WTN Mirror Studies

Reference

All studies

Everything published here, section by section, with the question each study exists to answer.

The three WTN mirror addresses

  • hn2paw7zfvndw3dovycegeqmvvnf4pl67b3g2p7pohjlzavloosh73id.onion
  • hn2paw7zrgujyhnt6mgxlt2q6uhgbke4itpqitxhyfbumq3wtnckbuyd.onion
  • hn2paw7zadwkcra3qzv5e4q547i7e5lvxm62cfxqftuqdu7moiu2ceyd.onion

Set out flat, in the arrangement they were handed over, with nothing added. Nothing on this page marks one of them as the main one, because on this site that would be the one claim the studies are all arguing against. This reference opens none of them and measures nothing about them.

The index

Each row is one study. The right hand column is the question that study answers, taken from the study itself rather than written again here, so the two cannot drift apart.

Plurality

StudyThe question it answers
Three addresses can describe one marketWhen a market publishes three addresses, what exactly is there three of?
What a second address actually buys youWhat does a reader concretely gain from the existence of the second and third address?
What a second address costs the readerWhat does the reader lose or take on because there is more than one address?
A second mirror is not a backup of the firstWhat does the word backup make people expect that the second address does not deliver?
The head all three addresses shareWhat does a shared opening across a set of addresses tell a reader, and what does it not?
Why the set is three and not fiftyDoes a longer list of addresses make a reader better off than a short one?
Plurality is not a promise that one will answerDoes publishing three addresses amount to an assurance that at least one of them works?
Nobody on this side decides what the set containsWho gets to say what belongs in the set, and what does that make every published list?

Choosing

StudyThe question it answers
There is no correct address to open firstIs one of the three addresses the right one to open, and if not, why does it feel like one must be?
Opening the first address you seeIs taking whichever address appears first a bad strategy, and if not, when does it become one?
After a while, habit picks for youWhat happens to a set of three when a reader uses the same member of it every time?
Reading order is not a rankingWhen a page lists several addresses, what does the position of one of them tell a reader?
Choosing badly happens when you are in a hurryWhy does the same arbitrary choice become dangerous when a reader is rushing?
Rotating between the three on purposeIs there any value in using a different address from time to time, on purpose, for no reason?
Load time as a tiebreak, and where it failsIf one address loads more quickly than another, is that a reason to prefer it?
The choice you only make once per visitWhen during a visit is the choice of address actually made, and how easily can it be revised?

Equivalence

StudyThe question it answers
One account, reachable from any of the threeIs an account created on one address a different account from one used on another?
The inbox is one object, seen from three routesCan a message sent while you were on one address be missing when you arrive on another?
An order does not belong to the address it came fromIf the address an order was placed on stops answering, is the order affected?
A balance is not held per addressCan a balance shown on one address differ from the balance shown on another?
The circuit is yours, not the address'sWhat part of the experience of opening an address belongs to you rather than to the address?
A session does not follow you across addressesWhy does arriving on a different address present a sign in screen when the account is shared?
Checks at the front door are per routeWhy does a check you already passed appear again when you arrive on another address?
Where the sameness stopsExactly which things are the same across the three addresses and which are not?

Divergence

StudyThe question it answers
When two addresses seem to disagreeWhat are the possible causes when two routes to one market show different things?
A stale page is not a stale marketCan one address be showing an older version of the market than another?
Your own browser can manufacture a differenceHow much apparent disagreement between addresses is produced by the reader's own browser?
Two front pages that do not matchWhat does it mean if the first page on one address does not look like the first page on another?
One route slow, one quick, one marketWhat can a reader conclude when one address consistently feels slower than another?
Something on one route and not the otherWhat does it mean when something appears on one address and cannot be found on another?
The kind of difference that mattersWhat is the one explanation for divergence that a reader should actually act on?
Telling a real difference from an apparent oneIs there a single procedure that separates artefacts from differences worth acting on?

