The poor Leu last second sniper ...

[Heliodromus]

This has been talked about before, both here and on other forums, but here we go again...

I was bidding in recent Leu (fixed closing time) Web auction # 30, and once again there are a bunch of funny results...

Some poor "person" keeps sniping in the last few seconds (often last 1-3 seconds!) of the auction, and losing, despite putting in super high bids, well ahead of the field. They keep losing because despite their last second snipe, they don't *quite* bid high enough. I say not *quite* high enough, because in previous auctions where I've been the (unfortunate?) high bidder, this last second sniper has pushed me all the way to my own maximum snipe of a few seconds earlier. Just one increment higher and they would have won ...

Here's are some examples (Leu Web 30).

Note closing time of lot at bottom of screen grab.

The winner, sniping at T-18 sec would have won the lot for one increment over EUR 300, except for Bidder 4 sniping at T-5, pushing them up to EUR 650.

The following lot, same rare type, in moderately worse condition hammered for EUR 85.

Here's another:

Bidder 4 was all set to win for EUR 70 before Bidder 5 bids at 1 second before closing, pushing them up to EUR 110.

This last second sniper sure is fast, but these last second bids never win ...

Here's another at 2 seconds before the close.

Some at 3 seconds before the close

I didn't bid in any of the above lots, but it didn't take much looking at some of the lots going higher than expected to find these.

There are more, but you get the idea. Last second very high losing bids, which have the effect of pushing the winner to pay much higher than they would have otherwise, maybe all the way to their "hidden" maximum.

Despite being an inveterate sniper myself, my advice for fixed end time Leu actions would be not to snipe, at least not at a price well above the current bid. A safer strategy, albeit one that carries the risk of losing, is to wait until last 30 seconds or so, then run the bid up in a number of smallish steps until you are the winner. Put in a final slightly higher bid if you feel like it, but in doing so you risk that this higher amount is what you will unnecessarily pay.

 

 

Edited July 22 by Heliodromus

[Salomons Cat]

I understand your concern. But I also bid in this way. When I really want a coin, I wait until the last seconds to place a bid that I consider high enough. Therefore, I don’t think this observation can prove anything. However, shill bidding is always a possibility and not difficult to do. I just wouldn’t accuse anyone publicly because it’s impossible for us to prove.

[kirispupis]

I've written about this before. Essentially there are four types of shill bidding:

1. Consignor shill bidding. This typically happens without being condoned by the auction house. Most bidding systems make it possible for a bidder to continuously raise his bid until hitting the bidder's max.
2. Auction house shill bidding. From my suspicions, this is extremely common. There are only two auction houses that I'm pretty sure don't do this to some extent. If you search my posts here, I did some data analysis on different houses where I've bid. There are some where the shill bidding is quite obvious, and others where it's uncertain.
3. Predictive auction house shill bidding. This is the same as #2, but the auction house knows your max. When you see someone winning a coin at 100 with a max of 1200, and at the last second there's a bid to bring it up to 1200, that's suspect.
4. Automated predictive auction house shill bidding. I've never proven that an auction house does this, but it's certainly possible. In this situation, bidding at the last moment won't save you because the computer is always faster. You'll put in a 1200 bid on a 100 coin at the last second, and you'll win it for 1200...

Personally, I combat these the way most others do:

1. Primarily bid live, except in the handful of cases where I trust the auction house
2. Bid at the last second or two if possible
3. Be prepared to pay your max. If that's too much, don't bid it.

[Heliodromus]

Edit: replying to Salomons Cat

Sure, but if we've got just 2 bidders bidding well ahead of the pack, then all things being equal I'd expect the person who snipes last to win at least 50% of the time ...

And, yes, impossible to prove any shenanigans (except statistically), since bidder's high bids are nominally private, but if you see this same pattern happening to coins you are personally bidding on, where you are privvy to the high bid, then you can make a better judgement.

In any case, the "run the bid up in small steps" strategy is a safer way to go. Anywhere else, I just snipe.

 

Edited July 22 by Heliodromus

[Salomons Cat]

They often have several hundred lots, so these examples don’t seem so surprising to me. But maybe I’m just a bit more hesitant with such accusations because I’m from Switzerland. It takes me only an hour by train to reach Leu, where I usually pick up my coins in person. The people who work there seem nice.

