Please explain the purpose of "sniping" (ebay)

Factual Questions
181

Yes, that’s right.
That’s twice now that someone has pointed out a “contradiction” that’s nothing of the sort.

I agree with all this. The point is, where did this $14.49 come from?
The proposition of many here is that one begins with such a value, independent of the market. It’s “folk economics” and it’s not true.

Imagine I take you to an alien world and pay you a salary of 4,000 Foobars a month. You need to pop to the alien hardware store and buy a wrench.
How many Foobars is your maximum, and why?

I put it to you that your maximum is set, not entirely by you, and not entirely by the market either, but the interaction between the two. You decide whether to trade X Foobars with a seller. You decide how many sellers to see, and whether to explore alternatives to buying a spanner. It’s this process by which you calculate your maximum.

Hearing you loud and clear fellas :slight_smile:

182

It has an impact because, the part I bolded, you got wrong. The bids are NOT evaluated “when the auction has ended.” Bids are evaluated when they are placed, and the current “high bid or second highest bid + increment” is constantly updated and provided to all bidders and potential bidders.

It is this fact that causes some bidders to increase their previous bids, driving the price higher. Sniping has an impact because the late bids do not change this public information in time for anyone to increase their bid as a result.

For all practical purposes, very late bids are evaluated after the ending time. If all the bids were placed in the last seconds, then your description would apply, and it is exactly this result that snipers are striving for. Earlier bids just provide a price floor above which this takes place.

Also, you have neglected to account for the existence of shilling. In order to shill, the seller has to enter a high bid, cancel it, and enter second bid before the auction closes. Sniping is an effective tool for combating shills, as it does not leave time to accomplish the last steps. Shilling sellers are then forced to win thier own auctions, and try to sucker the legitimate high bidder into accepting a “second chance offer” at their max bid level…which the bidder is free to reject.

Shilling not common you say? IME I get such second chance offers for about 5% of the auctions I don’t win. I win about 20-25% of the auctions I bid, so I suspect that I am not the highest legitimate bidder something like 4/5 to 3/4 of the time. That implies that second chance offers are made in around 15-20% of auctions I bid for. Sure, some of these may be ligit, but if even the majority are ligit, it tells me that shilling is rampant and worth the free and easy act of insuring against by sniping.

183

Well, part of the problem of “following” what you’re saying is that you’re kind of meandering all over the place. You didn’t present a clear argument. In post #121, you talk only about reinventing the manual proxy bid mechanism. This is a totally separate issue from “market research” to determine a value.

So… for the last few posts, you’ve abandoned the extended overtime auction idea and emphasized the notion of a fuzzy value that converges more precisely with additional market information. This is reasonable but you’ve failed to explain how live auctions help that.

You are mis-characterizing what people trying to explain. Yes, there is a value that can be determined from the overall market but messing around with manually adjusting bids on a live unfinished auction doesn’t tell you the that information. How can it when the other bidders max proxy bids are secret?

Hear you loud and clear sir but that’s not relevant to the context of a live auction. You have to clarify whether you’re bidding on an auction to actually win or doing informal “market research” (with the unexpected bonus of possibly winning.) If you’re truly doing “market research”, you’re really not the “victim” of any sniping because that aspect of losing the auction to occasional last-second bidders was part of your Grand Plan For Mastering The Universe all along. I already acknowledged a similar “rational” strategy in post #42.

Maybe you were trying to explain your logic as what I said in post #42. Well, the only thing that’s weird is that is that it doesn’t jive with your complaint about running out of time for additional bid increments in #121.

184

Didn’t read through the whole 4 pages. Has it been mentioned that sniping is a strategy to combat irrational/emotional competing bidders?

185

Well, I did re-read the post as you suggested, and found no place in that post where you acknowledge that people have an internal sense of value; you suggested that people have to know the market value before they can quantify the value to themelves.

Well, that is a sticking point on which we fundamentally disagree, so I won’t spend any more of our collective energy on it.

186

You think four pages of a discussion on ebay bidding didn’t cover that? Could you at least try to *skim *those four pages?

187

Once. In 1960. For twenty minutes.

188

Ha! You’ve got it all wrong.

Snipers are actually lemmings that unwittingly provide marketing data to ultra savvy ebayers ready to pounce and devour them. Snipers are merely being toyed with. There was a movie made about it:

Snipers: “You’re trying to trick me into giving away something. It won’t work.”
Vizzini: "IT HAS WORKED! YOU’VE GIVEN EVERYTHING AWAY! "

189

Apologies in advance if someone explained how sniping is useless in the following scenario, but I didn’t see it.

There’s an item up for bid that one or more have bid on. It’s currently at $5 and I’m willing to pay as much as $10 for. I use a sniping service to put a bid in of $10 at the last few seconds. I win the item for $5.25.

or

Same scenario but instead of using a sniping service, I put a bid in of $10 and the auction for this item ends a few days later. In the mean time, others that have bid see that they have been outbid and at least one of them aren’t going to have that and bid until they see they have outbid me. I don’t win the item.

