Living on the Links Spectrum

I'll disappoint the number-chasers straight away by saying that the purpose of this article is not to put a numeric total on it. Nor do I think such a total is even possible. What I will try to do is take an in-depth look at all the shades of grey, evaluate the courses considered to be on the links spectrum, and place each one somewhere along it. You can then draw your own line in the sand as to what you personally deem to be a true links.

To compile this piece I have considered over 300 courses and tried to place them somewhere along the continuum. Information has been drawn from several published sources, alongside my own personal experience of playing the vast majority of the venues — only Tralee remains on my own to-do list within Great Britain & Ireland. For the purposes of this essay I am focusing on courses in Great Britain & Ireland, including the Isle of Man and the Channel Islands. If you'd rather skip straight to the verdicts, jump to the complete filterable directory at the foot of the page.

The starting point is Classic Golf Links of Great Britain & Ireland by Donald Steel, published in 1992. In this excellent book Steel writes warmly about 75 courses he considers the finest seaside links. I've yet to encounter anyone seriously claiming that any of the courses Steel selected isn't a genuine links. Some embody the spirit of the links game more than others, but in defining each one as a links we are on safe ground. The only one I'd potentially question from personal experience is Felixstowe Ferry, which plays less links-like than the others for the majority of the round — at least until the closing few holes.

I don't intend to split hairs at the top end of the spectrum. Otherwise we'd be debating whether the three holes the other side of the railway demerit Tenby's links status; whether the front nine at Pyle & Kenfig really qualifies; where exactly is the sea Lytham; what to make of all the trees at Formby; how genuine a links the middle stretch is at St. Enodoc, or holes six to ten at Prestwick for that matter. Parts of Westward Ho! and Montrose will soon be underwater, so what about the new holes? Put any course under the microscope and you'll find flaws and blemishes.

As a side note, I do wonder whether West Cornwall and Great Yarmouth & Caister would still be included if the book were rewritten today, more than three decades on. Not because of their links credentials per se, but because it's debatable whether they'd still be considered 'classic' alongside the others. Crail (Balcomie) might have a similar, smaller question mark, as might the two Royal Channel Island courses. Undisputed links, all of them — but perhaps missing the 'classic' criterion now. It's also worth noting that The Machrie and Skibo Castle have both been significantly redesigned since the book was written; we'll come back to those.

There are of course many more links courses in the British Isles, and another book — True Links by George Peper and Malcolm Campbell, published in 2010 — does a superb job of shining a light on them. As you'd expect, none of the courses in Steel's book are omitted, and the authors add another 137 in GB&I. There are, however, a few debatable inclusions and a number of clear omissions, which we'll come to.

After rightfully appointing The Old Course at St. Andrews as the crucible of links golf, the book splits its picks into 'Icons' and 'Classics'. From these chapters Brora and the New Course at St. Andrews are bracketed as Icons; Castlerock (Mussenden) and Portstewart (Strand) sit among the Classics, alongside Murcar Links and St. Andrews Jubilee in Scotland, and Ballyliffin (Old), Enniscrone and The Island in the Republic. I'm happy to include all of these alongside Steel's selections at the upper end of the spectrum.

The book also includes a category of modern links — many built since Steel's book was written. There's no doubt these creations look, feel and play like a links (certainly to the untrained eye), but die-hard purists will perhaps question whether layouts sculpted with heavy machinery, often imitating rather than utilising natural landforms, can be considered 'true' links in the strictest sense. Some go as far as labelling them faux-links, which is why I list them separately. Before you object: this article isn't really about course quality — many of the courses below are exceptional — it's about authenticity as a 'genuine' links.

The story doesn't end with the True Links book, because on my travels I've encountered a number of links courses not mentioned in the tome. "No way," I hear you say. But it's true. I'm not sure whether they were considered and dismissed, or simply overlooked. I strongly suspect the latter. I'm not the only one to think this. The next source is the excellent Top 100 Golf Courses website, which produced a news article in 2017 (updated in 2021) listing the 240 'Real Links Courses' of Britain and Ireland. Just as usefully, the article includes a list of courses they investigated and dismissed.

From that list, Sheringham is the most controversial: they argue the chalk-cliff geology disqualifies it. Technically that may be correct, but centuries of windblown sand have given the turf and natural contours a quality matching some of the very best links courses, in my view. As we'll see, geology and topography can change over time.