Edited July 22 by Salomons Cat

[Heliodromus]

8 minutes ago, Salomons Cat said:

They often have several hundred lots, so these examples don’t seem so surprising to me.

They'd not be surprising if there were a similar number of lots where in similar circumstances (2 bidders ahead of the pack) the final 1-3 seconds bidder was the winner.

It's the statistical skew - in a number of axis - that makes this stand out.

BTW, I'm not paranoid about auction houses/platforms in general - I always snipe, and have never noticed anything improper looking elsewhere.

If there were any shennanigans, them I'm sure the front office staff are blissfully unaware.

[Salomons Cat]

I just took a quick look at the lots 1845 to 1945. I did not see anything suspicious there. In fact, many of the winning bids were from the last 10 seconds.
And anyway - why should a shill bidding system only work during the last seconds? 
If you want to drive up the prices with an automatic system, you should start earlier. In contrast, bidding in the last 10 seconds is typical human behaviour.

[Heliodromus]

41 minutes ago, kirispupis said:

Be prepared to pay your max. If that's too much, don't bid it.

That's certainly true, but the sniper's mindset is that you'll win some, lose some. You are necessarily bidding the absolute maximum you're prepared to pay, because that's what sniping calls for, but you expect a mix of outcomes. It'd obviously be extremely weird if you only ever lost or won at your exact maximum, and never won below your maximum!

Anyways, being of a scientific mindset, I've observed the evidence and am bidding in accordance to achieve best expected outcome (winning at market price +/-).

[Heliodromus]

14 minutes ago, Salomons Cat said:

In fact, many of the winning bids were from the last 10 seconds.
And anyway - why should a shill bidding system only work during the last seconds? 

Yes, I would expect most winning bids to be in last seconds. The whole purpose of a snipe is to not give another (human) bidder a chance to react, so 6-10 seconds is about right. Remember that it may take a couple of seconds for the server to receive and process your bid, depending on how far away you are (how many routers are you passing through), and what kind of load the server is under, so if you are unnecessarily aggressive and bid in last 2-3 seconds you run the risk of your bid not being accepted.

I suppose you could conceive of a shill bidding system that bid multiple times, and that would be harder to detect, but if you are greedy and want to take advantage of snipers, then you'd need to shill bid in last few seconds after the expected last 6-10 second snipes have been received. Of course you could also do both.

 

Edited July 22 by Heliodromus

[kirispupis]

10 minutes ago, Heliodromus said:

That's certainly true, but the sniper's mindset is that you'll win some, lose some. You are necessarily bidding the absolute maximum you're prepared to pay, because that's what sniping calls for, but you expect a mix of outcomes. It'd obviously be extremely weird if you only ever lost or won at your exact maximum, and never won below your maximum!

Anyways, being of a scientific mindset, I've observed the evidence and am bidding in accordance to achieve best expected outcome (winning at market price +/-).

In a fair world, yes. However, in my case there have been auctions where I've won everything at essentially my max (granted, I didn't snipe in those cases).

[Brennos]

It is not known whether the winning bid is the maximum amount bid by the buyer. However, it seems to be demonstrated that the "underbidder" is at best a hidden reserve price and at worst a kirispupis' type 3 shill bidding.

 

[Hesiod]

1 hour ago, Brennos said:

It is not known whether the winning bid is the maximum amount bid by the buyer. However, it seems to be demonstrated that the "underbidder" is at best a hidden reserve price and at worst a kirispupis' type 3 shill bidding.

 

It is when the bid is off increment – if you're the normal bidder and someone else bids the exact increment, it'll show an odd number (for example: 2796 has a bid 31)

I don't recall the exact increments at the levels in leu though.

[rasiel]

3 hours ago, Heliodromus said:

And, yes, impossible to prove any shenanigans (except statistically), since bidder's high bids are nominally private, but...

 

Not sure if this helps but I can relate that on the lots I've won from Leu I'm averaging 84.5% of max bid with the lowest being a 29% and (of course) several right at 100%.

