. During Matching Phase 140, a participant can cancel any unfilled orders individually or can cancel all unfilled orders.

[0086] In some embodiments, during Matching Phase 140, an attempt to fill a constrained order (for example, an order with a minimum fill requirement, an order with a minimum increment requirement, a swap or some combination of those constraints) will be made when price and time give the constrained order priority. However, if one or more constraints cannot be satisfied, then an attempt to fill the next prioritized order will be made.

[0087] In some embodiments, Matching Phase 140 includes one or more matching time slices. A matching time slice is an interval of time between a time when a matching engine wakes up, that is, commences matching activity, and a time when the matching engine goes to sleep, or ceases matching activity. Orders filled during a matching time slice are substantially simultaneously filled according to price/time priority. In some embodiments, the present method and platform include a look-ahead capability to ensure that all final matches in a matching time slice honor price/time priority.

Last Call Phase

[0088] In some embodiments, Last Call Phase 150 follows after Matching Phase 140. Last Call Phase 150 is held after Matching Phase 140 is completed. In some embodiments, Last Call Phase 150 is configurable based on prior auction activity. For example, in some embodiments, Last Call Phase 150 can consist of at least a 30-second batch session as opposed to a continuous session. In some embodiments, orders are matched in the same way as described above for Matching Phase 140. In some embodiments, Last Call Phase 150 involves only assets for which a bad mid, or bad mids, was/were calculated.

[0089] In some embodiments, a bad mid exists where there are only offers for an asset through the mid (above the mid) and no offers below or at the mid, or only bids for an asset through the mid (below the mid) and no bids above or at the mid. On the other hand, if there has been a trade of an asset at the mid, then the mid is a good mid. The fact that an asset does not receive any orders means that there was no participant interest in the asset, but does not mean that the mid is a bad mid. In some embodiments, if the Last Call Phase 150 is evoked because of a bad mid, there is no requirement that an amount (for example, ten million) of an asset be submitted at a corresponding mid.

Odd Lot Cleanup Phase

[0090] In some embodiments, Odd Lot Cleanup Phase 160 follows after Matching Phase 140. In embodiments with Last Call Phase 150, Odd Lot Cleanup Phase 160 follows after Last Call Phase 150. In some embodiments, mids for Odd Lot Cleanup Phase 160 are the same as the mids published for one or both of Breathing Phase 130 and Matching Phase 140, except that mids for assets which traded during Last Call Phase 150 are the weighted average levels of those trades.

[0091] In some embodiments, Odd Lot Cleanup Phase 160 does not include a requirement that an amount (for example, ten million) at a corresponding mid. In some embodiments, Odd Lot Cleanup Phase 160 does not allow one or more of a minimum fill constraint, a minimum increment constraint, and an order through the mid. In some embodiments, Odd Lot Cleanup Phase 160 does allow an order through a mid.

[0092] To further enhance efficiency, a participant can customize a view of assets in order to focus on one or more assets in one or more sectors in one or more upcoming call auctions or phases.

[0093] Also, in at least some embodiments, every action taken by every participant is logged. Corresponding data is stored in order to address, for example, actual or potential regulatory concerns or trade disputes.

The Platform

[0094] The method described above can be carried out on a platform. This platform can include a group of computers, servers, or both, networked together either physically and/or via the internet.

[0095] In some embodiments, participants will be able to access the platform via direct access or sponsored access. Sponsored access means that the participant is interacting with the platform via a third party. From a user perspective, in at least some embodiments, a participant using sponsored access will interact with the platform similarly to, or exactly like, a participant directly interacting with the platform via direct access.

[0096] In some embodiments, sponsors include brokers. In some

embodiments, sponsors handle post-trade allocations. [0097] In some embodiments, the platform can allow for simultaneous auctions of multiple sectors. In other embodiments, the platform can have the ability to restrict auctions to various sectors. In some embodiments, participants can view, on an Auction Calendar, assets, sectors or a combination of assets and sectors to be auctioned. In some embodiments, the Auction Calendar can include auctions that repeat on regular intervals (for example, hourly or daily).