Three courses on the Top 100 list are very much borderline links, and I've struggled to come down on either side: Wick, Golspie and Powfoot. I'm not convinced the turf at Wick is true enough; Golspie could arguably be classed as part-links; and Powfoot, in my gut, doesn't quite do enough — although it's very close. Still, I'm minded to side with the Top 100 panel here.

We're now hunting for courses not yet discussed in detail that may be links. This is where the real detective work begins. Surely there can't be any more, I hear you say. Let's see.

Our next reference is Another Journey through the Links by David Worley. Here we find inclusions of the Castle Course at St. Andrews, Crail (Craighead), Fairmont St. Andrews (Torrance & Kittocks), Ganton, Mullion, Old Head, Southerndown and Nefyn. I wouldn't deem any of these to be links, and neither does the Top 100 site — they all feature on its exclusion list. The Castle Course and both Fairmont layouts were built on cliff-top farmland; they may look a bit like a links, but the turf isn't links ground and doesn't play that way. Most of the holes at Nefyn play on similar non-links ground. Southerndown does play very links-like but is actually downland. Ganton is the odd one — some argue it should be classed as a links based on its proximity to the sea several thousand years ago(!). It's a logical argument I sympathise with, but ultimately I'm going to label it with a term I hate to use: inland links.

The following section concerns Scotland only, but it's an important one. It really emphasises that it's impossible to put an overall total on the number of links in existence, and to some extent highlights the folly of even trying to determine exactly what a links is. But further down the rabbit hole we go.

In Scotland's Golf Courses (last edition 2002), Robert Price takes a geological look at the land on which over 500 Scottish courses sit. As we've already touched on, different parts of the same course can play over different soils and terrain, but in summary Price identifies 113 of his 538 courses as having links characteristics — yet only 92 can be classed as genuine links.

"It is a gross over-simplification to put all links courses into one class. These are broad classifications because there are numerous and significant landform variations within each site." — Robert Price

Each course is assigned one of 11 'Landform' variables (links being one) and one of 4 'Vegetation' labels (parkland, woodland, moorland or links). The 92 courses Price highlights as the real deal score 'links' in both categories. That leaves 21 courses which are predominantly links in vegetation but not in landform. Things get complicated, because although you'd think a scientific assessment should prove more authoritative than personal opinion, it throws up some dilemmas based on what we've concluded so far.

Of the 25 Scottish courses Steel selected, Price classes only 23 as genuine links. Skibo Castle and Dunbar miss out, both falling into the 'raised beach' category for landform. As mentioned, Skibo has been significantly altered since both books were published, so let's leave it aside — although it's worth noting that both publications discuss the same course in time, and Price himself describes it as a 'traditional links' elsewhere. A classification error, perhaps. Dunbar is more surprising though — Price talks about the course as a links elsewhere in the book, and it clearly is one in my view.

From the Icons and Classics in True Links, Brora, Murcar and both the New & Jubilee at St. Andrews are all confirmed as links. Of the moderns mentioned, only Kingsbarns, St. Andrews (Strathtyrum) and Craigielaw existed when Price's book was published; all are listed as links (as was the now-extinct Kintyre at Turnberry). It would be interesting to know what Price would have said about Dundonald, Castle Stuart, Machrihanish Dunes, the Renaissance Club, Trump International, Dumbarnie Links and now Old Petty.

Most other True Links selections align with Price's assessment, but there are significant conflicts. Fortrose & Rosemarkie, Isle of Skye and Irvine are not links for either category (all 'raised beaches' for landform), while Durness, Hopeman and Strathlene Buckie are classed as 'raised marine platform' for landform and 'moorland' for vegetation — no mention of links at all. Balnagask (Nigg Bay) is listed as 'parkland' on a 'raised beach', and Cullen mixes raised beach & marine platform with links & moorland. Newburgh-on-Ythan is part-links anyway, with a mix of links/upland and links/moorland.

So in theory, those nine courses should perhaps be omitted from the 'true links' list despite being included in both True Links and the Top 100 list. At the very least, Price's analysis nudges them down the spectrum.

