UNDESIRABLE AND FRAUDULENT BEHAVIOUR IN ONLINE
AUCTIONS
Jarrod Trevathan
School of Mathematical and Physical Sciences
James Cook University
Wayne Read
School of Mathematical and Physical Sciences
James Cook University
Keywords:
Bid shielding, shilling, sniping, siphoning, non-existent/misrepresented items.
Abstract:
Online auctions are a popular means for exchanging items over the Internet. However, are many inherent
security and fairness concerns. Participants can behave in an undesirable and fraudulent manner in an attempt
to gain an advantage at the expense of rivals. For example, a bidder might seek to suppress the price by bid
sniping, or the seller could introduce fake bids to inflate the price. In addition, an outsider or rival seller can
lure away bidders by directly offering them better deals, or a malicious seller can auction mis-represented or
non-existent items. This conduct is a problem as it results in market failure, thereby inhibiting the usefulness of
online auctions as an exchange medium. While cryptography has been used to provide security in terms of bid
authentication and privacy, there is no documented means to prevent many of the aforementioned problems.
This paper investigates undesirable and fraudulent behaviour in online auctions. We examine the following
practices: bid shielding, shill bidding, bid sniping, siphoning and selling non-existent or misrepresented items.
We describe the characteristics of such behaviour and how to identify it in an auction. We also provide
recommendations for recourse against undesirable and fraudulent participants.
1 INTRODUCTION
Online auctions are one of the most popular destina-
tions on the web. Buyers and sellers can exchange
items amongst a worldwide audience from the com-
fort and privacy of their own homes. Participants re-
main largely anonymous and can bid in any manner
they desire. However, this freedom comes at the ex-
pense of new security risks and fairness concerns.
By behaving in an undesirable or fraudulent man-
ner, one party is able to gain at the expense of an-
other. For example, bidders can use practices such as
bid shielding and bid sniping to keep the price low.
Alternately, shill bidding is a strategy which a seller
may pursue, to artificially inflate the auction price. In
addition, siphoning is a tactic employed by an out-
sider, who is seeking to profit from an auction by of-
fering bidders a cheaper, identical item. Finally, a
seller might attempt to auction a non-existent or mis-
represented item.
Auction security has been previously discussed in
(Franklin and Reiter, 1996; Stubblebine and Syver-
son, 1999; Trevathan et al, 2005). Cryptographic
methods have been proposed to solve many of these
security issues (see e.g., (Franklin and Reiter, 1996;
Viswanathan et al, 2000; Trevathan et al, 2006)).
However, cryptographic solutions are generally lim-
ited to bid authentication and privacy. Protecting auc-
tion participants from the aforementioned problems
is a much harder task (i.e., a bidder’s bidding strategy
or misrepresented goods). Furthermore, none of the
proposed models in literature specifically address the
auction format used in online auctions.
This paper discusses undesirable and fraudulent
practices in online auctions. We describe the char-
acteristics of such behaviour, and show how auction
participants can identify if they are a victim. We
also provide recommendations/options for what to do
when such behaviour is encountered. We show how
behaviours are related, and when an auction exhibits
one form of undesirable behaviour, the other forms
soon manifest to counter balance it. This inevitably
leads to market failure, and retards the usefulness of
online auctions as an exchange medium.
This paper is organised as follows: Section 2 de-
scribes the basic format and rules of a typical on-
line auction. Sections 3 through 7 discuss each ma-
jor form of undesirable auctioning behaviour. Each
of these sections includes a description of individ-
ual behavioural characteristics, remedies/recourse for
450
Trevathan J. and Read W. (2006).
UNDESIRABLE AND FRAUDULENT BEHAVIOUR IN ONLINE AUCTIONS.
In Proceedings of the International Conference on Security and Cryptography, pages 450-458
DOI: 10.5220/0002100704500458
Copyright
c
SciTePress
Figure 1: An Example Online Auction.
Figure 2: Bid Shielding.
victims and relationships between certain types of
behaviour. Section 8 provides some concluding re-
marks.