[0098] In certain embodiments, the Auction Calendar includes the current date and time, as well as dates and times of upcoming auctions. In certain embodiments, the Auction Calendar includes a countdown clock to an

upcoming auction. In certain embodiments, participants are notified on the Auction Calendar when an auction is cancelled. In certain embodiments, the Auction Calendar is interactive, allowing participants to gain more knowledge about upcoming auctions, including, but not limited to, particular assets, sectors, or both, that are to be involved in an auction. In certain embodiments, the Auction Calendar has the ability to notify participants of upcoming auctions that include assets, such as bonds, that the participant owns, in which it has expressed interest, or both. In certain embodiments, a participant can request that the platform add a particular asset or sector to an upcoming auction. In certain embodiments, the platform automatically decides whether to add the participant-requested auction based on various factors, including, but not limited to, the number of other participants requesting an auction for the same asset, sector, or both, the time since the asset, sector, or both, were last auctioned, or some combination thereof.

[0099] In at least some embodiments, the platform can determine how long auctions will last. In some embodiments, the auction can last 5, 8, 10, 15, 30 or 60 minutes. In certain embodiments, an auction can continue past its time limit if trading is still taking place or one or more orders are still being entered over a certain period of time. For example, in some embodiments the platform can keep the trading open for a given amount of time after a trade has occurred. In certain embodiments, the platform can cancel an auction before its time limit has expired.

[0100] Pricing Phase Menu 400 is illustrated in FIG. 4. In certain

embodiments, the Pricing Phase 120 lasts three minutes, although other durations are possible. In some embodiments, participants can enter levels expressed as yield spreads to the sector benchmark or OTR. In certain embodiments, participants can enter bids, offers or two-way levels. In certain embodiments, such as the one illustrated in FIG. 4, levels are entered in basis points with 1 decimal place of precision (tenths of a basis point).

[0101] In at least some embodiments, the levels for the benchmarks are snapped at the beginning of Pricing Phase 120 and remain constant throughout Pricing Phase 120. The benchmark levels can be referred to as the "lock" price and yield. In some embodiments, the lock price of the benchmarks is displayed in 32nd notation to the nearest 32nd. In some embodiments, the lock yield of the benchmarks is displayed as a percent with 3 significant digits.

[0102] As set forth above, in preferred embodiments, a participant has no ability to view levels entered by other participants during Pricing Phase 120. In some embodiments, the platform can notify the participant if they attempt to enter a level significantly out of range (out of band) relative to pre-calibrated mids. Pre-calibrated mid are mids calibrated prior to the start of Pricing Phase 120. In some embodiments, the participants can override the notification. In some embodiments, the participant will be informed only once if its price is out of range to try and prevent, or at least dissuade, participants from using this information to attempt to determine the pre-calibrated mid.

[0103] In at least some embodiments, participants can attempt to game the platform by entering prices skewed towards their preferred mid. In at least some embodiments, the platform can detect this gaming and take actions to prevent, or at least mitigate, further gaming.

[0104] In some embodiments, the platform gives participants the ability to import data from various spreadsheet formats, including, but not limited to, Excel and comma-separated values (CSV), to enter pricing data. In some embodiments, participants can add, change or remove pricing levels until Pricing Phase 120 ends.

[0105] As set forth above, the platform has the ability to calculate a mid. In at least some embodiments, the platform employs four different components when calculating the mid: sponsor contributions, participant contributions, calculation of an arithmetic average, and a calibration of that average based on previous auction results. An example of this is illustrated in FIG. 2B. In at least some embodiments, the platform is designed to produce a usable mid even if no participants contribute prices during Pricing Phase 120. In some embodiments, the platform can use data from sponsors, third parties or both in calculating the mid. In certain embodiments, it is preferable to have data from at least three sponsors to calculate a mid for a given asset. In some embodiments, sponsors can provide bids, offers, pricing data or some combination thereof. In certain embodiments, bids and offers from sponsors can be treated in the same way as pricing data.

[0106] In some embodiments, prior to Pricing Phase 120, the platform performs the mid-calibrating algorithm to compute, or calculate, a pre- calibrated curve of mids. In most embodiments, the pre-calibrated mids are not displayed to the participants. In some embodiments, the pre-calibrated mids can be shared with sponsors. Sharing the pre-calibrated mids with sponsors can help address information asymmetry. The pre-calibrated mids allow the platform to exclude potentially erroneous, that is, out-of-range data that participants may enter in an attempt to game the platform.