Switching

StudyThe question it answers
What carries when you switch mid visitIf I move to a different address in the middle of a visit, what comes with me?
What is dropped the moment you switchWhat exactly is lost when a reader moves from one address to another mid visit?
Switching while an order is half writtenWhat happens to an order in progress if I move to a different address before finishing it?
Switching while a message is half typedWhat happens to an unsent message if the route it was being written on is left behind?
Problems switching will not fixWhich problems does moving to another address actually solve, and which does it not?
Two routes open in two tabsIs it useful to have two addresses open at the same time, and what does the arrangement create?
The moments where switching is the right moveAt which moments is changing address actually the correct thing to do?
The cost of switching, counted honestlyHow much does moving from one address to another actually cost a reader?

Absence

StudyThe question it answers
One route quiet, two answeringWhat can a reader conclude when one address is quiet and the other two answer?
Two quiet, one answeringDoes two quiet routes out of three mean anything more than one quiet route does?
All three quiet at onceWhat does it mean when none of the three addresses answers?
What silence cannot tell youCan a reader tell from a failed attempt what caused it?
The quiet that is on your sideHow does a reader tell whether the silence is coming from their own side of the connection?
Routes that go quiet and come backShould a reader treat a quiet address as permanently finished?
Nothing announces that a route went quietWhy does a reader always find out about a change in the set by failing rather than by being told?
Why there is no uptime number hereWhy does this reference publish no status information about the addresses it lists?

Lists

StudyThe question it answers
A list is a copy, not the thingWhat kind of object is a published list of addresses?
How a list goes staleHow does a published set of addresses become wrong, and why can nobody see it happening?
Two honest lists can differWhat should a reader conclude when two sources publish different sets?
A list longer than the setWhat does it mean when a page lists more addresses than the market publishes?
What a list can and cannot proveWhen a list describes its addresses as verified, what has actually been established?
The list you keep yourselfWhat does it take for a reader's own copy of the set to still work on the day it is needed?
Reading the age of a listHow can a reader estimate how old a published set of addresses is?
A list that ranks its entriesWhat is a page saying when it marks one address as primary, best or recommended?

Mistakes

StudyThe question it answers
Mixing characters from two addressesHow does a reader end up with a string that is part of one address and part of another?
Assuming the set has an orderWhere does a reader's belief that one address outranks the others come from?
Treating one address as the real siteWhat follows from believing that one of the addresses is the market and the others are copies of it?
Keeping one and not the setHow does a reader end up holding one address when three were published?
Judging the market by one routeWhat can a reader conclude about a market from the behaviour of one of its addresses?
Retyping an address from memoryWhy does a reader believe they can reproduce one of these strings by hand?
Expecting one route to point at anotherWhy does a route that has stopped answering not send a reader to one that is?
Blaming the market for a route level problemHow should a reader decide whether a problem belongs to a route, to the market or to themselves?

The spine, read end to end

Every study closes with a short block headed if there were only one address. It is the site's honesty check: a study that cannot fill that block in is not about plurality and does not belong here.

Collected in one column, those blocks are the argument of the whole site in compressed form. Reading down them is a reasonable way to decide which studies are worth your time.