2 ONLINE AUCTIONS
There are many types of auction (e.g., Vickrey, CDA,
etc.). The most popular type of auction is the English
auction. In an English auction, bidders outbid each
other in an attempt to win an item. The winner is
the bidder with the highest bid. English auctions are
commonly employed in online auctions such as those
offered by eBay
1
and ubid
2
Online auctions differ to traditional auctions in that
the auction ends after a given period of time. Fig-
ure 1 illustrates a typical online auction (such as that
offered by eBay). Time flows from left to right. Bid-
ders can only submit bids between the auction start
and end times. Each bid must be for an amount that is
greater than the current highest bid. When the auction
terminates, the winner is the bidder with the highest
bid (in this case $75).
eBay and ubid offer a mechanism for automatically
bidding on a bidder’s behalf. This is referred to as the
proxy bidding system on eBay, and the bid butler on
ubid. A bidder is only required to enter the maximum
price they are willing to pay. The bidding software
will then automatically outbid any other bid until the
maximum, potentially saving the bidder money. Such
1
http://www.ebay.com
2
http://www.ubid.com
mechanisms remove the need for a bidder to be con-
stantly watching an auction for bidding activity.
3 BID SHIELDING
Bid shielding typically involves several bidders work-
ing in collusion with each other, or a bidder with ac-
cess to multiple accounts. The first bidder enters a
bid for the amount they are willing to pay for an item.
A second bidder then immediately enters an exces-
sively high bid in an attempt to deter other bidders
from continuing to bid. As the auction is drawing to
a close, the second bid is retracted. When this occurs,
the next highest bid remaining (i.e., the first bid) then
becomes the winning bid.
Figure 2 illustrates a bid shielding scenario. Ini-
tially regular bidders make a few bids ($10 and $15 in
this case). The first colluding bidder enters his/her bid
for $20. This is referred to as the shielded bid. The
second colluding bidder then enters a bid for $250,
which is well above the current bidding range, and is
unlikely to be outbid. This is known as the shielding
bid.
In the example, the shielding bid deters regular bid-
ders from bidding again. Prior to the end of the auc-
tion, the second colluding bidder retracts the shielding
bid for $250. The winning bid then falls back to the
shielded bid for $20. As the bid is retracted near the
end of the auction, other bidders do not have time to
respond. By this stage, most regular bidders have lost
interest in the auction anyway.
Bid shielding generally cannot be accomplished as
shown in the example due to the presence of software
UNDESIRABLE AND FRAUDULENT BEHAVIOUR IN ONLINE AUCTIONS
451
Figure 3: Bid Shielding in Auctions with Automated Bidding Agents.
bidding agents (i.e., eBay’s proxy bidding system).
As soon as the shielding bid is entered, the bidding
history won’t show this as a bid for $250. Instead, this
will be listed as the minimal amount required to out-
bid the shielded bid (e.g., $21, as $1 is the minimum
increment). Rival bidders will then be inclined to con-
tinue bidding, which in turn drives up the shielding
bid’s value.
Bid shielding in the presence of automated bidding
agents can be achieved by using two shielding bids.
Figure 3 illustrates this scenario. As in the previ-
ous example, a shielded bid for $20 is submitted. A
shielding bid for $250 is entered followed immedi-
ately by another shielding bid for $260. As the sec-
ond shielding bid outbids the first, the bidding his-
tory immediately reflects the value of the first shield-
ing bid (i.e., $250), thus giving the appearance of a
high price. Near the end, both of the shielding bids
are retracted. The winning bid then drops to $20 (i.e.,
the shielded bid).
This sort of behaviour disadvantages the seller in
terms of suppressing the price. Honest bidders are
also disadvantaged as they are forced out of the auc-
tion.
There are no mechanisms in place to prevent bid
shielding. Most auctioneers keep a record of the num-
ber of bid retractions made by a bidder. However, the
purpose for the record keeping is actually to deter bid-
ders from reneging on winning bids (i.e., win and then
later decide they don’t want the good). Retractions
that occur before an auction terminates are also typi-
cally deemed less suspicious than those that occur af-
terwards. Furthermore, keeping records on retraction
rates is largely useless as bidders can simply register
under different aliases.