[0107] In at least some embodiments, the pre-calibrated curve of mids is generated by the identical algorithm that is employed after Pricing Phase 120 has ended. In at least these embodiments, if no participants contribute levels during Pricing Phase 120, the pre-calibrated and post-Pricing-Phase curves are identical. If participants do contribute levels during Pricing Phase 120, their prices are accounted for when calibrating the mids. In at least some

embodiments, the platform attempts to remove data from those trying to game the platform by removing out-of-range prices. In some embodiments, the platform discards the top and bottom X% of participant prices, where X is configurable by the platform. In some embodiments, X is 20%. In some embodiments, where there are few participant prices, the platform can use all of the participants' prices.

[0108] In at least some embodiments, after calculating the arithmetic mid which, in at least some embodiments, is based on the average of the average participant prices (after removing, or discarding, the out-of-range prices), the average sponsor prices and the average third party prices, the platform calibrates each mid. The mid can be calibrated based on data including, but not limited to, the published mid from one or more previous auctions; the volume and the volume- weighted mid of the order imbalance at the end of the previous auction; the volume and the volume-weighted mid of the matches from the Matching Phase 140 of the previous auction or various combinations.

[0109] In at least some embodiments, there is no commitment to execute any bids or offers entered during Pricing Phase 120. In certain embodiments, Mid- Calibrating Algorithm 210 is employed to obtain a mid. In at least some embodiments, the platform publishes the mid for the participants at the end of Pricing Phase 120, at the beginning of Breathing Phase 130 or at both times.

[0110] Breathing Phase Menu 500 is illustrated in FIG. 5. After the mids have been calculated and calibrated, the mids are published and viewable by participants and Breathing Phase 130 begins. In certain embodiments, Breathing Phase 130 lasts two minutes, although other durations are possible. In at least some embodiments, Breathing Phase Menu 500 contains a countdown clock that shows when Breathing Phase 130 ends. In certain embodiments, a lock price, a yield tied to a benchmark or both are snapped at the beginning of Breathing Phase 130 and remain constant until the end of Breathing Phase 130.

[0111] In at least some embodiments, Breathing Phase Menu 500 contains the amount of credit a participant has. In at least some embodiments, submitting an order during Breathing Phase 130 does not affect the participant's credit.

[0112] During Breathing Phase 130, participants can enter orders. In at least some embodiments, orders are not matched during Breathing Phase 130. In some of these embodiments, the orders are assigned a random time stamp to remove advantages for an order that is "first in", including an order that is entered ahead of another order because of a faster typist, a faster internet speed or other similar advantage.

[0113] In certain embodiments, Breathing Phase Menu 500 can highlight an asset, such as a bond, for which a participant entered a price level during Pricing Phase 120. In some embodiments, Breathing Phase Menu 500 can indicate how a participant's price level compares to the calibrated mid.

[0114] In some embodiments, participants can enter orders via Breathing Phase Menu 500. In some embodiments, a participant can indicate a minimum fill and minimum increment he would accept for buying or selling a given asset. In some embodiments, if the minimum fill, minimum increment or both are left blank, the platform automatically assigns to each criterion a predetermined amount. In certain embodiments, the platform allows participants to import data from various spreadsheet formats, including, but not limited to, Excel and CSV, to enter ordering data.

[0115] In certain embodiments, participants can enter swaps on Breathing Phase Menu 500. In certain embodiments, the orders entered during Breathing Phase 130 are simultaneously submitted at the start of Matching Phase 140.

[0116] In some embodiments, the platform's matching engine enforces price/time priority. In these embodiments, a participant might be motivated to improve its level, by a trivial amount, in order to increase its odds of being matched. Such trivial increases or decreases in price are known as pennying.

[0117] Some embodiments of the platform include mining data from past and current call auctions, and taking action to prevent a participant from further attempts to game the platform through pennying. In some embodiments, different rules against pennying are applied to different assets.

[0118] FIG. 6 shows Matching Phase Menu 600. In certain embodiments, Matching Phase 140 is a continuous matching session during which participants receive real-time notification of their filled orders, but no indication of what other participants have traded or the state of the order book. In other

embodiments, Matching Phase 140 is a batch session.

[0119] In some embodiments, after the orders entered during Breathing Phase 130 are entered into an order book, the platform's matching engine attempts to match them. In certain embodiments, Matching Phase 140 lasts for three minutes or until trading activity has abated. In some embodiments, the abatement of trading is determined based on a given time period without any trades or entries or orders. In some embodiments, the platform employs a countdown to indicate the end of Matching Phase 140, regardless of the number of trades made. In some embodiments, the clock is set to 30 seconds. In certain embodiments, Matching Phase Menu 600 can container an Order Book 700, such as the one shown in FIG. 7, which can show, among other things, the orders a participant has entered.

