“Hardly ever” is irrelevant. I’m talking about worst case.
I don’t know how to respond to this because it completely ignores everything I said in the previous post.
Either you haven’t understood my post, or you are unwilling to take it on.
It was a hypothetical set of events in a real auction. It simply isn’t the same to say “just imagine that happened in your head”.
Again, with the real world shopping example, I can walk into the first shop I see and imagine that the price there is the cheapest in all of Londinium and make a decision there and then. It would not make paying U suddenly the best strategy (unless I wanted to do_or_die anyway).
No I’m not.
If I bid $1500, then see that the current highest bid is $1502, I can indeed bid $1504, and if the current highest bid is higher than $1504 then I am immediately and automatically out-bid.
The only thing I “risk” by bidding $1504, apart from the money, is that my bid is insufficient to win – but I was already in that position.
The only difference (in a rationality sense) would be if the actual outbid in the auction gave you more information than the virtual outbid in your head. This would certainly be true if ebay told you the real bid value of your opponent. But instead all you know is that it is more than yours. It should be no trouble to imagine your response to that knowledge.
How about this analogy? I offer you to play a game where you make bids on some product you want. I tell you there’s a card face down on the table that has some amount on it, completely unknown. You are allowed to make as many bids as you want and I will tell you if it is higher or lower than the amount on the card. If it is higher, you pay the card amount and take home the item. Otherwise, you can stop at anytime.
But secretly I’m rigging the game for a psychology experiment. Not matter what you say, I’m going to tell you your bid is low. Then when you finally give up, I’m going to have the card show $1 dollar above your last bid. Assuming you thought I was being honest you’re obviously disappointed, but do you suddenly think you weren’t rational when making bids?
What if after I show you the card I put another card down and say it’s possibly even higher, but again, unknown. Do you see why it makes no sense now to want to just add $2 to your last bid and then call it quits again if it doesn’t win?
I didn’t “ignore” your previous post; what I said is simply based on an insight that has been demonstrated in this thread long ago but which you refuse to incorporate into your posts - namely, that it is in your interest to bid what the item is really worth to you and not any figure above or below that amount.
All you are saying here is that your initial bid of $1500 was not your subjective valuation of the item; in fact, the correct number was $1504 to begin with. You could and should have bid that amount straightaway rather than bidding a lower amount and the upping your bid.
Firstly, your usage of “U” as this extra do-or-die price is really not adding anything to the analysis. You could’ve just called the average-price-you-want-to-pay as your U and nothing would be lost in the conversation. We don’t have to strictly treat U as this exalted number you’d exchange one of your kidneys for but merely as any arbitrary price you would like to pay. This would have saved us a ton of tangent paragraphs about U.
Ok, let us hyperfocus on this one decision here as The Big Moment of Truth.
$1500 Mijin
$1502 sniper
$1504 Mijin
$1506 sniper
$1508? hmmm, I give up and I’ve decided NOT to submit $1508
Somehow, you took inventory of whatever knowledge was in your brain at the time you saw $1506 and made a yes-no decision right there.
The perplexing question is why must your eyeballs actually see the digits $1506 on the screen for that decision? Is it psychological or do you really have new information? You emphasize that your “approximately average” price in mind is not precise enough. And yet somehow, the digits 1-5-0-6 you see on the screen triggers something in your mind to say, “ok I stop.” But, you can see that 1-5-0-6 in your minds eye anytime you want because you have no difference of information between bids of $1500 and $1506. (For that matter, you have no extra information of bids between $0.01 and $1506.) Where did this newfound “precision” suddenly materialize from for you to finally decide “stop”? That $1506 you see is the result of your bidding boundary and your actions and not the competitor’s boundary. To pile on the uncertainty to $1506, the auction isn’t even over yet. This means your confidence with your “winning-at-the-moment” bid at$1500 was misplaced because the auction was not even finished.
Yes, you now have supposedly new knowledge of the intermediate win-loss values of that particular timeframe but the only thing that matters is the final auction outcome.
