Why the bloody hell did you change your website?
"If it ain't broke, don't fix it!" - postscript to a number of feedback emails over the past two weeks.
Two weeks ago, we launched a major upgrade of the Swellnet website. Overall we've had an enormous amount of positive feedback on the new layout and products, and also some negative feedback. This is completely expected - any website that undergoes a structural transformation that forces users to alter their habits will provoke some kind of response.
Anyway, it's probably worth some explanation around the changes, and why we've done 'em.
There are two major components to the Swellnet upgrade, (1) the website itself which publishes surf reports, surfcams, news, photos and forums, and (2) a new computer model-based surf forecast system which replaces our old manual forecasts. This new forecasting system has been more than a decade in planning, and we're really stoked with the results.
There are two reasons for the upgrades. The old website was poorly designed and built, and suffered from too many technical problems. We tried valiantly to remediate these issues over the course of several years, but in the end it was clear that we'd be better off starting again. The old website was broke, to counteract the old saying. So we moved to a new publishing platform and rebuilt Swellnet in a completely new database environment.
In all honesty, this doesn't matter much to the viewer - it's moreso a benefit to the business and our staff, so that we can do things more efficiently. Like upload a gallery in thirty seconds, rather than thirty minutes.
We also gave the old logo the boot (reluctantly, from my end) because there were times when it really didn't do the job properly. So there's a new look 'n feel to sit on top of the new backend.
At the same time we also decided to launch our new forecasting system, which we've been slowly working on for many years. This a fundamental change to our core product, The Surf Forecast, which I started way back in 1999 as a fax service to a couple of Adelaide surf shops.
Because Swellnet's old forecasting system required a surf forecaster - an actual person - to manually update the forecast data every couple of days, it wasn't a scalable service. The Swellnet website was limited to thirty forecast regions in Australia; any expansion beyond this would require more staff, which wasn't cost effective.
The forecasting upgrade is not a change we made lightly. We actually had the ability to roll out an automated forecast many years ago, but back then we weren't happy with the output - the computer model data had to accurately replace the manual forecasts with similar wave heights and winds, and we hadn't quite got the formulas right.
Since then we've worked really hard to create a unique surf forecast service that mirrors the accuracy of our manual forecast. In many cases - because it's updating four times per day rather than three times per week - it's often outperforming our old manual forecasts.
But more importantly, we're now able to automatically convey the same kind of surf information we'd just have previously delivered manually, which means we can now spend time building more forecast tools and generating other content for Swellnet.
And the new forecasts scale right around the world - we've got over two thousand locations in the back end, each of which has been specifically calibrated to its own coastline. We are confident that it's the most accurate and useful forecast system in the world.
I know that the new forecast system will take a little getting used to, but we're here to help. We've already written a few articles on some aspects of the system that you should check out:
South facing vs north facing: what is your beach? (how swell directions are colour represented on the new site)
Understanding our new forecasting system
Sneak peak of new Swellnet forecasting tools
Otherwise, if you've got any queries or questions, just fire them off below or start a new thread in our forums.
And.. there's much more to come over the coming months as we continue to roll out new forecast tools and products. We'll announce them in due course - make sure you sign up to Swellnet so that you're completely in the loop. //BEN MATSON
Comments
Four clicks to get the most accurate forecast in my part of the world for free?
You monsters!!
Mate, we'll have it down to two within the next few weeks. Just click on the ads, will ya?
Seriously though, I love the feedback - good or bad. Helps us strive to create a better product.
Really? If I click on the ads, you guys make extra money?
Done!
Come to think of it, I need a new pair of Crocs.
Nah, it doesn't quite work like that. But still, if you see something you're genuinely interested in, it's nice to support those advertisers that support us. So clickety click...
Looking at those screenshots I must have been following Swellnet since 2003... from back in the old Burleighcam days haha. Cheers fellas I have got some great knowledge of the australian coast from this site over the years...