[0120] In certain embodiments, if a participant changes its order parameters, it loses its time priority in the order stack. In some embodiments, participants are given real-time notification of filled orders on Matching Phase Menu 600.

[0121] The platform's matching engine can identify one or more market gaps and generate one or more KeyChain orders to fill the one or more market gaps during Matching Phase 140.

[0122] In at least some embodiments, the matching engine performs credit checks prior to matching two orders. In certain embodiments, sponsors assign a credit limit to each of their sponsored client firms and client firms assign a credit limit to each of their traders. In some embodiments, the sum of the traders' limits can exceed the credit limit the sponsor has assigned that client, but the smaller limit is enforced. In some embodiments, credit limits are defined across the entire platform. In some embodiments, the platform evaluates credit by looking at a balance sheet (un-netted) and/or risk (netted). In at least some embodiments, credit is not checked at the time of order entry, but rather, at the time of matching. In some embodiments, the ordered is filled up to the credit limit, the participant is sent a notice as to the filled order, or portion thereof, and a remaining order, or portion thereof, is cancelled. In some embodiments, firms can prioritize a client's credit limits.

[0123] In some embodiments, the platform will not interact directly with a sponsored client. [0124] In some embodiments, the platform allows one or more of the same types of orders as the above-described method.

[0125] In at least some embodiments of the platform, price always takes priority. If there are multiple orders at the same price, time takes priority among them. In at least some embodiments, size has no priority because a size of an order can be relatively easily gamed and, in the process, undermine the platform's efficiency. Therefore, time trumps size, and price trumps time. The platform attempts to fill a constrained order (for example, an order with a minimum fill requirement, a minimum increment requirement, a swap, or some combination of those constraints) when price and time give them priority.

However, if one or more constraints cannot be satisfied, then the platform attempts to fill the next prioritized order.

[0126] As mentioned above, Matching Phase 140 is a continuous match session during which the algorithm attempts to match as many orders as possible. The matching engine goes to sleep when it has matched everything it can. The matching engine does not wake up until a new order enters the order book or new levels for the benchmarks are snapped.

[0127] At the start of Matching Phase 140, there can be a mini batch auction that takes place before the continuous matching begins, but the same matching rules remain in force. In some embodiments, Matching Phase 140 includes one or more matching time slices. The orders that match engine fills during a matching time slice are filled substantially simultaneously according to price/time priority. Each successive matching time slice, however, has its own time stamp. Within a matching time slice, the matching engine can look ahead in ways that violate time priority so long as the final results from that matching time slice honor price/time priority. [0128] It is possible that, during a matching time slice, a participant gets matched multiple times. In some embodiments, the matching algorithm can lead to size-based starvation (for example, a large order with time priority can starve out other orders at the same price). This potential issue can be overcome if a participant improves its price.

[0129] In certain embodiments, when Matching Phase 140 has ended, a participant's partial filled orders for the same asset, such as a bond, are aggregated into a single average-price trade. In some embodiments, the platform sends out a trade ticket at the end of Matching Phase 140.

[0130] Some platforms utilize a Last Call Phase Menu. The platform may produce a bad mid for one or more assets. In certain embodiments, Last Call Phase Menu will appear when a bad mid was produced. In some embodiments, Last Call Phase 150 lasts thirty seconds, although other durations are possible. In at least some embodiments, Last Call Phase 150 is a batch auction and follows after the same procedure as Matching Phase 140, with the exception that the platform's matching engine does not run until the end of Last Call Phase 150. At the end of Last Call Phase 150, all of a participant's filled orders for the same asset are aggregated into a single trade at an average price.

[0131] Some platforms utilize an Odd Lot Cleanup Phase Menu. In certain embodiments, Odd Lot Cleanup Phase 160 follows after Last Call Phase 150 or (if there is not a Last Call Phase 150) then Matching Phase 140. In at least some embodiments, this is no minimum size constraint during Odd Lot Cleanup Phase 160. In at least some embodiments, Odd Lot Cleanup Phase 160 employs the same matching algorithm as Matching Phase. In some embodiments, at the end of Odd Lot Cleanup Phase 160, all of a participant's partially filled orders for a single asset are aggregated into a single trade at an average price. In certain embodiments, if a participant traded the same asset in both Matching Phase 140 and Odd Lot Cleanup Phase 160, the filled orders can be aggregated to produce a single trade at an average price. In some embodiments, entering orders off the mid is not allowed in Odd Lot Cleanup Phase 160. As with the other phases, certain embodiments of the platform may not employ Odd Lot Cleanup Phase 160.

