Where to Surf in July and August: 75,932 Spot-Days of Data

Surf travel advice is mostly folklore. Someone scored Uluwatu in August once, so Uluwatu becomes "the August spot." A magazine ran a Skeleton Bay cover, so Namibia goes on the list. Almost none of it is checked against what the ocean actually did.

So we checked.

We took 1,097 of the surf spots in our database and ran every one of them through the same scoring engine that powers Strike Mission, against ERA5 reanalysis data (the European Centre for Medium-Range Weather Forecasts' reconstruction of past ocean and atmospheric conditions) for 75 days spread across two July-August windows: 29 June to 26 July 2024, and 30 June to 15 August 2025.

That is 75,932 spot-days. Every one scored on swell direction, swell size, swell period and wind, adjusted for wind history, windswell, secondary swell and tide, against that specific break's own optimal windows. A slab is judged by slab standards. A beach break is judged by beach-break standards.

Here is what came back.

Finding 1: Surfing is mostly bad

Category	Spot-days	Share
Poor (0-44)	54,441	71.7%
Fair (45-64)	12,759	16.8%
Good (65-79)	6,838	9.0%
Firing (80-94)	1,857	2.4%
Epic (95-100)	37	0.05%

Nearly three quarters of all spot-days were Poor. Only 11.5% reached Good or better. Firing days were 2.5% of the sample.

Epic days happened 37 times out of 75,932. That is one in every two thousand spot-days. Across 1,097 breaks and 75 days, only 20 spots recorded a single Epic day.

If you have ever felt like you keep missing it, this is the reason. You are not unlucky. The base rate is just brutal.

Finding 2: Half the world's surf spots had zero good days

Of the 1,015 spots with a full sample, 483 of them (47.6%) did not record one single Good day in the entire 75-day window. Not one.

Only 282 spots managed even one Firing day.

This is the number that should change how you plan a trip. Picking a destination for July or August is not a matter of finding somewhere nice. Nearly half the options are simply switched off, and no amount of luck or local knowledge turns them back on.

Finding 3: The famous spots were dead

This is where the folklore falls apart hardest. Here is how the marquee breaks actually performed, measured as the share of days that reached Good or better:

Spot	Country	Good+ days	Firing+ days
Pipeline	USA	0%	0%
Nazare	Portugal	0%	0%
Mavericks	USA	0%	0%
Supertubos	Portugal	0%	0%
Mundaka	Spain	0%	0%
Lower Trestles	USA	0%	0%
Skeleton Bay	Namibia	0%	0%
Desert Point	Indonesia	0%	0%
Cloud 9	Philippines	0%	0%
The Wedge	USA	0%	0%
Cloudbreak	Fiji	1%	0%
Bells Beach	Australia	3%	0%
Snapper Rocks	Australia	14%	3%
Teahupo'o	French Polynesia	15%	0%
Jeffreys Bay	South Africa	19%	0%

Ten of the fifteen most recognisable waves on earth produced nothing. Not a bad run of luck: a structural mismatch between when they work and when people take their summer holidays.

Some of this is obvious once you say it out loud. Pipeline and Mavericks are North Pacific winter waves, and July is their off-season. But plenty of surfers still book Portugal in August expecting Supertubos, or Siargao in July expecting Cloud 9, because the break is famous and the timing never gets questioned.

Big-wave spots as a class were the starkest: across the 26 in the sample, 0.1% of days reached Good and none reached Firing. They are winter instruments.

Finding 4: The hemisphere gap is enormous

Hemisphere	Spots	Good+ days	Firing+ days
Southern	453	17.2%	3.9%
Northern	562	7.0%	1.3%

Southern hemisphere spots were roughly two and a half times more likely to be worth surfing, and three times more likely to fire. July and August are deep Southern Ocean winter, which is where the groundswell is manufactured.