Another type of price suppressing behaviour is re-
ferred to as pooling or bid rigging. Two or more bid-
ders collude, and agree not to bid up the price. How-
ever, this attack is only effective if the number of bid-
ders is small, or the majority of bidders are in collu-
sion.
4 SHILL BIDDING
Shill bidding (or shilling) is the act of introducing
fake bids into an auction on the seller’s behalf in or-
der to artificially inflate the price of an item. Bidders
who engage in shilling are referred to as shills. To
win the item, a legitimate bidder must outbid a shill’s
price. If one of the shills accidentally wins, then the
item is re-sold in a subsequent auction. Shill bidding
is a problem as it forces legitimate bidders to pay sig-
nificantly more for the item.
In March 2001, a U.S. federal grand jury charged
three men for their participation in a ring of fraud-
ulent bidding in hundreds of art auctions on eBay
(see (Schwartz and Dobrzynski, 2002)). The men cre-
ated more than 40 user IDs on eBay using false regis-
tration information. These aliases were used to place
fraudulent bids to artificially inflate the prices of hun-
dreds of paintings they auctioned on eBay.
The men hosted more than 1,100 auctions on eBay
from late 1998 until May 2000, and placed shill bids
on more than half of those auctions. The total value
of the winning bids in all auctions which contained
shill bids exceeded approximately $450,000. The to-
tal value of the shill bids in these auctions exceeded
approximately $300,000 (equivalent to 66%).
To be effective, a shill must comply to a particular
strategy which attempts to maximise the pay-off for
the seller. This section provides an insight into the
general behaviour of shills. It describes a shill’s char-
acteristics and strategies, presents examples of shill
behaviour in an auction, and discusses shill detection
techniques.
4.1 Shill Mindset
The main goal for shilling is to artificially inflate the
price for the seller beyond the limit that legitimate
bidders would otherwise pay to win the item. The
pay-off for the seller is the difference between the fi-
nal price and the uninflated price. A shill’s goal is to
lose each auction. A shill has an infinite budget. If the
shill wins, the item will have to be re-auctioned. Re-
sale of each item costs the seller both money and ef-
fort thereby eroding the possible gains from shilling.
SECRYPT 2006 - INTERNATIONAL CONFERENCE ON SECURITY AND CRYPTOGRAPHY
452
Figure 4: Aggressive Shill Bidding.
The shill faces a dilemma for each bid they submit.
Increasing a bid could marginally increase the rev-
enue for the seller. However, raising the price might
also result in failure if it is not outbid before the auc-
tion terminates. The shill must decide whether to take
the deal or attempt to increase the pay-off.
On the contrary, a bidder’s goal is to win. A bidder
has a finite budget and is after the lowest price possi-
ble. Increasing a bid for a legitimate bidder decreases
the money saved, but increases the likelihood of win-
ning.
4.2 Shill Characteristics and
Strategies
A shill has the following characteristics:
1. A shill usually bids exclusively in auctions only
held by one particular seller, however, this alone is
not sufficient to incriminate a bidder. It may be the
case that the seller is the only supplier of an item
the bidder is after, or that the bidder really trusts
the seller (usually based on the reputation of previ-
ous dealings).
2. A shill tends to have a high bid frequency. An
aggressive shill will continually outbid legitimate
bids to inflate the final price. Bids are typically
placed until the seller’s expected payoff for shilling
has been reached, or until the shill risks winning
the auction (e.g., near the termination time or dur-
ing slow bidding).
3. A shill has few or no winnings for the auctions par-
ticipated in.
4. It is advantageous for a shill to bid within a small
time period after a legitimate bid. Generally a shill
wants to give legitimate bidders as much time as
possible to submit a new bid before the closing time
of the auction.
5. A shill usually bids the minimum amount required
to outbid a legitimate bidder. If the shill bids an
amount that is much higher than the current highest
bid, it is unlikely that a legitimate bidder will sub-
mit any more bids and the shill will win the auction.