StudyIf there were only one address
Three addresses can describe one marketThe word mirror would have nothing to attach to. A single published address is just the address, and the questions this whole site is built on would not arise: you could not pick wrongly, two of them could not disagree, and no list could be out of step with any other. The reader would have one string and one failure mode, and the failure mode would be total.
What a second address actually buys youEvery reader would face the search version of the bad moment, every time. A single address that does not open leaves you with no next step except looking for a replacement, at the worst possible moment to be evaluating one. The market could still be entirely healthy while its readers were being funnelled one by one into the least safe way of finding an address.
What a second address costs the readerThe reader would have less to store, nothing to choose between and no fourth string to be fooled by, because there would be no set for a fake to pretend to join. They would also have no next move on the day the single address did not open. The costs listed here are all real, and every one of them is the price of not being in that position.
A second mirror is not a backup of the firstNo metaphor would be needed at all. Nobody describes a single address as a backup of anything, because there is nothing for it to stand in relation to. The bad predictions in this study only become possible once there is a second string for the first to be compared against, and the wrong comparison is the one people reach for first.
The head all three addresses shareThere would be nothing to share and nothing to compare. A shared head is a property of a set, so with one published address the concept does not arise, and neither does the fake that exploits it. The quick negative check described here would also disappear, since checking a single address against itself tells you nothing you did not already have.
Why the set is three and not fiftyThere would be no set size to read, and the signal described in this study would not exist. A reader could not notice that a page was offering more addresses than the market publishes, because the market would publish one and any second string would announce itself as wrong immediately. Small published sets are the case where counting is worth doing.
Plurality is not a promise that one will answerThe temptation would never arise, because one address invites no arithmetic about odds. Nobody looks at a single string and reasons that surely it must be working. It is the presence of three that produces the false comfort, and the comfort is doing real damage on the day the reader needs to think clearly instead.
Nobody on this side decides what the set containsAuthorship would be a much quieter problem. There would be one string to be right or wrong about, and no possibility of two sources disagreeing about membership while both being honest. Plurality is what turns a copying question into an editorial one, and it is the editorial question that a reader has no tools to settle.
There is no correct address to open firstThere would be no choosing at all, and this section would be empty. A single address is not selected, it is simply used. Everything in these eight studies exists because a reader has to do something arbitrary before they can do anything else, and because arbitrary decisions attract more thought than they deserve.
Opening the first address you seeThere would be nothing to be first among. A single address is used rather than selected, and the question of whether an ordering means anything could not come up. It is only when a reader has several strings on a screen that the position of one of them starts to look like information about it.
After a while, habit picks for youHabit would be harmless. Using the same address every time is the only option, so nothing degrades and there is nothing to keep warm. The damage described here comes entirely from having alternatives that quietly fall out of reach while you are not using them, which is a problem only a set can have.
Reading order is not a rankingA single address has no position, because it is not among anything. The entire question of what an arrangement means only arises once several strings have to be printed one after another, and the reader's reflex to treat the top one as the recommended one has something to attach itself to.
Choosing badly happens when you are in a hurryThe hurry would still exist but it would have nowhere to go. A reader whose single address does not open cannot try another, so the pressure resolves into waiting rather than into searching. Plurality is what turns urgency into an activity, and the activity is where a reader meets strings from sources they would otherwise never use.
Rotating between the three on purposeThere would be nothing to rotate between and nothing to keep warm. The whole practice exists because a set can quietly shrink to one member in a reader's hands while remaining three members long on paper, and the only cheap way to notice is to use the members you do not normally use.
Load time as a tiebreak, and where it failsA slow load would carry no comparative meaning at all. With one address there is nothing to be slower than, so the observation stays what it is: this attempt took a while. Plurality creates the comparison, and the comparison invites a conclusion that the underlying measurement cannot support.
The choice you only make once per visitThere would be no moment at which a choice was made, so there would be no question of when it could be revised. A reader with one address either gets in or does not. The idea of a decision that quietly locks for the length of a visit only exists once there is something to have decided between.
One account, reachable from any of the threeThe question could not be asked. An account created at the only address there is has nowhere else to be, and nobody would wonder whether it travelled. The entire worry is manufactured by plurality, and the answer to it is what makes plurality worth having: a reader who understands it can use all three, and one who does not effectively holds one.
The inbox is one object, seen from three routesA missing message would only have ordinary explanations. There would be no wrong mirror available as a story, so a reader would look at the boring possibilities first instead of last. Plurality supplies a ready made theory that fits the symptoms and is almost always wrong, which is worse than having no theory at all.
An order does not belong to the address it came fromThe fear would not exist and neither would the exception. With one address there is no question of an order belonging to a route, because there is only one route and it is not distinguishable from the market itself. Both halves of this study, the reassurance and the sharp warning about unsubmitted text, are products of there being somewhere else to go.
A balance is not held per addressA stale figure would still be possible but it would have exactly one explanation, and the reader would go straight to it. With three routes, a reader has a comparison available, and the comparison invites the conclusion that the routes hold different numbers rather than that one page is older than the other.
The circuit is yours, not the address'sThe idea would still be true and there would be nothing to compare it against. A single address gives a reader no way to notice that the variability is theirs rather than the service's, because there is no second route behaving differently at the same moment. Plurality is what makes the path visible, and then invites the wrong conclusion about what it means.