Comparing further with the Top 100 'Real Links' list — and particularly courses not in True Links — Rosehearty and Monifieth (Ashludie) are confirmed as links, and the Balgove at St. Andrews belongs too. Further proof for Machrihanish (Pans), Royal Aberdeen (Silverburn) and Kilmarnock (Hillhouse) can be found, and our earlier borderline cases — Golspie, Wick and Powfoot — are confirmed as genuine links by Price. Covesea isn't mentioned in the book, but it's mostly links; given it's only nine holes, the part-links section may suit it best.

As for the courses I personally ruled out: Archerfield wasn't built in time to feature, so we're none the wiser. Scotscraig, however, is deemed to be a links on both landform and vegetation. I don't see it myself, but I can't argue with the geology, so it possibly does belong on the list. Eyemouth is confirmed as raised beach / moorland — not a links. Westray isn't classed as a links (undulating / moorland). Crail (Craighead) is listed as a links on both fronts, contrary to general consensus — and to my own view.

The book also allows us to rule out any notion that the Fairmont St. Andrews courses are links. And although vegetation at Anstruther is deemed links, the landform is raised beach, so we can dismiss this one too. The Glen (raised marine platform & links) and Longniddry (raised beach & links/parkland) cannot be given 100% links accreditation either. Interestingly, Traigh is listed as links for both landform and vegetation, contrary to the Top 100 view.

There are courses on the Top 100 exclusion list that do make Price's links list. I haven't seen any of them: Isle of Barra, Fraserburgh (Rosehill), Nairn (Cameron) and Peterhead (New).

Price gives us some other, perhaps controversial, considerations, stating that these courses meet all the criteria for genuine links: Benbecula, Elie Sports, Gairloch, Murcar (Strabathie), Scoonie, Solway Links, Troon Fullarton and Vaul. I've seen only the Strabathie at Murcar (it would just about pass) and Troon Fullarton, which is certainly a links and is now included in the updated Top 100 list. I have my doubts about some of the others, particularly Gairloch and Scoonie.

I'd like to throw in some courses not previously mentioned that are definitely not links, but which appear misleadingly on a Wikipedia page detailing "all the links golf courses": Royal Cromer, Caldy and Machynys Peninsula. The article admits it also includes "links-style" courses and courses that share many features of links courses, which pretty much discredits it for our purposes.

That said, the same article does throw up some genuine links the major sources missed. Among them: Warkworth (9 holes, now in the updated Top 100 list), Alnmouth Village (9 holes, now in the updated Top 100 list), Buncrana (9 holes) and Kinghorn (now in the updated Top 100 list). Robert Price doesn't have Kinghorn as a links, but its classification (raised beach / moorland) matches some courses in True Links.

Girvan is down as links vegetation but not links landform by Price, and there's no doubting that the first eight holes are pure links — the final ten are certainly not. Part-links is where it belongs. Newport Sands (Newport Links) in Pembrokeshire is clifftop for the front nine and borders on links on the back nine — part-links is again the fair call. St. Davids City, also in Pembrokeshire, is a lovely nine-hole links and definitely belongs in the proper links lists, not part-links.

Maryport has the strongest case among the ambiguous Wikipedia entries. Robert Kroeger, in Golf on the Links of England, describes it well — it won't compete with a Rye or a Littlestone, but it does offer links golf. I'm sympathetic to the case for inclusion, although it sits very close to the line.

Other courses claimed as links by Wikipedia, but which research suggests probably aren't (or where it's inconclusive), include: Brighouse Bay, Buckpool, Bute (Kingarth), Craignure, Machrie Bay, Garmouth & Kingston, Portpatrick (Dinvin), St. Medan, Tobermory, Royal Tarlair, Stonehaven, St. Bees, Falmouth, Ilfracombe, Freshwater Bay, Gosport & Stokes Bay, Clacton on Sea, Frinton (Havers & Kirby), Gorleston, Blackpool North Shore, Iona, Uig Lodge, Cairndhu, Larne, Otway, Kilkee and Wicklow.

Brighouse Bay, Buckpool, Bute (Kingarth), Craignure and Machrie Bay are all mentioned in Robert Price's book as having links vegetation but not links landforms, so I'm happy to rule these out — your own criteria may differ. It's also worth noting that Brodick, Carradale and the Isle of Eriska fall under a similar heading, though they're not mentioned in the Wikipedia article. Garmouth & Kingston, Portpatrick (Dinvin), St. Medan, Tobermory, Royal Tarlair and Stonehaven are not predominantly links for either landform or vegetation in Price, so they should be disregarded entirely. Iona and Uig Lodge aren't in his book.