6. A shill’s goal is to try and stimulate bidding. As
a result, a shill will tend to bid more closer to the
beginning of an auction. This means a shill can
influence the entire auction process compared to a
subset of it. Furthermore, bidding towards the end
of an auction is risky as the shill could accidentally
win.
The most extreme shill bidding strategy is referred to
as aggressive shilling. An aggressive shill continu-
ally outbids everyone thereby driving up the price as
much as possible. This strategy often results in the
shill entering many bids.
In contrast, a shill might only introduce an initial
bid into an auction where there has been no prior
bids with the intent to stimulate bidding. This kind
of behaviour is a common practice in both traditional
and online auctions. However, most people typically
do not consider it fraudulent. Nevertheless it is still
shilling, as it is an attempt to influence the price by
introducing spurious bids.
This is referred to as benign shilling in the sense
that the shill does not continue to further inflate the
price throughout the remainder of the auction. A be-
nign shill will typically make a “one-off” bid at or
near the very beginning of the auction.
Regardless of the strategy employed, a shill will
still be a bidder that often trades with a specific seller
but has not won any auctions.
Another factor that affects a shill’s strategy is the
value of the current bid in relation to the reserve price.
For example, once bidding has reached the reserve
price, it becomes more risky to continue shilling.
However, this is conditional on whether the reserve is
a realistic valuation of the item that all bidders share.
4.3 Shill Bidding Examples
Figure 4 illustrates an example auction with aggres-
sive shilling. The shill aggressively outbids a legit-
imate bid by the minimal amount required to stay
ahead, and within a small time period of the last bid.
The shill bids force the other bidders to enter higher
bids in order to win. The shill does not win the auc-
tion despite the high number of bids;
UNDESIRABLE AND FRAUDULENT BEHAVIOUR IN ONLINE AUCTIONS
453
Figure 5: Benign Shill Bidding.
Figure 5 illustrates an example auction with benign
shilling. Initially no bids have been made. A shill
bid is entered for $10 to try stimulate bidding. After
seeing that there is some demand for the item, other
bidders eventually submit bids for the item. The shill
does not enter any further bids.
4.4 Shill Detection
There is often much confusion regarding what con-
stitutes shill behaviour. Bidding behaviour that might
seem suspicious could in fact turn out to be innocent.
Furthermore, a shill can engage in countless strate-
gies. This makes it difficult to detect shill bidding.
While the online auctioneers monitor their auctions
for shilling, there is no academic material available
on proven shill detection techniques.
(Wang et al, 2002) suggest that listing fees could
be used to deter shilling. Their proposal charges a
seller an increasing fee based on how far the winning
bid is from the reserve price. The idea is to coerce
the seller into stating their true reserve price, thereby
eliminating the economic benefits of shilling. How-
ever, this method is untested and does not apply to
auctions without reserve prices.
We are developing techniques to detect shill bid-
ding (Trevathan and Read, 2005). Our method ex-
amines a bidder’s bidding behaviour over several auc-
tions and gives them a shill score to indicate the de-
gree of suspicious behaviour they exhibit. Bidders
who engage in suspicious price inflating behaviour
will rate highly, whereas those with more regular bid-
ding behaviour will rate low. A bidder can examine
other bidder’s shill scores to determine whether they
wish to participate in an auction held by a particular
seller.
The shill score has been tested using simulated auc-
tions involving real world people. To facilitate testing
we implemented an online auction server (see (Tre-
vathan and Read, 2006)). Several types of tests have
been conducted. The first type involves auctioning
fake items to real bidders. Each bidder is allocated
a random amount of money, which they use to bid
in the auction. Even though bidders don’t actually re-
ceive the item, these tests manage to recreate the men-
tal drive and desire to win. Winners are excited and
often boastful after a hard fought auction. One per-
son (namely the author) is tasked with being a shill
in order to stimulate bidding. The shills goal is to in-
crease the price as much as possible, without actually
winning the auction.
When the shill score is used on these auctions, it
clearly identifies the shill bidder. It also exonerates in-
nocent bidders that bid in a regular manner. The shill
scores for a series of tests is given in Figure 6. In this
case, the bidder known as Shelly is the shill bidder.