Here is the scenario where you imagine $1506 in your head and proxy bid $1504:
$1498 – don’t know if any intermediate winning-losing bids will land here, but I don’t care
$1500 – don’t know if any intermediate winning-losing bids will land here, but if it does I might win
$1502 – don’t know if any intermediate winning-losing bids will land here, but if it does I might win
$1504 – don’t know if any intermediate winning-losing bids will land here, but if it does I might win 1506 -- don't know if any intermediate winning-losing bids will land here, , but if it does I'll lose
??? – don’t know what the last value will be 1 second before auction ends or who else will submit a bid and don’t know if I’ll be the winner
final outcome – gee I guess I lost auction
Here is the scenario where you bid $1500 at first and manually increase it:
$1498 – don’t know if any intermediate winning-losing bids will land here
$1500 – I see that I’m the highest bidder at this point
$1502 – I see that I’m losing – let me respond with $1504
$1504 – I see that I’m the highest bidder at this point 1506 -- I see that I'm losing again; I give up
??? – don’t know what the last value will be 1 second before auction ends or who else will submit a bid and don’t know if I’ll be the winner
final outcome – gee I guess I lost auction
Both outcomes are the same. Your supposedly “extra information” of the temporary boolean win-lose adds nothing to your knowledge of what $1504 or $1506 feels like or if it will actually win the auction in the end. Only when the auction truly ends do you know the winner. All the intermediate lead changes are irrelevant.
You mention that you know how to write pseudocode (software programming code?). If so, try to imagine an algorithm with a decision matrix to choose bid-yes-or-no between $1500 and $1506. There is no live minute-by-minute heuristic you can use that only looks at those price amounts as they happen for any artificial intelligence decision. There’s simply no usable information content there.
It was necessary here as people kept saying that it is easy to determine how much to bid: you simply work out the value, or utility, of an item to you, and bid that.
The reality is of course that there’s a layer of strategy between that figure and how much we, say, bid in auctions (not to mention that I’m skeptical that such a value is even consciously known).
For the reason I gave after the line you’ve quoted. Those events alter my behaviour and so imagining that they have happened I may not do what is strategically best.
Again, if I walk into a shop and see that they are selling screws at U, I could choose to imagine that this is the cheapest price in town, and I either buy here or do without screws forever more.
However, it may actually be the case that the store next door is selling them cheaper, and my imagination has led me astray.
You’ve both done the same thing here which is interpret my example as saying that $1506 is my real maximum. That’s not what the example was about.
Basically if I’ve bid $1500 and I’m beaten by a bid of $1502, I want to see whether I’ve been outbid by a genuinely higher value. To the level of precision that I know the average price and what I’m willing to pay, $1502 is equal to $1500. In the bidsniping case, I don’t get to see whether I’ve been beaten by a genuinely higher price or an approximately the same price.
Now you may say “But you have to draw the line somewhere!”, and of course one does.
The point simply being that someone acting rationally, with a rational strategy, may be disappointed to see that they got outbid by a minute amount in the last second. “If only I’d picked a strategy of 80% of U instead of 75% U” – which is of course the same disappointment one may have in the non-sniping case, though you have time to think about and potentially change strategies in the normal case.
Or, “If only I’d got one final bid in to see if the high bid genuinely was higher or if the other guy just typed in the minimum figure”.
So, it is because there is no emotional element to a thought experiment.
Ebay is not a retail store, and the price an item sells for on ebay is set by the second-highest bidder at one bid increment over that bidder’s offer. The situations are not remotely equivalent.
We are asking you to envision (or imagine) your reaction to another purchaser offering more than you did, and to place your initial bid as if you have already been outbidded by a trivial amount.
Why? Personal satisfaction? The need to know that the winner wanted it a LOT morre than you did or just enough more to win that item?
… missed edit window – insert this into the part about eBay not being a retail store –
Also, it is trivially easy to find what similar widgets have sold for on eBay over the past month or so, so you can tell if there is a good chance you will get an item at your target price or whether you can expect others to outbid you.
You’ve used analogy of real in-store shopping several times. We have to adjust the scenario to more closely match the dynamics of ebay.
You walk into a shop and see an item you like for $1500.
While you look at it, the owner stops by and crosses out $1500 and writes $1502.
You think to yourself, $2 increase is no big deal and walk up to the owner telling him you want to buy it at that price.
The owner then tells you that price is just a “maybe” price. He won’t know for sure until 9:00pm closing time. You incredulously ask why he has to wait until 9pm. He says he doesn’t know what price he might fetch at 9:00pm.
You ask if anyone else will come in that might pay more. The owner doesn’t know.
You then ask why he even bother writing $1502 on the price tag if that amount is undetermined until 9pm?
To psyche you out, that’s why.
