Dutching, explained: the data behind it

Dutching means backing every selection in the same market, spread across bookmakers, so that whichever one wins returns the same result. You stake each selection in proportion to its price. When the numbers line up, the outcome is settled before the event starts: any winner pays the same. It is the technique you reach for when there is no exchange to lay against, and the data behind it is simple to state and hard to keep good: the best back price for every selection, complete across the market, and fresh.

What is dutching?

Dutching covers a whole market by backing each outcome, rather than backing one outcome and laying it on an exchange. In a two-way market you back both sides. In a three-way market such as football match_odds you back home, draw and away. You split your stake so the return is the same regardless of which selection comes in, and you pull each price from whichever book is showing the best number for that selection.

The maths is the sum of the implied probabilities. Take each selection's best back price, convert it to 1 / odds, and add them up. If the total is below one, the market can be dutched for a positive position; if it is above one, the book's margin is against you and there is no edge. Stakes are then set so each selection returns the same amount. The whole method lives or dies on one input: the best available back price for every selection at the moment you act.

How is dutching different from laying?

Laying and dutching solve the same problem two different ways: both remove your exposure to which selection wins. Lay-based matched betting backs a selection at a bookmaker and lays that same selection on an exchange, so the exchange carries the other side. Dutching never touches an exchange. Instead of one back and one lay, you place several backs, one per selection, across the books you cover.

That difference decides the data you need. A lay-based approach needs a matched pair per selection: the bookmaker back price and the exchange lay price, plus the liquidity behind that lay. Dutching needs the opposite shape: no lay side at all, but every selection in the market priced at its best back, with none missing. One method wants depth on both sides of a single selection; the other wants breadth across the whole market. We set out the lay-based side in the matched-betting data guide.

When is dutching used instead of laying?

Dutching is the tool for markets where no exchange price exists, or where the exchange has no usable liquidity. Laying only works if there is an exchange with a real lay price and enough money behind it. For UK football and racing that is usually the case, and Betfair, Smarkets or Matchbook will give you a lay to work against. Elsewhere it often is not.

The clearest example is emerging markets. In South Africa or Nigeria, domestic books that the big aggregators skip carry prices worth covering, but there is no deep local exchange to lay on. Laying is off the table; dutching across the domestic books is how you cover the market instead. That is a large part of why a dutching feed type exists at all: it serves the regions and markets where a lay side simply is not available.

What data does dutching need?

Dutching needs three things from the data, and all three at once: the best back price per selection, completeness across the market, and freshness. Miss any one and the position breaks.

• Best back price per selection: for each outcome, the highest back price across the books you cover. A worse price on one leg eats the whole position, so the feed has to compare books and surface the best, not just any price.
• Completeness across the market: every selection present. Dutching is a statement about the *whole* market, so a single missing outcome makes the stakes wrong and the cover partial. A dropped selection is more dangerous here than in lay-based work.
• Freshness: prices seen recently. The market moves, and a stale leg quietly changes the sum of implied probabilities, so an old price looks fine and isn't.

OddsRelay carries 60+ UK books with bet365 included, which is what makes the best-price comparison across a market meaningful: the more books priced for the same selection, the better the top back price you can surface for each leg. The same coverage extends to the domestic books in emerging markets where dutching is the route that works. You can see what is live on the coverage dashboard.

What does dutching data look like?

Here is one football market shaped for dutching: every selection carries its best back price and the book showing it, and there is no lay side, because dutching does not use one. This is an illustrative shape, not live data:

{
  "event": "Kaizer Chiefs vs Orlando Pirates",
  "market": "match_odds",
  "region": "za",
  "selections": [
    { "selection": "Kaizer Chiefs", "back": { "bookmaker": "hollywoodbets", "odds": 2.55 } },
    { "selection": "Draw",          "back": { "bookmaker": "betway_za",      "odds": 3.30 } },
    { "selection": "Orlando Pirates","back": { "bookmaker": "supabets",       "odds": 2.90 } }
  ]
  // no lay side: dutching backs every selection; region + freshness fields elided
}

The contrast with a lay-based row is the point. A matched row pairs one back with one lay, and carries a rating and a qualifying_loss derived from that pair. A dutching market carries no lay, no liquidity, and no per-pair rating: the figure that matters is the sum of 1 / odds across the selections, computed over the whole market once every leg is present. Different shape, different fields, different check.

The honest catch: spread prices move

Dutching is more exposed to price movement than laying, and it is fair to be plain about it. Your position depends on several back prices staying put across several books while you place several bets. Between the moment the feed reads a price and the moment your last leg is placed, any book can shorten a price and turn a positive sum negative. The more legs and the more books, the more the spread can move under you.

That is why freshness matters more for dutching than for almost anything else, and why we are careful about how we describe speed. Our posture is pre-match polling on roughly a few-second cycle, which suits pre-match dutching. We do not claim real-time or in-play pricing, because that is not what we ship, and a promise of live streaming would be the wrong thing to lean a multi-leg position on. Freshness and latency are published rather than asserted, so you can judge whether the cycle fits your strategy. For the full set of tests a feed should pass before you trust it, see what makes an arb feed usable.

Where dutching fits alongside arbitrage

Dutching sits close to arbitrage: both back several prices so the result is settled in advance, and both need best-price coverage that is complete and fresh across a market. The difference is scope. An arb pairs the best back with the best opposing price on the same selection; dutching covers every selection in one market at its best back. The data requirements rhyme, which is why the same feed serves both. We cover the shared foundation in how arbitrage data works.

OddsRelay delivers the inputs for both: best back prices across 60+ UK books with bet365 included, exchange lay prices from Betfair, Smarkets and Matchbook where a lay makes sense, and market-complete back coverage in the emerging-market regions where it does not. It powers a leading UK matched-betting platform today.

See it against your own markets

The fastest way to judge whether the dutching data holds up is to point it at the markets you care about. A free trial gives you the feed to test completeness and freshness against your own strategy, and the coverage dashboard shows what is live right now before you commit.