Shelly engaged in aggressive shilling behaviour and
consequently has a shill score that is over nine. This
is clearly much higher than the other bidders. The
results from these tests are thus far encouraging.
We have also conducted similar auctions using real
items (e.g., bottles of wine and collector’s edition
playing cards), where the winner was required to pay
real money. In these settings, bidders are more cau-
tious. However, the shill was still able to influence
the auction proceedings and also was detected by the
shill score. (Note that all shilling victims were fully
reimbursed!) Furthermore, the shill score has been
tested using commercial auction data and simulations
involving automated bidding agents.
5 BID SNIPING
Bid sniping is another undesirable type of bidding be-
haviour. A bidder who employs a sniping strategy is
referred to as a sniper. A sniper will only bid in the
closing seconds of an auction, thereby denying other
bidders time to react. This essentially prevents the
sniper from being outbid.
Figure 7 illustrates the mechanics of bid sniping.
Regular bidders enter their bids as normal. In the clos-
ing seconds, the sniper enters a bid for the minimum
required to win (i.e., $41). The bid entered by a sniper
is referred to as a sniper bid. None of the other bid-
ders have time to outbid the sniper bid, and therefore
the sniper wins.
Sniping behaviour is the exact opposite of shilling.
A sniper’s goal is to win the auction for the lowest
SECRYPT 2006 - INTERNATIONAL CONFERENCE ON SECURITY AND CRYPTOGRAPHY
454
Figure 6: An example of the Shill Score when run on auction simulations containing one aggressive shill.
Figure 7: Bid Sniping.
price. Whereas a shill’s goal is not to win and to in-
flate the price. Sniping is often used as a preventa-
tive measure against shilling. A sniper may not be
able to prevent shilling occurring during an auction.
However, the sniper can prevent themselves from be-
ing shilled.
Sniping disadvantages regular bidders in that they
are denied the opportunity to respond to the sniper
bid. Bidders are typically frustrated when they realise
that sniping has occurred. This is especially the case
when a bidder has observed an auction for a long pe-
riod of time, only to be beaten in the closing seconds.
The seller is also potentially disadvantaged by snip-
ing, as the sniper bid does not stimulate rival bidding,
that might have occurred, had other bidders been able
to respond.
Sniping is permitted on eBay, although its use is
discouraged. Instead eBay recommends that a bidder
should only place a single bid at his/her maximum
valuation using the proxy bidding system. Despite
this recommendation, sniping is rampant, and is now
considered as a natural part of the online auctioning
experience. (Shah et al, 2002) performed a study into
the amount of sniping in 12, 000 eBay auctions. Their
results showed for the majority of auctions, a signifi-
cant fraction of bidding occurs in the closing seconds.
A sniping agent is a software bidding agent that fol-
lows a sniping strategy. The sniping agent constantly
monitors an auction, and waits until the last moment
to bid. Many companies now exist such as Bidnap-
per.com
3
, ezsniper
4
and Auction Sniper
5
which of-
fer sniping agents for use on eBay auctions.
uBid auctions differ to eBay in that auctions termi-
nate using a timeout session. Once the ending time of
an auction has been reached, the auction is extended
for ten minutes for each bid received. This limits
the effectiveness of sniping, but can lead to a show
down between snipers. As a result the auction can run
for a lot longer than the seller anticipated. The seller
can enforce a maximum extension limit to prevent the
auction continuing indefinitely. However, this results
in the auction being essentially the same as a normal
auction without a timeout session.
The only preventative measure for a bidder against
sniping, is to “out-snipe” the sniper. However, this
often results in there being multiple snipers in an auc-
tion. This behaviour leads to failure of the English
auctioning process. If everyone engaged in sniping,
the auction would essentially become a sealed bid
auction. In a sealed bid auction, bidders submit their
bids secretly during a bidding round. At the close of
3
http://www.bidnapper.com
4
http://www.ezsniper.com
5
http://www.auctionsniper.com
UNDESIRABLE AND FRAUDULENT BEHAVIOUR IN ONLINE AUCTIONS
455
bidding, the Auctioneer determines the winner. En-
glish auctions on the other hand are open bid, and al-
low bidders to bid multiple times. In a traditional of-
fline English auction, sniping cannot occur. Sniping is
a feature unique to online auctions. Sniping behaviour
blurs the boundaries of an online auction between the
type of auction it is and the rules that govern it.