You can further embellish the in-store analogy with some kind of proxy scenario such as pacing around the store to see what any new walk in cusomters might do or you can tell the owner your final price and go home. He can then call you at 9pm to notify that your final price was acceptable.
You explain to the owner that you’d love to be notified if any walk in customer is willing to pay more than $1502. The owner is willing to accommodate your request but only until 8:59:59 … keep adding to this analogy blah blah blah
The key is that you’re dealing with a bunch of “maybe” prices.
I’m sensing you have some indescribable feeling if you “know” you were beat by a “genuine” price. You’re not satisfied with historical prices or amounts you see at other places such as amazon marketplace or craigslist. It has to be a price that’s related to the auction you’re in and it has to be a real competitor’s number. It feels more satisfying to have some type of “real” situation actually happen instead of an imaginary one, yes?
But I say you’re still exercising your imagination – it’s just at a different point in time. What if you were able to find out that the sniper that beat you sent an email to the seller apologizing, “ooops, I’m really sorry but I made a mistake by bidding $1506. It was a typo and I really meant to enter $156 into my sniping software” As a result, the sniper and seller never complete the transaction and no money changes hands.
In light of that, is that $1506 “genuinely” higher? How will you ever know? All you can do is imagine that all those winners actually followed through on their commitments to buy and sent that money you “see” to the seller. You’re right back to setting a price based on what you think will happen in the real world. You’re not really sure what happens as the auction proceeds. You’re not really sure in the aftermath of the auction either. All you’ve got for sure is whatever comfortable price level in your head. This is the same information you started with.
As mentioned before, you don’t have enough information at the end of the auction to use describe it as “outbid by a minute amount.” That’s a false description. Therefore, any resulting disappointment is also false.
Because it’s worth seeing if the highest bid is only 1 bid increment above me, or significantly above me. In the former case, I’ll now be the winning bid again.
And it’s relevant from a rational-buyer’s POV because the only discrete maximum is U. Since people don’t typically bid U, they are bidding an arbitrary figure, usually a round number, somewhere below U. But there is no qualitative difference between the amount they are bidding and an amount +/- one bid increment, say.
You didn’t provide the “answer” to your analogy.
If I’ve left my house intending to pay up to $15 for a set of headphones, say, and I travel around and find the lowest price is $15.02, I have two options.
I can treat $15 as an absolute maximum, and go home.
But, $15 is not really an absolute maximum. It’s probably far below U. So my other option is to declare that $15.02 is approximately the same as my maximum – certainly to within the precision that I set such a figure – and buy.
This is the kind of buying decision people make every day, and sniping (among other things) impacts upon this behaviour. It is not irrational, though it is “fuzzy”*.
Of course they don’t know that: that’s my point.
In the normal case, I can go over by one bid increment and see if someone came in a negligable amount over me (in which case I’m now winning again), but I’m denied that option in the sniper case.
I think part of the problem with people following the abstract logic of valuation and buying is that people are thinking first of situations where they have considerable market data.
For instance, if I were to really buy screws, my maximum price would be very low indeed as I know industrial societies can mass produce screws very cheaply.
But, by analogy, my past experience of market data is like “earlier in my shopping trip”. Of course I’ll put a hard limit of $2 for a box of screws when I’ve seen them for that price previously.
That “or” clause is not correct. You can’t determine that. You keep typing out statements that contradict how ebay’s algorithm actually works. Are you absolutely sure you understand the hidden max proxy system?
..and if you’re not, you still won’t know whether it’s “trivial” or “significant”.
Yes, I understand the fuzzy valuation concept. Let’s stipulate that it’s “not irrational” to feel that way. What’s irrational is that you require real competitive bidders to help you eventually say “ok I stop manually incrementing my bid” way under “U” to feel good about a deal.
Those competitive bidders may have submitted a high bid with a typo. They may not follow through on the transaction and stiff the seller. Maybe the other bidder that submitted $5000 thinks he’s going to win a “gold watch” but didn’t read the ad carefully enough to see that it’s actually “gold plated watch.” You keep calling these “genuinely” higher prices. You’re adjusting your behavior against a universe of possibilities you can’t ever know. The only thing “genuine” is your own valuation in you head. And that valuation may be fuzzy which is* not irrational*.
You’re manually adjusting your bid increments as a reaction to:
ebay’s escalation algorithm
stupid bidders
deadbeat bidders
careless fatfinger bidders
maybe a shill bidder
possibly some reasonable bidders
Tweaking your bids against all these unknowns is rational?