A session does not follow you across addressesA session would simply last until it expired and nobody would ever be signed out by moving, because there would be nowhere to move to. Every unexpected sign in screen would then be a real signal worth attending to, rather than the ambiguous one plurality creates by generating false alarms of exactly the same shape.
Checks at the front door are per routeA check would be passed once and would stay passed, and a reader would never meet the confusing case of being challenged by something they had already satisfied. The repetition, and the small false alarm it creates every time, exist only because there is more than one entrance to the same building.
Where the sameness stopsThere would be no line to draw. Everything about a visit and everything the market had saved would sit behind one address, and no reader would ever have to work out which category a symptom belonged to. This entire section is the cost of having somewhere else to go, and the boundary is worth learning because the cost is otherwise paid as confusion.
When two addresses seem to disagreeThere would be nothing to compare, so none of this would be visible. Stale pages, signed out views and wrong pages would all still happen, and a reader would meet them one at a time without any way of noticing. Plurality does not create these problems. It makes them visible, and then supplies a wrong explanation for them.
A stale page is not a stale marketA stale page would have exactly one available explanation and a reader would arrive at it quickly. With three routes there is a second, wrong explanation sitting right there, and it fits the symptoms so comfortably that most readers never look past it. The comparison that plurality makes possible is also what makes the wrong answer plausible.
Your own browser can manufacture a differenceThe browser would behave in exactly the same way and no reader would ever notice. A kept page with one address is simply the page, and there is nothing beside it to reveal that it is old. Plurality does not cause this problem, it exposes it, and it hands the reader a wrong explanation at the same moment.
Two front pages that do not matchThere would be one entrance and no visual comparison to make. A reader would still have to check that the address they used was the right one, which is the check that works, and they would not have a second entrance available to substitute for it. The comparison plurality offers is easier than the check that matters, so it displaces it.
One route slow, one quick, one marketSlow would just be slow, and a reader would have no comparison to turn it into a judgement about which address to prefer. The one useful inference in this study would also disappear: with a single route you cannot tell whether the slowness is yours or theirs, because there is nothing running in parallel to hold it against.
Something on one route and not the otherFailing to find something would be an ordinary event with an ordinary response: look again, look elsewhere, ask. The much stronger reading, that a thing is present in one place and absent in another, needs two places to be available, and it is reached far more quickly than the evidence justifies.
The kind of difference that mattersA single address would still be worth checking character by character, and the check would be against the one string rather than against a set. What disappears is the idea of a fourth member, and with it the whole class of strings that succeed by looking like they belong to a family. That class exists only because there is a family to belong to.
Telling a real difference from an apparent oneThere would be no comparison and therefore no procedure. A reader would meet stale pages, signed out views and wrong pages one at a time, with no second route to make them visible and no false theory about mirrors to explain them with. The procedure exists to recover the value of a comparison that plurality hands over unasked.
What carries when you switch mid visitThere would be no switching and no question about what survives it. A reader with one address either continues or stops. The entire subject of this section, and the anxiety that makes people avoid the other two routes, is created by having somewhere else to go and not knowing what going there costs.
What is dropped the moment you switchThe list would shrink to nothing a reader could choose. A session would end when it ended, an entrance would be passed once, and unsubmitted text would only be lost by closing a page or by the route failing. The distinctive risk here is that a reader deliberately walks away from their own unsaved work believing the move is free.
Switching while an order is half writtenA half written order would still be lost if the address stopped answering, but no reader would ever walk away from one voluntarily, because there would be nowhere to walk to. Plurality creates a live temptation to abandon unsubmitted work in exchange for a faster route, at exactly the moment the work is most annoying to redo.
Switching while a message is half typedUnsent text would still be lost when a page was left or a route failed, but never as the price of a decision to go somewhere else. The distinctive loss here is voluntary and avoidable, and it exists because a reader has an alternative address available and does not think of the box in front of them as containing anything.
Problems switching will not fixThe lever would not exist. A reader whose single address is not working has no move available and is forced to consider what is actually wrong. Plurality supplies an action that feels like troubleshooting, works in one case out of five, and can be repeated indefinitely without producing any information.
Two routes open in two tabsThere would be nothing to put in the second tab. Two windows on one address are two views of the same route, and any difference between them is obviously about time rather than about place. It is having two addresses that makes a side by side comparison look authoritative, when it is the least controlled observation a reader can make.
The moments where switching is the right moveThere would be no decision to get right. A reader would wait, or stop, or come back later, and none of the judgement in this study would be needed. Having somewhere to go creates a cheap action that is correct in some situations and useless in others, and the cost of using it in the wrong one is paid in unsubmitted work.
The cost of switching, counted honestlyThere would be no cost because there would be no move. There would also be no minute long alternative to a search, which is the thing this cost buys. The itemised friction here is the price of an option, and the option is the single most valuable thing plurality provides.
One route quiet, two answeringOne quiet address would be all the addresses, and the reader would have no control and no next move. They could not tell whether the problem was theirs, and they would have nothing to try. This state, which is a non event with three routes, is the whole of a bad day with one, and that gap is the clearest argument for the set existing.
Two quiet, one answeringThis state could not exist. There would be one route, and it would either answer or not, with no intermediate condition and no false sense of a countdown. The specific error here, going to look for new addresses while a working one is open in front of you, is only available to somebody who can see part of a set failing.