6 SIPHONING
Siphoning (or bid siphoning) refers to the situation
where an outsider observes an auction and contacts
bidders offering them an identical item at a better
price. The outsider is referred to as a siphoner, and is
said to “siphon” bids from the auction. The siphoner
benefits in that he/she does not incur any of the costs
involved with organising and advertising an auction.
Siphoning disadvantages the Auctioneer through
lost revenue. That is, the siphoner does not have to
pay the Auctioneer to list/advertise an item. Siphon-
ing also disadvantages the seller whose auctions are
being siphoned. This is in a form of price under-
cutting (i.e., the siphoner offers the item at a better
price), and reduces the demand for the seller’s items.
A bidder who does business with a siphoner loses the
protection offered by the auction, and exposes them-
selves to fraud (e.g., misrepresented or non-existent
items).
Siphoning may also be used in conjunction with
shilling. When a shill bid accidentally wins, the seller
of the item can contact the next highest bidder, and
directly offer them the item. This saves the seller the
time and expense of re-auctioning the item.
Consider the following scenario: A seller is auc-
tioning off a traditional Japanese sword. A legitimate
bidder enters a bid for $1000. The seller then enters
a shill bid for $1200. The legitimate bidder refrains
from bidding and the shill bid wins. Later on the le-
gitimate bidder is approached by the seller. The seller
claims that the winner is a dead beat
6
, or has backed
out of the deal due to financial or other personal rea-
sons. The seller then offers the item to the bidder at,
or near, the bidder’s price of $1000.
Siphoning combined with shilling disadvantages
both the bidders and the Auctioneer. Bidders are
forced into paying an inflated price due to the shill
bids. The siphoning component this time does not
affect the seller, but rather the Auctioneer. This is be-
cause the seller does not have to repeat the auction and
is denying the Auctioneer revenue from listing fees.
The seller in effect has “siphoned” bids from his/her
own future auctions.
6
A bidder that has won an auction and to fails to make
payment.
Siphoning is impossible to detect by the Auctioneer
alone. Instead bidders must report the behaviour once
it has happened to them. However, most bidders are
aware of what siphoning is, and probably wouldn’t
recognise that they have been siphoned. Furthermore,
if a bidder receives a better price, or an item they re-
ally want, then they would not see such behaviour as
undesirable. In this case they are unlikely to report it.
There is no clear law regarding siphoning. The
only advice to bidders is to decline communication
with anyone that contacts them outside of the official
channels. In addition, if you were not the winner of
an auction, don’t accept an item from the seller after
the auction has ended.
7 NON-EXISTENT OR
MISREPRESENTED ITEMS
A dubious seller might attempt to auction a non-
existent item, or misrepresent an item. In the first
instance, the seller accepts payment from the buyer
but doesn’t deliver the item. In the second, the seller
misrepresents the item by advertising it as something
it isn’t, or delivers an item of lesser value. In the shill
case described in Section 4, the men misrepresented
paintings as being significant works when they were
inexpensive replicas (see (Schwartz and Dobrzynski,
2002)).
To ensure that an item conforms to its description,
eBay recommends that buyers study the seller’s pho-
tos. However, this is unsatisfactory as the seller can
simply copy pictures from a legitimate item, and post
these for the non-existent/misrepresented item.
eBay also recommends asking the seller questions
regarding the item. However, a seller may simply lie.
Furthermore, some bidders might be reluctant to ask
questions as they desire anonymity. For example, if
the bidder has a high profile, they might want to con-
ceal the fact that they are bidding. This might oc-
cur in the case where the bidder feels that their pri-
vacy would be compromised in some manner (e.g., re-
veal that they have a fetish for an item), change other
bidders’ perceptions of the item (e.g., stimulate un-
wanted bidding), or influence the seller to raise the
reserve price.