[0132] Some platforms utilize a Post Trade Phase Menu. In at least some embodiments, Post Trade Phase 170 follows after one or more of Odd Lot Cleanup Phase 160, Last Call Phase 150 and Matching Phase 140. In certain embodiments, Post Trade Phase Menu shows, among other things, individual fills, aggregate, average-price trades, fee(s), accrued interest, settlement amount, session duration, trades not filled and/or duration-weighted amount of the hedge. In certain embodiments, Post Trade Phase Menu also shows a

countdown clock for a next scheduled auction. In some embodiments, the countdown clock is specific to a particular sector. In some embodiments, Post Trade Phase Menu allows participants to download auction results into a CSV file.

[0133] In certain embodiments, the platform may allow OTR-only auctions. OTR-only auctions are separate auction from auctions that include OFTR's, TIPS, or both OFTRS and TIPS. The only bonds in an OTR-only auctions are one of more of the six OTR's. There are no outright or hedged orders, only swaps. In some embodiments, there is no Pricing Phase for an OTR-only auction. Each published mid is the arithmetic mid of the sponsor feeds. The published mid is the yield spread between the mid yields of two bonds. The mids can be updated every second. The minimum order size is typically 10 million. Because all orders are swaps, every participant must accept odd lots. A minimum fill constraint is allowed. A minimum increment constraint is not allowed. There can be a KeyChain associated with the OTR Auction. [0134] The platform can log each action taken by each participant.

[0135] The above described embodiments can be employed individually or in various suitable combinations. Listed steps can be excluded or additional steps added without departing from the scope of the present disclosure.

[0136] While particular elements, embodiments and applications of the present invention have been shown and described, it will be understood that the invention is not limited thereto since modifications can be made without departing from the scope of the present disclosure, particularly in light of the foregoing teachings.

Claims

CLAIMS What is claimed is:

1. A computer- implemented method for an anonymous multi-phase, call auction of at least one asset, the method comprising:

(a) initiating the auction for the at least one asset;

(b) conducting a pricing phase for a first period of time;

(e) conducting a matching phase over a third period of time;

(f) receiving at least one buying order and at least one selling order; and

2. The method of claim 1, further comprising filling one or more of the at least one buying order and one or more of the at least one selling order.

3. The method of claim 1, further comprising:

(h) identifying one or more market gaps;

(i) generating one or more KeyChain orders to fill one or more of the one or more market gaps; and

j) completing a KeyChain of one or more buying orders, one or more selling orders and one or more KeyChain orders.

4. The method of claim 1, further comprising:

(h) calculating and publishing at least one pre- calibrated mid curve for the at least one asset.

5. The method of claim 1, wherein the at least one buying order and the at least one selling order are entered during the breathing phase and assigned a first random stamp and a second random stamp during the breathing phase.

6. The method of claim 4, wherein the at least one buying order and the at least one selling order are submitted at substantially the same time at the outset of the matching phase.

7. The method of claim 5, wherein the at least one asset comprises one from the group of Treasuries and other bonds.

8. The method of claim 7, wherein the at least one asset spans multiple sectors.

9. The method of claim 5, wherein the matching phase lasts for at least 3 minutes and ends 30 seconds after a last new order is entered.

10. The method of claim 5, further comprising:

(h) a last call phase.

11. The method of claim 5, further comprising:

(h) an odd lot cleanup phase.

12. The method of claim 5, wherein pennying by one or more participants is mitigated.

13. The method of claim 1, wherein the at least one asset is an illiquid asset.

14. The method of claim 1, wherein the pricing phase includes:

(i) evaluating; and

15. The method of claim 1, wherein any orders entered during the matching phase are prioritized based on price/time priority.

16. The method of claim 1, wherein the matching phase involves performing a credit check on a participant prior to matching any order submitted by the participant.

17. The method of claim 1, wherein orders entered during matching phase are matched continuously in real-time.

18. The method of claim 1, wherein each order is filled at a displayed price or filled at an updated price more favorable to the participant that placed the order.

19. The method of claim 1, further comprising:

(h) determining if there is sufficient participant interest for the auction of the at least one asset.

20. The method of claim 1, wherein the auction is all-to-all.