Your store analogy is fundamentally different because you actually walk of out of the store with the item in hand once you strike an acceptable deal. There are no open ambiguous possibilities – the item is in your hand. That’s not ebay.
This background data is very straightforward and easily understood but it’s not relevant to an optimal strategy for reacting to the various timing of other bidders.
Yes, I can know whether the bid is a little over me, by attempting to bid over it.
A coloured font doesn’t change that.
This is a spurious point. Of course other people’s bidding affects my options, which is partly why bid cancellation is not something taken lightly under eBay’s terms.
The most trivial example of this is if the current high bid is way above what I’m willing to pay. Even if I could be bothered to enter, to try to bid my max amount (on the off-chance that there are “accidental” bids above me), the eBay interface does not even allow me to do so.
I was explicitly pointing it out because your statements are not stated in terms of mathematical precision and rigorous logic.
You said: “it’s worth seeing if the highest bid is only 1 bid increment above me, or significantly above me.”
The correct statement is: “it’s worth seeing if the highest bid is only 1 bid increment above me, or an unknown amount above me.”
There is no “of course” about it! That’s one source of your irrationality! You’ve got a whole army of people in this thread saying the opposite: “of course other people’s bids DO NOT AFFECT my decision strategy.”
Hey, it’s a free country and you’re certainly within your rights to believe that other people’s bidding affects your determination of a stopping point but that’s the very behavior that everyone is labeling as irrational.
Maybe it all boils down to the negative connotation of “irrational” that’s bothering you. If you want to use the word “rational” to simply to describe “what a lot of other people typically do”, and that’s what you also do to feel comfortable then just say so.
However, the other definition of “rational” is determining an optimal value (even fuzzy on) based on specific real information. Intermediate winning-losing statuses before the auction finally finishes is spurious information. Competing bids that you label as “genuinely” higher is spurious information since you’ll never really know what truly happened after the auction was done. As a seller, I’ve actually had more deadbeats than bid cancellations. I pity anyone who manually increased their bids or felt disappointed after losing because the deadbeats’ bids weren’t even “real”. The only thing a rational bidder can feel good about is sticking to their guns regardless of competitors – the number in their head is the only reliable information for decision making they can count on.
…which doesn’t matter because you should have your “fuzzy” value in your head and can run through the simulations to converge on a cut off value.
Let me ask you this: what if ebay had an extra “simulations” webpage that took the current auction item you were interested in and had robots “submit” bids against your pretend bids? You can rerun the simulations as many times as you like. Would that help you find a cutoff value?
This thread is starting to get a little ridiculous. An earlier post I made has what I think is the most straightforward rebuttal of this idea of “I want to have the chance of adding just a few dollars to see if their bid is really close to mine” as being rational. It’s all I’m really interested in seeing a response to.
Yes, that is what I mean. It’s debatable who’s using the more rigorous language however.
Since “an unknown amount” could mean zero, 1 bid increment, or, technically, less than 1 bid increment among other values.
Whereas I’m explicitly saying greater than the insignificant amount. Sure, “insignificant” is a subjective term, but I’ve defined it tightly enough for our discussion here.
<slow claps>
What was my very next line after the line you’re quoting?
You were talking about how other people’s bids – which may be in error – should not affect my bidding.
I pointed out, in that line that you’ve decided to ignore, that “error bids” prior to me entering an auction, not only affect my decisions but may prevent me even entering.
Ergo, as a practical matter you are simply incorrect.
No, we’ve gone through this lots of times now.
There is no discrete cutoff until we get to U. As a buying strategy, I don’t wish to come in at U.
Now you may say: won’t I go up to U incrementally, in the simulation?
No: after being outbid a couple of times my expectation that I am close to the true high bid is low and I quit.
Of course this analogy is different enough to an auction situation to be meaningless.
There’s a static price (at least the participant believes that there is), no other participants, and the two cards are very misleadingly representing a single bid in the closing lines of the description.
I’ve also got a U similar to yours. Actually, I’ve got a bunch of them:
U[sub]1[/sub]: upper bound max based on zero effort in self introspection, studying game theory, or researching any previous market prices
U[sub]2[/sub]: upper bound max if I shake the couch cushions for loose coins and give up one Starbucks cappuccino and apply the extra funds towards the auction
U[sub]3[/sub]: upper bound max before wife notices the missing funds and we have an argument about it
U[sub]4[/sub]: upper bound if I buy it, use it for a short time, and resell it on ebay all before the credit card bill comes due. If the credit-card limit is $10k, then U4 is $10k. (This boundary is a little tricky because it assumes you can resell it for the same U4 price that you bought it for.)