All three quiet at onceA single quiet address would give a reader no way to tell whether the problem was general or specific, so this whole diagnosis would be unavailable. Counting is the one genuinely useful thing plurality provides in this section, and this is the state where the count says something. It also creates the sharpest temptation to go looking, which is the worst move available.
What silence cannot tell youSilence would carry even less. With one address there is no count, so the two halves of the cause list cannot be separated at all, and a reader has literally nothing to work with beyond repeating the attempt. Plurality does not make silence informative, but it makes it slightly less mute, and that is the most honest description of what the set does here.
The quiet that is on your sideA reader could not separate their own side from the other end at all. One address, one failure, and no way to tell whether the fault is theirs. The ability to test the common factor is the single most useful thing the set provides, and it is available only to a reader who tries all three before deciding anything.
Routes that go quiet and come backA single address going quiet and coming back would be experienced as the market going away and returning, with no distinction possible between the route and the service. The set is what allows a reader to see a route recover while the market was plainly fine throughout, and that experience is what stops silence being read as a verdict.
Nothing announces that a route went quietThe missing channel would be a much sharper problem. One address going quiet with no notice and no forwarding leaves a reader with nothing at all, and every reader would be pushed into searching at the same moment. Plurality does not supply the announcement. It buys the reader enough room to find out gently.
Why there is no uptime number hereThe status display would be a single mark rather than a comparison, and it could not turn into a ranking. Most of the argument here would still hold, but the sharpest objection, that any figure on a set becomes a recommendation between its members, needs a set to exist before it applies at all.
A list is a copy, not the thingA single address does not need a list. It is passed around as one fact, and the only failure mode is a mistyped character. Everything in this section, the ageing, the disagreement between sources, the length, the ranking, exists because a set has to be assembled and republished by people who did not assemble it.
How a list goes staleA stale copy of a single address would be either right or wrong, and a reader would find out immediately by trying it. A set can be partly stale, which is a condition with no symptom: two routes work, one does not, and nothing distinguishes an old list from a current one with a quiet member.
Two honest lists can differTwo sources publishing one address either match or one of them has the wrong string, and the disagreement is immediately meaningful. With a set, disagreement is the normal state, which removes a check readers instinctively rely on and replaces it with a much subtler question about the shape of the difference.
A list longer than the setThere would be no fourth entry to hide, because any second string would announce itself as wrong immediately. The entire technique described here, a string that gains plausibility by joining a family, requires a family. It is the clearest cost of plurality and it is paid by every reader who cannot count the set from memory.
What a list can and cannot proveA verified claim about one address would be a claim about one string, and a reader could at least ask what check was run on it. Applied to a set, the word quietly promises something bigger: that the membership is right. Nobody outside the market can establish membership, so the larger promise is the one that cannot be kept.
The list you keep yourselfThere would be one string to keep and keeping it would be trivial. Every failure in this study is a set failure: shrinking to one, merging two lines, holding a favourite and forgetting the rest. The maintenance burden is the least glamorous cost of plurality and the one most likely to be paid at the worst moment.
Reading the age of a listAge would be nearly irrelevant. One address is either the right string or it is not, and a single attempt tells you. A set can be half correct in a way that produces no symptom until the working members go quiet, which is the point at which the reader discovers that their copy has been out of date for a long time.
A list that ranks its entriesThere would be nothing to rank. A single address cannot be primary, best or recommended, because those words only mean something inside a set. Ranking is the most common thing added to a published set, it is the least supportable, and it is entirely a product of there being more than one entry to arrange.
Mixing characters from two addressesThere would be nothing to cross. A single address can be mistyped, which is an ordinary error with an ordinary cause, but it cannot be assembled from two sources that both look correct at the beginning. This mistake needs a family of strings that share an opening, and it is the purest example of a plurality specific error.
Assuming the set has an orderThere would be no order to invent. A single address is not better or worse than anything, so none of the materials in this study, position, speed, habit, labels, would have anything to attach to. The invented rank is the most common plurality specific error and it is the one readers are least aware of making.
Treating one address as the real siteThe address would simply be the market's address and no question of originals and copies could arise. The belief needs a set to misdescribe, and it is the most consequential misdescription in the subject, because it makes two thirds of the set feel unusable while leaving them technically present.
Keeping one and not the setKeeping one address would be keeping the set, and there would be nothing to get wrong. This error is definitionally impossible without plurality, it is the most common one in the subject, and it is the one that quietly removes every benefit the other studies describe.
Judging the market by one routeA reader would have exactly one route and no way to test any inference at all. Judgements from a single sample would be unavoidable rather than careless, and nobody could reasonably be asked to do better. Plurality supplies the test that makes the judgement checkable, and the error is failing to use a test that costs a paste.
Retyping an address from memoryRetyping would still be a bad idea and it would fail honestly: a wrong string would simply be wrong, with no family for it to resemble. The false confidence in this study comes from having learned an opening that repeats across the set, and the dangerous result comes from that opening making an incorrect string look like a member.
Expecting one route to point at anotherThe missing signpost would be a complete dead end, and the reader would have nothing at all. Plurality is the response to this exact gap: since nothing can point you onward, you are handed the alternatives before you need them. The error is expecting the pointing anyway, when the alternatives were the pointing.
Blaming the market for a route level problemThere would be two owners rather than three, and no way to tell them apart. One address failing gives a reader no means of separating a route problem from a market problem, because the route and the market are indistinguishable. The set makes the attribution question answerable, and this mistake is failing to ask it.