Another recommendation is to check seller feed-
back. eBay has a system where buyers and sellers
can leave feedback regarding their dealings with each
other. A party to a transaction can rate their experi-
ence as either good, neutral or bad. An individual is
given a rating based on the feedback received. How-
ever, feedback ratings are not a reliable measure of
an individual’s integrity, and feedback can be falsely
generated using multiple bidder accounts.
Online auctioneers offer a dispute resolution proce-
SECRYPT 2006 - INTERNATIONAL CONFERENCE ON SECURITY AND CRYPTOGRAPHY
456
dure where a buyer can initiate action against a seller
if an item was not received, or if the item is different
from what was expected. If a seller constantly does
not deliver items, the Auctioneer can revoke his/her
account. However, the seller can simply re-register
under a different alias.
Insurance can be offered to buyers for items not de-
livered. eBay has created a successful off-shoot busi-
ness, PayPal
7
, for guaranteeing online transactions.
However, launching an insurance claim can be an ar-
duous process, and might not fully compensate the
victim. Furthermore, it can require the compensated
funds to be spent by participating in future auctions.
In addition, it is only a solution that can be used after
fraud has occurred.
Escrow fraud is an emerging threat to online auc-
tions. A fraudster sets up a fake escrow service for
paying a seller for an item. When the winning bid-
der wires the money to the escrow agency, the agency
vanishes along with the payment. Neither the seller
receives the payment, nor does the winner receive the
item. The following are typical characteristics of an
escrow scam:
1. Check for poor grammar on the escrow site.
2. Although site may look authentic, it is usually
copied from a legitimate site such as Escrow.com
8
and Auctionchex
9
.
3. There are obvious give-aways in the Terms page,
which is generally stolen from another site.
4. A site will often leave hints of what its previous
incarnation was - especially if they’ve just changed
domain names recently.
5. Be wary if the seller insists on using a specific es-
crow site. Sellers don’t usually press for escrow,
buyers do.
Bidders must be wary of the escrow service they use,
and not do business with unknown or un-trusted es-
crow vendors.
8 CONCLUSIONS
Online auctions are susceptible to undesirable and
fraudulent behaviour. Such behaviour is designed to
influence the auction in a manner that either favours
a bidder, the seller, or an outsider seeking to profit
from the auction. This results in market failure, and
reduces the usefulness of online auctions as an ex-
change medium.
This paper investigates undesirable and fraudulent
trading behaviour, and discusses its implications for
7
http://www.paypal.com
8
http://www.escrow.com
9
https://auctionchex.com
online auctioning. We show how to identify such
behaviour and give recommendations for recourse
against undesirable and fraudulent participants.
Bid shielding is a practice employed by one or
more bidders to suppress bidding and keep the price
low. There is no means to protect against bid shield-
ing, other than to exclude users with a high bid retrac-
tion rate. However, a bidder can use a fake name to
re-register.
Alternately shill bidding is a practice employed by
the seller to artificially inflate the price by introduc-
ing spurious bids. Shill bidding is strictly forbidden
by commercial online auctions, and is a prosecutable
offence. There is limited material on shill bidding and
prevention/detection techniques. We are presently de-
veloping a method to detect shill bidders by giving
them a ranking called a shill score. A bidder can use
the shill score to decide whether they want to partici-
pate in an auction held by a particular seller.
Bid sniping is a strategy employed by a bidder to
prevent being outbid. A sniper submits a bid dur-
ing the closing seconds of an auction, thereby deny-
ing other bidders time to react. Many bidders engage
in sniping behaviour in an attempt to prevent them-
selves from being shilled. Sniping is permitted in on-
line auctions, but is discouraged. Commercial sniping
agents are available that engage in sniping behaviour
on the bidder’s behalf. Sniping can be reduced by en-
forcing a timeout limit which extends the auction by
several minutes for each new bid received after the
auction’s termination time. The prominence of snip-
ing behaviour raises concerns regarding the effective-
ness of online auctions to conduct English auctions
according to traditional rules.