The "modern" or "faux" links category is itself a spectrum, and lumping every recently-built links-style course together does some of them a disservice. Knowing how a course was built really matters here.

Trump International (Old) in Aberdeenshire is the most instructive case. Despite being a thoroughly modern build by Dr. Martin Hawtree (opened 2012), the routing was threaded through naturally occurring dune systems on a 4,000-year-old SSSI-designated dunescape — the same kind of towering coastal dunes that gives character to nearby Cruden Bay, Murcar Links and Royal Aberdeen. The fairways are sand-capped for drainage, the stacked-sod bunkering has an "American" feel that some critics flag, and the SSSI status has subsequently been revoked because the course "partially destroyed" parts of the dune system. But the underlying landform is genuine, pre-existing linksland. By the geological tests Price applies, the Old is as much a true links as several courses we've already counted. I'd nudge it up from the moderns and into the upper bracket, with a caveat about the construction style.

Trump International (New) — the MacLeod Course, which opened in July 2025 — is a different beast. Only six of its eighteen holes occupy the coastal dune system. Holes 11–13 give North Sea views, but the remaining layout traverses heather-clad heathland and expansive wetlands. Even the developer describes it as a "links-heathland-wetland hybrid". It deserves a place in the moderns rather than alongside its purer sibling.

Castle Stuart (Cabot Highlands), by contrast, was explicitly "fabricated from farmland along the coast of Moray Firth" — Mark Parsinen and Gil Hanse's 2009 creation. Beautiful and impeccably routed, but built rather than discovered. Same with Tom Doak's neighbouring Old Petty, which opened in 2025 on similar sand-capped farmland. Kingsbarns, where Kyle Phillips "shifted hundreds of thousands of tonnes of earth" onto largely unremarkable coastal farmland in 2000, sits at the same end of the spectrum. These three are exceptional modern courses — but they are designed and built links, not the natural article.

Dumbarnie Links (Clive Clark, 2020), the Renaissance Club (Tom Doak, 2007), Dundonald, and Machrihanish Dunes sit in similar territory: high-quality modern designs on sandy coastal terrain that's been substantially shaped. Doonbeg, the Glashedy at Ballyliffin, Sandy Hills at Rosapenna, and St. Patrick's are the equivalent Irish moderns — all magnificent golf courses, all built. None of this is a criticism — the moderns are some of the most enjoyable rounds you'll have anywhere — but a purist asking "which of these is a true links?" needs to know that the answer is more nuanced than the label suggests.

A few other courses come up in general discussion. From personal experience, I'll immediately rule out Aldeburgh, North Foreland, Whitby and Holyhead. I can't say for certain about Axe Cliff, Bigbury, Carlyon Bay, Magdalene Fields, Mundesley, Lydd, Whitstable & Seasalter, Cooden Beach and Bridport & West Dorset. Some might be best described as 'cliff-top links', but we can confidently remove the word links from most of them.

A note on Ardfin on the Isle of Jura, opened in 2017: this Bob Harrison creation is regularly hailed as one of the most spectacular new courses in the British Isles, but it isn't a links. It's a clifftop course built on peat and rock, with thousands of tonnes of imported sand. It deserves a mention here only because the question is asked often enough that it's worth being clear.

No discussion of the modern links debate would be complete without mentioning Coul Links, near Embo and just up the coast from Royal Dornoch. Plans for an 18-hole course on the site (originally a Bill Coore / Mike Keiser project, more recently revived by Communities for Coul) have been mired in planning controversy for the best part of a decade, with Highland Council approval in 2023 followed by call-in by Scottish Ministers, a public inquiry in November 2024, and final submissions in early 2025. At the time of writing, the application is awaiting a final decision.

If it ever gets built it would be one of the most exciting new links in living memory. If it doesn't, that may also be the right outcome — Coul is a designated SSSI, SPA and Ramsar site of considerable ecological importance, and reasonable people are deeply divided.

Well done if you've read this far — you are clearly a links aficionado. You've heard what I think; what I now want to know is what you think. Are there any courses listed below that you don't think are links? Are there any I've missed? Which ones would you move up or down the spectrum? Drop me an email at info@golfempire.co.uk.

This article is reviewed and updated regularly. Last revised May 2026.