If you don’t think sniping stops the second scenario from occurring in many circumstances, please explain it to me.

But first see the following thread:

Thanks for the great app, ebay (assholes) - The BBQ Pit - Straight Dope Message Board

The OP was angry that he wasn’t informed in time that he was outbid. Had he been outbid by someone not using a sniping service, he would have had time and placed a higher bid and possibly outbid the other bidder. Good thing for the other bidder he used a sniping service.

190

Mijin, you still haven’t indicated in any way why you need external bidders to influence your decision? Why not just take out a piece of paper, write down your starting bid, and then write down a one increment higher bid next to it. Pretend that someone sent in that bid. Decide whether you would meet that bid. Hem, haw, take as long as you want. If you feel that seeing someone bid higher would cause you to up your bid then write down a bid one increment higher than that. Repeat this until you would no longer feel that you would bid any higher. Then enter this highest bid. Now Ebay will run that procedure for you in the auction and you can relax and play warcraft or whatever. I don’t see how this will cause you to bid more $ or lose more often than your current strategy of manually upping your bid.

In the earlier part of the thread there were some who argued in favor of incremental bidding in the cases of multiple copies of the same items on sale. In that case incremental bidding makes sense as you want to move your high bid from auction to auction. But if there is only one item to bid on, I fail to see how seeing another’s bid after your bid of X tells you any more about the market other than there is at least one person in the world who is willing to spend X+1.

191

Maybe some people would understand better if we used “emotional bidders” instead of “irrational bidders”.

192

We may be getting somewhere now, since a lot of people are framing the discussion like this, and I think it’s the sticking point.

I’m not saying that at one instant my maximum is $15, then when someone puts in a bid of $15.02, suddenly I wish to increase my maximum.
This is the “emotional bidder” straw man, and it’s not the situation that I’m trying to describe.

I’m trying to describe a game theory situation, which can potentially be exploited in proxy auctions (and probably other types of auction).

There are two truths to my bidding strategy:

[ul]
[li]A single bid increment is immaterial to me. It doesn’t make the difference between buy / refuse until we get to my absolute upper bound.[/li][li]My absolute upper bound is actually a slight overvaluation. I don’t want to regularly bid my upper bound.[/li][/ul]

Now, many posters remain convinced that I’m being irrational / emotional, but I remain convinced that it is how we all think.

Let’s try yet another hypothetical:

Imagine you are at an interview for what is your dream job. You have been told that either you or a second candidate will be given the role.
And to decide who gets the job, you both have to make a bid, which will be known to both candidates.
Whoever bids the least, gets the job.
And, unlike the real world, there is no minimum wage and no matter how low you say, if you win you will be compelled to work the job.

And in this hypothetical, importantly, the other guy will get to bid after you.

Now, what the only-emotional-bidders-can-get-sniped brigade are effectively saying is that there’s no advantage in the other guy getting to go second. You should just bid the least you’d be willing to work for, and it’s fair whether you win or lose.

But I think trying to say what your absolute lower bound is, is very hard. I’d rather be the second guy, with the option of undercutting his opponent by $1, or walking away and making me actually work for whatever absurd low figure I said.

(to make it directly analogous to proxy bidding we may need to stipulate that there are multiple people going for the job, and it’s one bid increment below the second lowest bid etc etc. I’m trying to simplify, the point simply being; I’d still rather go last)

193

Mijin, your hypothetical isn’t the same game dynamics as ebay.

To make it a comparable study, both candidates need to submit hidden lowest bids.

It is not your visible bid against his visible bid.
It is your invisible proxy bid against his invisible proxy bid.

The interviewer then compares both invisible proxy bids and determines the winner.

The invisible proxy aspect makes the concept of “going last” for a chance to undercut irrelevant. Explain mathematically how there is a sequential aspect of going “last” when hidden proxy bids are involved.

I still say you’re missing a fundamental aspect of the hidden proxy system. You keep inventing examples and analogies based on the competitive amounts you “see” instead of the reality of hidden amounts you “can’t see.” It leads you to believe you have extra information about your own relative position that’s not really there.

194

But your hypothetical has almost no relation to how ebay works. The issues:

  1. The second guy sees what you bid. But in ebay, the guy only sees the current bid, not your upper (or lower in the hypothetical) limit.
  2. You say the second guy has the option of leaving you with your low amount, implying that whatever you bid IS the wage you will get. Obviously this is not how ebay works with automatic incremental bidding.

I suspect that if you played the wage game according to actual ebay rules, you would in fact make a bid that was completely independent of what the other guy bid, for the simple fact that (as multiple people have pointed out) the other guy’s bid gives you no additional information on what you are willing to work at your dream job for.

195

I think I see the problem that Mijin is running into. Let’s keep using L and U, and say that V is the magical (and probably purely theoretical value) that the widget is worth to me, and we can clearly say that L <= V <= U.

The problem arises in saying that we’re willing to pay V, but when we see the auction go for V+1, we’re unhappy that we lost and would have been willing to pay it. Thus, we’ve just contradicted the V is the value of the item, that V is actually at least V+1 in this case and, in fact, L is equal to at least what we thought V was.

