Set a ceiling once and stop watching the auction

The expensive part of bidding on anything is not the money. It is being awake. A slot lost at three in the morning is lost because nobody was watching, not because nobody was willing to pay — and watching a board for twelve hours is a worse deal than almost any price you would have paid. Outbid Guard is the fix: one ceiling, stated in advance, and the board acting on it while you do something else.

How it works

Set a maximum for one of your brands. From then on, whenever somebody outbids that brand on a slot it is holding, the board works out what taking the slot back would cost and does it — provided the cost is at or under your ceiling.

If taking it back costs a cent more than your ceiling, nothing happens and you get an email saying so, with the number. That is not a compromise; a guard that crept over its limit "because it was close" would be worse than no guard, because you would have to check it, which is the thing you paid to stop doing.

The ceiling is checked twice: once when the defeat is noticed, and again inside the database lock that actually places the bid. Between those two moments the price can move — that is what an auction is — and only the second check is binding.

What it spends, and what it does not

Outbid Guard spends house credit. It cannot charge a card.

This is a real limit and worth being blunt about rather than discovering later. Charging a stored card while its owner is asleep requires an off-session payment mandate — a saved payment method, explicit prior consent to charge without you present, and a path for the bank to challenge it. Stripe documents what that involves under saving cards for future payments, and none of it is optional. This board does not hold one, and building it badly is the kind of mistake that shows up as a chargeback rather than a bug report.

So a guard defends your slot as long as your credit balance covers the next bid. Every account starts with house credit, and a slot successfully defended costs the same as one bought the normal way. When the credit runs out, the guard does not fail silently — it emails you the reason and the amount it would have needed.

Card-funded guards are the obvious next step and are not shipped. When they are, this article will say so.

Everything it declines to do, it says out loud

There are four ways a guard can decline, and all four send you an email naming which:

• the price went above your ceiling
• your credit did not cover it
• the slot had already closed
• it had already made five attempts on that one slot

The last one is a circuit breaker. Two people can point guards at the same slot, and in theory that terminates on its own — every bid has to beat the last one, and both ceilings are finite, so the escalation runs out of room. Five attempts is the limit for the case where that reasoning turns out to be wrong at three in the morning, which is exactly the hour this feature exists for.

The reason all four are emailed rather than logged is simple: a guard that quietly did nothing would be indistinguishable from a guard that was broken, and you would end up checking the board to make sure — which is the whole thing you were trying to avoid.

What it does not turn the board into

A guard is not a bidding robot and there is no strategy in it. It does not bluff, it does not bid early to establish a position, it does not read the other side. It answers exactly one question — "somebody took your slot, do you want it back at this price?" — with the answer you gave it in advance.

That matters for what the board is. A bid ceiling makes an auction fairer because it stops escalation from being a test of who is paying least attention; a personal ceiling does the same thing one bidder at a time. The person who wins a contested slot should be the one who valued it most, not the one who happened to be at a keyboard.

Set it before you need it

The mechanism only helps if the ceiling exists before the defeat, which means the moment to set it is when you place the bid, not when you lose. Pick the number the honest way: work out what a click is worth to you, multiply by the click-throughs a slot in your category has actually delivered, and set the ceiling somewhere below that. The price index shows what slots have been settling at, which is the other half of the arithmetic.

Then stop watching. That was the point.

The honest summary

One ceiling. Credit-funded. Never exceeded, never silent, capped at five attempts per slot. It removes the 3am problem and does not pretend to remove anything else — and when it declines, it tells you why before you find out from the board.