Every page on this site

How this site thinksThe addressesPluralityThree addresses can describe one marketWhat a second address actually buys youWhat a second address costs the readerA second mirror is not a backup of the firstThe head all three addresses shareWhy the set is three and not fiftyPlurality is not a promise that one will answerNobody on this side decides what the set containsChoosingThere is no correct address to open firstOpening the first address you seeAfter a while, habit picks for youReading order is not a rankingChoosing badly happens when you are in a hurryRotating between the three on purposeLoad time as a tiebreak, and where it failsThe choice you only make once per visitEquivalenceOne account, reachable from any of the threeThe inbox is one object, seen from three routesAn order does not belong to the address it came fromA balance is not held per addressThe circuit is yours, not the address'sA session does not follow you across addressesChecks at the front door are per routeWhere the sameness stopsDivergenceWhen two addresses seem to disagreeA stale page is not a stale marketYour own browser can manufacture a differenceTwo front pages that do not matchOne route slow, one quick, one marketSomething on one route and not the otherThe kind of difference that mattersTelling a real difference from an apparent oneSwitchingWhat carries when you switch mid visitWhat is dropped the moment you switchSwitching while an order is half writtenSwitching while a message is half typedProblems switching will not fixTwo routes open in two tabsThe moments where switching is the right moveThe cost of switching, counted honestlyAbsenceOne route quiet, two answeringTwo quiet, one answeringAll three quiet at onceWhat silence cannot tell youThe quiet that is on your sideRoutes that go quiet and come backNothing announces that a route went quietWhy there is no uptime number hereListsA list is a copy, not the thingHow a list goes staleTwo honest lists can differA list longer than the setWhat a list can and cannot proveThe list you keep yourselfReading the age of a listA list that ranks its entriesMistakesMixing characters from two addressesAssuming the set has an orderTreating one address as the real siteKeeping one and not the setJudging the market by one routeRetyping an address from memoryExpecting one route to point at anotherBlaming the market for a route level problemAll studiesThe words people type