You can also work that backwards. We know we’re unwilling to pay U, but what about U-1? If we’re willing to pay U-1, then U-1 is V, if not, then U is actually at most U-1, and possibly less. As such, you’ll actually find that V=U, and if it doesn’t, then your estimate on U isn’t useful. Interestly enough, you’ll find that L is pretty much useless because anything less than V is a deal, but if you’re having trouble working down from U, it might help to work up from L instead.

This is where the store analogy is useful. I walk into a store and see a widget for sale. Obviously, if I see it for anything less than my estimate for V, great, I buy it on the spot, I’m happy because I got a deal. Similarly, if I see it for U+1, I should be happy to walk away because U is the point where I walk away; if not, my U is low and I should be willing to buy it. As such, U=V.
Or, for a more intuitive explanation, the idea of a maximum bid isn’t the point at which you feel like you’re ending up gaining value from the transaction. That idea comes from being used to shopping in stores where I see an item for $13.99 and I buy it, I may actually be willing to pay more, but that’s where it is, so I don’t have to test it. Rather, a maximum bid is the point just below which you break absolutely even simply because if you’re breaking absolutely even, why bother with the effort of the deal when you’re no better off for having done it. That’s you’re upper limit, but it’s also your personal value because, by definition, that upper limit IS your maximum.

Yes, it’s probably hard to narrow down, but I think that’s precisely because we look at the going value and add up. If it makes it easier, you might be better off just starting from your U estimate and working your way down. Are you uncomfortable spending that much for it? Yes, decrease and try again; no, bid, that’s your maximum value. The important thing to keep in mind is that you’re not necessarily paying that much, so you very well may get a deal, you just can’t depend on it.

196

How is the newly revised knowledge that you’d be willing to pay V+2 have any meaning if you don’t even know if the competitor’s hidden max proxy bid is V+100 ?

If one is going to be unhappy, he should at least dignify that suffering with a real reason instead of an invented one.

For example, be unhappy that you don’t know the other guys secret max bid and wish you could wiretap his computer so you could find it out. Or be unhappy that the winning bidder has a cooler sounding username that rolls of the tongue. These are real reasons you can actually glean from ebay.

Don’t be unhappy about losing to a V+1 amount! That number is meaningless!

197

Yes, V+1 is completely determined by your V. The other bidder’s offer could have been V+ whatever, and you can chase that value until you reach your OMG!

Every eBay auctio will be won at one bid increment over the second person’s bid. That is not any indication of what the winner’s maximum bid was, it is just how much eBay charged him to beat your offer.

198

How about if the second person has a couple of shots at it, and gets to see if her bid was insufficient?

I was with you until the last few words.

I’m describing what I believe is a fuzzy logic situation. Below L we are 100% happy to buy. Above U we are 100% happy to walk away. And inbetween is a region of increasing happiness to walk away / decreasing happiness to buy.

Now, an actual decision has to be made, so this model has to be collapsed into a Boolean at some point. But it would be a theoretical mistake to therefore declare that the situation is Boolean and V = (L + 1) = (U - 1)

Not necessarily.
Say I go to eBuyer and find an SSD for $100. I decide to look at a few other websites. I find that eBuyer is the cheapest price, so I go back there and buy it.
Why did I not “buy it on the spot”, since it must clearly be < V?

So like most others here you have concluded that always bidding (the true) U is the best strategy. I don’t believe it is, and like most people, in practice, I would rather be sniped from time to time than frequently pay U.

I think part of the problem is the subjective nature of wanting an item. Posters here are assuming that if I let an item go, it’s because I didn’t want it sufficiently.
But you can want an item plenty, and for strategic reasons walk away.
e.g. delaying buying from eBuyer, even though their price was < V.

199

Ok, but isn’t your argument here (feel free to correct me) that sniping is a bad thing that leads to disappointed bidders who were behaving perfectly rationally?

You had mentioned previously that an alternative system might involve not allowing last second bids (somehow) so that bidders have a chance to respond to any other high bids before the end. To me this says, “I would rather pay L but I’ll increase my bid up to U if I need to.” But if that’s true, what possible reason do you have for not putting U as your one and only bid?

The only possible difference between the two situations is if an outbid somehow gives you more information you can use to determine your amount. And with ebay, the only information is that there is someone out there willing to pay something more than what you typed in. How does that help you determine a new higher bid to make?

200

Not always rationally, but potentially rationally, yes.
The rest of the thread seemed to assume that only rubes can get sniped.

Risk.

Remember, I’m suggesting that U is actually a slight overvaluation. I’m saying that if you were in the process of buying an item at a fair price, and then the price went up by a trivial amount (and you knew it was for a legitimate reason; you’re not being scammed), you may be willing to pay the higher price.
This is partly because you have invested time in the current transaction. Putting a price on your time is not in itself irrational, and typically this price is greater than 1 bid increment.

However, if I were to bid U every time then in the worst case I would pay over the odds in all of my auction transactions.