Between the event and your bet there are six stages, and you own the slow ones

§1Six stages, one of them yours

The chain from an event to an accepted bet runs: the event happens; the official clock or scoring system records it; a data provider publishes it; the operator ingests it and its model reprices the market; the new price reaches you, usually alongside a picture that is older than the price; you decide and submit; the operator applies its bet delay and accepts or rejects. That is six stages, and the first four of them are finished before you have any information at all.

The figure is the argument of this page. It is not that any individual stage is slow; several of them are measured in milliseconds. It is that the whole front of the chain is finished before the back of it starts, and the back of it is you.

§2Stream delay is a design decision, not a fault

The picture you are watching is delayed by seconds to tens of seconds, and that delay is deliberate. Broadcast delays exist to protect rights holders, to allow for compliance and commercial windows, and to give producers a margin to cut away from something they should not have shown. Every legitimate streaming route — an operator's own player, a broadcaster's app, a free stream — carries a delay, and the sizes differ by route rather than by the quality of your connection.

This is the reason that "watching the game" is not the informational advantage it feels like. By the time the picture reaches you the market has seen the event through the provider, priced it, and possibly already moved again. You are not betting on what is happening. You are betting on a recording of something that has already finished happening in the market's world.

§3The round trip: your click and the bet delay

When you submit a live bet, the request and the answer are only part of the elapsed time. The operator's bet delay is an explicit interval, held between receipt and acceptance, and it is applied at the operator's discretion. Its length is a policy choice: long enough to protect against a fast counterparty, short enough that the product still feels responsive.

For you, the delay has three consequences. It makes the price you saw non-binding. It means a bet can be accepted at a materially different price without you having changed your mind. And it means that if you submit a bet while a market is about to suspend, you are asking the operator to decide what to do with it, under rules you may not have read.

The delay is also not the last stage of uncertainty. A bet can be accepted at the price you clicked and settled against a result that a later correction changes. The chain from the event to the settlement is longer than the chain from the event to the price.

§4Latency is never symmetric

It is tempting to think of latency as a shared tax: everybody waits the same amount, so nothing changes. The distribution is not like that at all. The operator sits at the front of the chain by construction: it receives the provider's feed directly and reprices from it. You sit at the back. The gap between those two positions is not a constant you can close by paying more for internet; it is the structural relationship between a market maker and a price taker.

There is a second asymmetry that matters more in practice. A trading desk can choose not to quote during the window when its information is stale. You cannot choose to be told when its information is stale — the price you see looks equally valid whether it is current or a second out of date. The information about which prices are trustworthy is itself held by one side.

§5What the chain costs, in the only currency that matters

The cost of the delay is not abstract. It shows up as slippage: the difference between the price at the moment you decided and the price at which your bet was actually accepted. Cycle after cycle, almost all of that difference runs one direction, and the size of it is a function of the market's volatility rather than of your judgement.

Very few bettors measure their slippage, and the reason is instructive: measuring it requires recording both prices, and the second one is only visible in the bet receipt. Anyone who has done it for a few hundred live bets knows the answer already, which is that the slippage is real, systematically negative, and larger when they were reacting fastest.

§6The parts of the chain you can actually change

• Stop treating the picture as the present. The one adjustment that costs nothing and changes everything. Whatever you are watching, the market has already seen it.
• Leave the reaction window empty. If your reason for a live bet is an event you just saw, the reason is stale. Deciding rules in advance is the only thing that survives the delay.
• Record both prices. Not to improve your results — to see the slippage. Recording the price you saw and the price you got is the fastest way to lose the belief that you are fast.
• Prefer markets where the delay matters less. Long-horizon markets — totals, period winners, markets decided over minutes rather than seconds — are less sensitive to being thirty seconds behind than a next-goal market is.
• Accept the asymmetry and size accordingly. You cannot win a race against the repricing engine, so a stake that only makes sense if you can is a mis-sized stake regardless of how good the reasoning was.