Finding 5: The best places are not the famous ones

Ranked by the share of days that reached Good or better:

Spot	Country	Good+ days	Firing+ days
Cimaja	Indonesia	91%	37%
Nosara	Costa Rica	89%	23%
Jimmy's Right	Indonesia	89%	53%
Carrizalillo	Mexico	88%	43%
Honey Smacks	Indonesia	86%	43%
Camana	Peru	86%	25%
El Transito	Nicaragua	85%	39%
Asuchillo	Nicaragua	83%	56%
Poneloya	Nicaragua	82%	37%
Playa Chinchorro	Chile	81%	16%
Lacerations	Indonesia	81%	51%
Balangan	Indonesia	81%	49%
Bingin	Indonesia	79%	40%
Las Gatas (Zihuatanejo)	Mexico	78%	19%
El Brazo	Chile	81%	12%

Asuchillo in Nicaragua fired on 56% of days. Jimmy's Right in Indonesia on 53%. For comparison, Cloudbreak and Teahupo'o managed 0%.

Bali's Bukit shows up repeatedly, which will surprise nobody. What should surprise you is how much of the list is Central and South America. Nicaragua's beach breaks and Chile's northern coast quietly out-delivered almost every famous wave on the planet during these months.

Finding 6: The country table

Averaged across every spot we track in each country, for countries with at least five spots in the sample:

Country	Spots	Good+ days
Costa Rica	15	36%
Indonesia	79	34%
Nicaragua	26	32%
Chile	49	31%
New Zealand	17	26%
Angola	9	22%
El Salvador	7	21%
Peru	54	21%
Mexico	59	19%
Panama	11	17%
Australia	95	13%
Fiji	9	11%

At the other end: the USA averaged 1%, Portugal 1%, the UK 1%, Spain 3%, Morocco 2%. Europe and North America are, statistically, places to be *from* in July and August, not places to travel to.

What this data can and cannot tell you

We would rather state the limits than have you find them.

This is a July-August sample, and nothing else. 75 days across two windows in the northern summer. It says nothing about where to surf in January, and it is not a ranking of the best waves in the world. Nazaré scoring 0% here is a statement about August, not about Nazaré.

It is reanalysis, not a session report. ERA5 is a physical reconstruction of wind and swell, and our engine scores that. It does not know about a sandbar that moved, a red tide, or the fact that a spot is unsurfable at low tide on a Tuesday because of a boat channel. It also does not know about crowds, which for several spots on that top-15 list is the real limiting factor.

Tide modelling is uneven. Our tide data is calibrated against official tables for Australian regions and modelled elsewhere, so tide-sensitive spots outside Australia carry more uncertainty than the rest.

The bottom of the range is compressed. Our scores floor at 0, so a genuinely unsurfable day and a very unsurfable day both land at 0. That is why every figure here is a *frequency of good days* rather than an average score. Averages would be distorted by the floor; frequencies are not.

Two summers is a small sample. 2024 and 2025 were what they were. A different pair of years would move the individual numbers. The structural findings, hemisphere gap and famous-spots-are-off-season, are not close calls and would survive.

The practical version

If you are planning a July or August trip:

Go south, or go to the tropics with a Southern Ocean swell window. Indonesia, Central America's Pacific coast, Peru, Chile.

Ignore the name of the wave. Fame is a measure of how good a spot is at its best, not how often it is good. Cloud 9 is a magnificent wave. In July it is not a wave at all.

Assume most days are bad, and give yourself more of them. With an 11.5% base rate for Good conditions, a four-day trip is a coin flip and a ten-day trip is a plan.

Check the forecast, not the reputation. That is the entire reason we built Strike Mission: it scores every spot we track, every day, and ranks them, so the question becomes "where is it good in the next ten days" instead of "where is it usually good."

Methodology: 1,097 spots scored against ERA5 reanalysis for 75 days (29 June to 26 July 2024, 30 June to 15 August 2025), 75,932 spot-days total. Spots needed at least 60 scored days to appear in per-spot rankings, leaving 1,015. Scoring uses the same engine that runs in the product, described in our scoring breakdown. Happy to share the aggregate dataset with anyone who wants to check the work.