# Prestige Fillies' Stakes 2026 Preview: Nine Lightly Raced Fillies on Soft Goodwood Ground _Goodwood, Sun 30 Aug, 15:10. Prestige Fillies' Stakes, Group 3, 7f, soft. Aperoll heads the market at 3/1 but our model prefers Bayside View._ *Source: [https://stablebet.co.uk/racing-news/goodwood-prestige-fillies-2026-preview/](https://stablebet.co.uk/racing-news/goodwood-prestige-fillies-2026-preview/) ยท last updated 2026-08-30* ## Prestige Fillies' Stakes 2026: an open Group 3 on testing ground **Sunday 30 August 2026, 15:10 BST. Goodwood. Prestige Fillies' Stakes, Group 3, seven furlongs, soft.** Nine fillies go to post for Goodwood's Group 3 on ground officially described as soft, and the interesting thing about the field is how little any of them have raced. Eight of the nine have run three times or fewer, and only Eabha has as many as five career starts. Coaxed has run exactly once, and won. That matters for a Group 3. When a field is this lightly exposed there is less form to separate them, the official ratings sit in a narrow band, and both the market and any model are working from thin evidence. Our own model puts the whole field between 15% and 9%, which is about as flat a spread as you will see in a pattern race. Aperoll heads the market at 3/1. She won last time out for Richard Hannon and has won two of her three starts over seven furlongs, which is the strongest trip record in the race. Bayside View is next at 7/2, with Oisin Murphy booked for Michael Bell. The two most interesting names may be further down the market. Nuit D'eclair carries the highest official rating in the field at 100, and Coaxed, the once-raced Gosden filly, is unbeaten. Both are available at 5/1. Prices quoted here are the early ones we hold at the time of writing. They will move, particularly on ground that could change again before the off. Nothing on this page is a tip. ## The nine, and what the form actually says **Aperoll** (3/1, R Hannon, P J Dobbs) is the market's pick and has the clearest case on paper. Three runs, a win last time out, and two wins from three attempts at seven furlongs. Rated 95. If the ground holds up and she repeats her last start, she is hard to fault. **Bayside View** (7/2, M L W Bell, Oisin Murphy) was second last time and has won one of her three runs over the trip. Rated 94, so a pound below Aperoll on official figures. Her last three finishing positions average 1.67, level with Aperoll and Eabha, though Coaxed and Redline both average 1.0 off shorter records. **Nuit D'eclair** (5/1, K P De Foy, D Egan) is the highest rated of the nine at 100. Three runs, second on her latest. On ratings alone she should be shorter than 5/1, and the market appears to be marking her down for something the bare figures do not show. **Coaxed** (5/1, J and T Gosden, W Buick) has the shortest and cleanest record here: one run, one win, over this trip. A Gosden filly with Buick booked and an unbeaten record is the sort of runner a market prices on potential rather than evidence, because there is barely any evidence either way. **Light Of Dawn** (5/1, K R Burke, C Lee) was third last time out and is rated 99, second highest in the field. Two career runs. Behind the 5/1 group the prices lengthen quickly. **Apollo's Comet** is 10/1, **Siena Storm** 16/1, **Redline** 25/1 and **Eabha** 33/1. Eabha is worth a note for one reason. At 33/1 she is the longest price in the race, but our model rates her at 10.8%, which is higher than it rates Apollo's Comet, Siena Storm or Redline, all of whom are shorter. That is the biggest gap in the field between a price and a model estimate that sits above it. The market disagrees hardest in the other direction on Aperoll, whom it rates 9 points higher than the model does. It is not a recommendation, and the model has no special knowledge of why she is friendless in the betting. ### The ground Soft is the headline condition, and it is the thing most likely to reshape this race. None of these fillies has a deep body of form to tell you how they act on testing ground, which is another way of saying the ground introduces uncertainty rather than resolving it. Check the going update before the off. ## Where our model disagrees with the market Our [AI Race Predictor](/ai-race-predictor/) prices every runner independently of the bookmakers, then compares the two. On this race it does not agree with the favourite. The model's top-rated runner is **Bayside View** at 14.8%, ahead of **Nuit D'eclair** on 12.6% and the market leader **Aperoll** on 11.8%. The market makes Aperoll a 21% chance. That gap, model 11.8% against market 21%, is the largest disagreement in the race, and it is driven by the model weighting her three-run record less heavily than the market does. Two things are worth saying plainly about that. First, the model is not telling you Aperoll will lose. A 21% chance and a 12% chance both describe a horse that probably does not win. The disagreement is about how much probability to assign, not about the outcome. Second, the bookmaker margin on this race is about 19%, which is wide. On a nine-runner Group 3 that is the cost of doing business built into the prices, and it is the single biggest reason a punter loses over time. Our [betting-margin explainer](/ai-race-predictor/lab/understand-odds/) shows how that number eats returns regardless of which horse you fancy. ### How much weight to give this Not much, on its own. Over 9,111 races the model's top-rated pick has won 24.6% of the time, and flat staking those picks at starting price returns a **loss of 13.7%**. Over the last 30 days it is running at minus 20.7%. The model is not a tipping service and it does not beat the market: on calibration it scores a Brier of 0.100 against the market's 0.092, so the market is still the better forecaster overall. What the model is useful for is spotting where the market's confidence looks thin. In a race like this, where nine lightly raced fillies sit in a narrow band and the book is 19% over, the honest read is that this is a hard race to price and an easy one to lose money on. This race is not unusual in that respect. We looked at [how often the model and the market disagree across today's whole card](/racing-news/model-vs-market-30-august-2026/), and on the races we can price honestly it is 18 times out of 19. Full figures, updated after every meeting, are on the [model track record](/ai-race-predictor/track-record/). > **Responsible betting reminder:** Set a budget before you bet and stop when you reach it. Free, confidential help: [BeGambleAware.org](https://www.begambleaware.org) or the National Gambling Helpline on 0808 8020 133.