This seems in line with other major auction houses like NAC (89%), N. Naumann (82%), Roma (91%) and Rauch (87%). Oddly, CNG, the only US-based auctioneer I buy from frequently, is quite a bit higher. 100% in fact. Just shitty luck, I guess.

Rasiel

[Postvmvs]

I think we have all had a "feeling" that there is a sinister force that bids almost up to our max just to make us pay more, but is there really anything to it?

Just historical interest, but per the Wikipedia for Heritage Auctions:

In 2009, Heritage Auctions faced a lawsuit from a former employee named Gary Hendershott. The allegation made in the lawsuit was that the company had participated in fraudulent activity by employing a "shill" bidder under the pseudonym "N.P. Gresham". This shill bidder was purportedly used to manipulate bidding prices artificially, an action claimed to be in violation of anti-racketeering laws.^[1][2] The lawsuit was dismissed with prejudice.^[3]

[Heliodromus]

16 minutes ago, Postvmvs said:

I think we have all had a "feeling" that there is a sinister force that bids almost up to our max just to make us pay more, but is there really anything to it?

If there's no evidence, just a feeling, then it's probably just your paranoia!

 

[ominus1]

43 minutes ago, Postvmvs said:

a sinister force

...well hey...ya gots ta have a a Mojo....:)

[Heliodromus]

1 hour ago, rasiel said:

Oddly, CNG, the only US-based auctioneer I buy from frequently, is quite a bit higher. 100% in fact. Just shitty luck, I guess.

That sounds very odd. Are we talking about the same thing here - submitting a single bid ahead of time (whether days or seconds), well above the current bid, then eventually ending up winning it at exactly your max bid... and ALL your bids at CNG have been like this?! More than just a 2-3? Have you lost any bids at CNG?

[kirispupis]

FWIW, here's my analysis of shill bidding.

In terms of CNG, I buy there often and they're one of the few I'm quite certain does not shill bid. The competition is simply high IMHO. I've lost many coins there, but many of those I've won are not near my max.

 

[Nerosmyfavorite68]

Or, just do like I do and only buy fixed price coins.  It's so much less stress and no shills!  If I think the price is somewhat reasonable, I buy the coin.

It also gives me the freedom of buying at a convenient time to me, buying as many or as few as I like, and having cost certainty.

[rasiel]

2 hours ago, Heliodromus said:

That sounds very odd. Are we talking about the same thing here - submitting a single bid ahead of time (whether days or seconds), well above the current bid, then eventually ending up winning it at exactly your max bid... and ALL your bids at CNG have been like this?! More than just a 2-3? Have you lost any bids at CNG?

Yes, exactly what I'm saying. And have lost a bunch too (at an average of 148% my max). I know they're squeaky clean. Just bad luck like I said.

[wuntbedruv]

13 hours ago, Heliodromus said:

This has been talked about before, both here and on other forums, but here we go again...

I was bidding in recent Leu (fixed closing time) Web auction # 30, and once again there are a bunch of funny results...

Some poor "person" keeps sniping in the last few seconds (often last 1-3 seconds!) of the auction, and losing, despite putting in super high bids, well ahead of the field. They keep losing because despite their last second snipe, they don't *quite* bid high enough. I say not *quite* high enough, because in previous auctions where I've been the (unfortunate?) high bidder, this last second sniper has pushed me all the way to my own maximum snipe of a few seconds earlier. Just one increment higher and they would have won ...

Here's are some examples (Leu Web 30).

Note closing time of lot at bottom of screen grab.

The winner, sniping at T-18 sec would have won the lot for one increment over EUR 300, except for Bidder 4 sniping at T-5, pushing them up to EUR 650.

The following lot, same rare type, in moderately worse condition hammered for EUR 85.

Here's another:

Bidder 4 was all set to win for EUR 70 before Bidder 5 bids at 1 second before closing, pushing them up to EUR 110.

This last second sniper sure is fast, but these last second bids never win ...

Here's another at 2 seconds before the close.

Some at 3 seconds before the close

I didn't bid in any of the above lots, but it didn't take much looking at some of the lots going higher than expected to find these.

There are more, but you get the idea. Last second very high losing bids, which have the effect of pushing the winner to pay much higher than they would have otherwise, maybe all the way to their "hidden" maximum.