U[sub]5[/sub]: upper bound if I win the lottery between now and the end of the auction. I wouldn’t even want to pay this even though I was a millionaire just on the principal of not paying a stupid price.
Your idea of “U” and all my various thresholds of U and everyone else’s pain points of U are irrelevant. It adds nothing to the analysis. I guess you believe this additional information about “U” will cause everyone to reclassify your behavior as “rational” but it doesn’t work that way.
All that matters is that one single decision where you decide to “stop.” You don’t need U for that because you don’t even want to pay U. You have an adhoc decision at price S. You hyperfocus on S and that is what you work backwards from.
Having any expectation that you’re close or not close to the true high bid is irrational. The simulation and the real ebay auctions don’t provide enough information for that feeling.
The point with U is that U is the only real threshold. There is a qualitative difference between U and U+1; in the latter case, I’m actually losing utility from the deal.
While debaters here assumed that the max price someone is willing to pay was U, it made sense that there was a point where a rational agent must definitely stop and not go a dime further.
I’m glad that now dopers have moved to my position (which is now being declared “obvious”) that people in the real world seldom bid U.
So…they bid S. The difference with S though is that it is only a hard threshold if you choose to treat it that way. There is no qualitative difference between S-1, S and S+1.
As I mentioned before, if I decide I’ll pay up to $15 for headphones, if the cheapest price I find is $15.02 I may say “No way, not a dime more than S!” or I may say “Meh, close enough”. Both decisions can be defended.
Wrong. Like most people in this thread you aren’t thinking this through, you’re sure you have the answer before the question is even asked.
If we imagine the two scenarios: I bid $4 and I’m outbid by what ostensibly is 1 bid increment. So I bid over by 1 bid increment but I’m immediately auto-bid out by 1 increment again. I repeat this pattern all the way to $4.80, where I am again auto-outbid by $4.82.
In the second scenario, I join an auction and bid the starting price of $4.80, which has held up for a while, but now I see I’m outbid by $4.82.
Now, you’re saying that the situation is effectively the same in the two scenarios: I have no idea what the max bid is, and it is irrational to think otherwise.
I’m saying that in scenario 2 there is a higher probability that the max bid is $4.82. That’s because in scenario 1 there is nothing significant about the price $4.82…I have no reason to suppose someone would bid that amount.
In scenario 2 however, $4.82 has significance: it is the amount that someone who simply typed in the minimum would bid. I suspect that a proportion of eBayers routinely do this, making the probability greater than in scenario 1.
And the “meh close enough” can be determined ahead of time without messing around with manually increasing bids. The “meh close enough” numbers already exist as discrete points on the number line before the auction and starts and finishes. The need to clarify “meh close enough” with real bidders is not necessary and irrational.
Are you saying you’re trying to see if your last bid happens to fall within the next pre-defined bid increment?
First, if that’s your “strategy”, then you still could have done that ahead of time by calculating S_my_comfortable_price+1_bid_increment. That $4.82 with significance could have had significance without a real person entering it. All your “rational” valuation can be done ahead of time by looking at the ebay increments chart. Bid once, bid late, bid your “max” – you’re done.
Second, rational bidders don’t rely on this behavior because the auction isn’t even finished. That’s just a meaningless transient “lead change.” Will that bidder that entered $4.82 add another bid that’s $5.99? What if a new bidder lurking in the shadows submit a bid at the last second for $10? What if the bidder who entered $4.82 doesn’t actually buy it from the seller? The fact that you’re specifically seeking these bid-increment numerical relationships and adjusting your micro behavior to it is irrational.
To summarize, it is understandable and reasonable to have “fuzzy” valuation numbers. It is irrational to clarify that fuzziness with real bidders’ amounts in an unfinished auction. If the valuation in your head (even if it’s fuzzy) is less reliable than the various numbers being thrown at you buy a live auction, that’s irrational.
I think the same points have been made, remade, rephrased, repeated, reiterated, rehashed, and restated for the last 100 or more posts. We’re long past the point where anybody’s mind is going to be changed.