Siphoning refers to the situation where an outsider
observes an auction and offers an identical item to the
bidders at a lower price. The siphoner avoids all of
the costs associated with conducting an auction, and
effectively profits from the seller. Siphoning is often
used in conjunction with shilling, and scams involv-
ing non-existent or misrepresented items. Siphoning
behaviour should be reported to the Auctioneer, how-
ever, there is little more that can be done.
A malicious seller can offer non-existent or misrep-
resented items to profit from unsuspecting buyers. It
is recommended that bidders inspect photos and ask
questions. However, a seller can post fake photos, lie
in response to questions, and a buyer might want to re-
main anonymous. eBay’s feed back rating is a mecha-
nism for gauging the integrity of a buyer/seller, how-
ever, it is dubious at best. Insurance can reimburse
victims of fraud, but is not a complete solution. Es-
crow fraud is emerging as a threat to insurance-based
fraud prevention measures and is difficult to identify.
The best advice is for a buyer to be wary of the goods
they are bidding for.
The popularity of online auctions and the over-
UNDESIRABLE AND FRAUDULENT BEHAVIOUR IN ONLINE AUCTIONS
457
whelming number of flawless transactions, is a tes-
tament to the success of online auctioning. Never-
theless, some participants are disadvantaged by unde-
sirable or fraudulent behaviour. Preventing such be-
haviour is a difficult task and existing solutions are
largely inadequate.
Buying items from online auctions is the same as
buying items anywhere online. A buyer cannot physi-
cally inspect the merchandise as in a “bricks and mor-
tar” store. The buyer is forced to rely on the item’s
description. Distances between buyers and sellers can
be vast. This makes it hard to police transactions that
go awry, especially when buyers and sellers cross po-
litical and cultural boundaries.
The best recommendation is caveat emptor. Don’t
deal with unknown sellers, or sellers who reside in
countries which do not have strict enforcement of in-
ternational commerce laws. Ensure that you thor-
oughly research the item you are bidding for.
REFERENCES
Franklin, M. and Reiter, M. (1996). The Design and Imple-
mentation of a Secure Auction Service, IEEE Trans-
actions on Software Engineering, vol. 22, 302-312.
Schwartz, J. and Dobrzynski, J. (2002). 3 men are charged
with fraud in 1,100 art auctions on eBay, in The New
York Times.
Shah, H., Joshi, N. and Wurman, P. (2002). Mining for Bid-
ding Strategies on eBay, in SIGKDD’2002 Workshop
on Web Mining for Usage Patterns and User Profiles.
Stubblebine, S. and Syverson, P. (1999). Fair On-line Auc-
tions Without Special Trusted Parties, in Proceed-
ings of Financial Cryptography 1999, vol. 1648 of
Lecture Notes in Computer Science, Springer-Verlag,
230-240.
Trevathan, J., Ghodosi, H. and Read, W. (2005). De-
sign Issues for Electronic Auctions, in Proceedings of
the 2nd International Conference on E-Business and
Telecommunication Networks (ICETE), 340-347.
Trevathan, J., Ghodosi, H. and Read, W. (2006). An
Anonymous and Secure Continuous Double Auction
Scheme, in Proceedings of the 39th International
Hawaii Conference on System Sciences (HICSS),
125(1-12).
Trevathan, J. and Read, W. (2005). Detecting Shill Bidding
in Online English Auctions, Technical Report, James
Cook University.
Trevathan, J. and Read, W. (2006). RAS: a system for sup-
porting research in online auctions, ACM Crossroads,
ed. 12.4, 23-30.
Viswanathan, K., Boyd, C. and Dawson, E. (2000). A
Three Phased Schema for Sealed Bid Auction System
Design, Proceedings of ACSIP 2000 - Australasian
Conference on Information Security and Privacy, vol.
1841 of Lecture Notes in Computer Science, Springer-
Verlag, 412-426.
Wang, W., Hidvegi, Z. and Whinston, A. (2002). Shill Bid-
ding in Multi-round Online Auctions, in Proceedings
of the 35th Hawaii International Conference on Sys-
tem Sciences (HICSS).
SECRYPT 2006 - INTERNATIONAL CONFERENCE ON SECURITY AND CRYPTOGRAPHY
458