Despite being an inveterate sniper myself, my advice for fixed end time Leu actions would be not to snipe, at least not at a price well above the current bid. A safer strategy, albeit one that carries the risk of losing, is to wait until last 30 seconds or so, then run the bid up in a number of smallish steps until you are the winner. Put in a final slightly higher bid if you feel like it, but in doing so you risk that this higher amount is what you will unnecessarily pay.

 

 

I had a similar experience at an online auction a few weeks back (quite a well regarded one, actually).

A number of lots, one of which I went for, went on for about 2-3 minutes of bidding due to an apparently extremely slow bidder countering leading bidders, then waiting until the very last seconds of the final call to bid again. I've never seen anything like it, almost seemed automated and robotic. You'd expect maybe that someone might pause to think before bidding again, but not leave it right to the last second every bid. Can't have been a pre-bid automated response either as that would be instant and increment up from the underbidder's bids more quickly.

I suspect something sinister, but of course I have no proof...

[Heliodromus]

8 hours ago, rasiel said:

Yes, exactly what I'm saying. And have lost a bunch too (at an average of 148% my max). I know they're squeaky clean. Just bad luck like I said.

I wouldn't characterize that as unlucky ... if you are either losing (by quite a bit) or winning at your max, it seems more suggestive of you not being overly aggressive with bidding (a good thing!), more like "barely winning" than "being run up to maximum".

I think the potential for an "exposed" high bid above the field bid being abused is more likely to occur when either the material being bid on is specialist and hard to price (eg. Ramskold coll. just sold at Leu - a collection of Constantinian Rome mint coins assembled since 1969), or maybe there is just an overenthusiastic bidder unaware of values!

 

[Heliodromus]

2 hours ago, wuntbedruv said:

A number of lots, one of which I went for, went on for about 2-3 minutes of bidding due to an apparently extremely slow bidder countering leading bidders, then waiting until the very last seconds of the final call to bid again.

That sounds more like a human bidding pattern to me - in an open ended (not fixed end time) auction I'll typically wait until last second of the 3-2-1 to bid, to allow other bids to come in first.

Note too the difference between the suspicious bidder being careful not to actually have the high bid (they typically don't want to win, unless this is the consignor implementing a reserve below which they'll buy it back), and what your describing where the counterbidder is outbidding others and therefore willing to win.

[Aleph]

Has anyone done the analysis of the hit rate of the Phantom of Leu?  Last minute snipes should be successful a fairly consistent amount of the time.  If last second snipes consistently fail, this is a real red flag.  For what it’s worth, I have no doubt Leu has been dishonest in the past and I significantly reduced my participation in their auctions as a result.  Roma did this too and we have all seen how moral their director is after the Eid mar aureus fiasco.

Be prepared to pay if you want to bid successfully in their auctions.

Edited July 23 by Aleph

[kirispupis]

33 minutes ago, Aleph said:

Has anyone done the analysis of the hit rate of the Phantom of Leu?  Last minute snipes should be successful a fairly consistent amount of the time.  If last second snipes consistently fail, this is a real red flag.  For what it’s worth, I have no doubt Leu has been dishonest in the past and I significantly reduced my participation in their auctions as a result.  Roma did this too and we have all seen how moral their director is after the Eid mar aureus fiasco.

Be prepared to pay if you want to bid successfully in their auctions.

See the research I linked to in my post above. The challenge is without knowledge of what the actual max bid was, it's difficult to gauge how "accurate" these last second snipes are. I performed some analysis on my own bids, where I obviously knew the actual max, and the results certainly suggest some predictive sniping, but weren't a smoking gun.

These auctions are obviously competitive, and a lot of people bid in the last seconds. Therefore, it's difficult to truly say that the results are due to shill bidding vs simple competition. However, I remain highly suspicious.

Regardless of whether such behavior takes place, it's caused me to hold back on Leu auctions. I similarly held back at Roma. Only if the coin is a top priority and unlikely to be procured elsewhere do I bite. In their last auction, several coins I was intrigued about went for prices considerably below what my max would have been. However, I have no regrets because I just couldn't trust that I would have wound up at my max. IMHO this hurts consignors.