Howabout updating the billiard-ball avatar benny .. get some locks on son, and a bit of wind in the wig, plus the red jacket needs a swing over that fence there behind you, maybe something in animal skin instead - now that sellnought growls so loud.
All you need is an iPad, VW trakka, swellnet, a few boards and happy wife.
You can't do today's job
With yesterday's methods
And be in business tomorrow
Great new website. Thanks swellnet team
new site's looking good guys.
just wondering if you've looked into creating a swellnet app?
Yep, the App's nearly there. It's been half built for ages, waiting for the new site (and database) to go live. Now we're finally able to do the last mile and push towards getting it released. Not sure on an ETA but it's a big priority.
i used to use this site for a quick view of swell conditions.........since the changes, not any more, is now too slow with too much unecessary info.
Who ever has made the decision that this aspect of the site is improved, is quite frankly, full of shit
There are times I despair for the human race, usually around election time when I realise how many hay-fed automatons are in my presence, but this upgrade is serving as another fresh reminder. Stretch, you need only make two mouse clicks to every single piece of information, prioritised by bolding and order, a surfer could ever want! It is a hundred head and shoulders above anything else available on the web, Magic Sea Weed included.
Oh god I just can't help it anymore...
Hey Ben and Craig. I have an idea to keep everyone like strrretch happy... Have an alternate forecasts page for 'special' people with a different expression of the graphs. I am thinking maybe it could be done with coloured crayons and really large fonts? And maybe a gif. of Mr Squiggle could be beside it with his nose pointing in the direction of the wind... and whooshy wind sounds corresponding to wind strength. And the wiggles telling Jeff to wake up cause the surf is good... or stay in bed cause it's shit...
Though I guess that would take too long to load too. Maybe just a sad face or happy face would suffice?... or a 'yewww' for 'great conditions', a 'shaka' for 'average', and a 'bong' for 'don't bother'?
I disagree that shaka should be for average conditions.
Shaka is fun, double shaka is going off, and Yewww is mental, and Yew to the Yewwwww is once every two years mental.
Thumbs down means onshore, sad face means onshore & rainy and a red monkey's penis means the trifecta of grief. Onshore, rainy and flat.
The website is amazing. It's freely available everyday and it's a credit to you. Well done.
Seems legit- except the new colour scheme. It is now reminiscent of Kiwi ocean colour, not the blue of Australian ocean colour. Fully sick cuzzy brew!
Y'know what? I just googled the colour teal (which is apparently the greeny colour on our logo and Nav bar), and Wikipedia gave me this:
"In computers, the color teal represents software that is well-written and efficient with no glitches."
Ha!
I am super impressed with the new site - the new long term forecasting system is better than th old one - easier to follow and hopefully accurate. The issues in being able to actually see the long term forecast while annoying are an inevitable part of every site upgrade, rebuild etc. I am stoked with it. And access to more South Australian data is awesome in a state that is often an after thought in surfing media.
Thanks CG. We've still got a LOT of more work to go to get things the way we want it, most of which revolves around navigation/user experience etc. So I think any niggling concerns of yours will be elimated over the coming weeks and months.
But, of even greater importance is getting everyone to feel confident with the accuracy of the new system.
We've spent a lot of time tweaking the algorithms - and there are still many more tweaks to be implemented - however we'll do this in a transparent way so that everyone's aware that we have also identifyied those areas where the forecast sometimes gets it wrong (for example, it overcalled this week's south swell in SE Qld, but nailed almost the entire NSW coast).
We're aware of almost all of the (few) forecasting limitations of this new system and are working steadily behind the scenes to improve on it wherever possible.
How often does the forecast reset?
For example, here today, your report is pretty well spot on. As forecast, 1-2ft at south facing beaches. Just surfed my local at 1-2ft.
Wind-wise though, quite often the wind strength is under-called. A few times now, the wind has been called at Nth at 7 kts. Today you've forecast NNW at 7kts but on the ground here we've had northerly winds overnight and at dawn they shifted to a breezy SW (I estimate at about 10kts +) and currently the wind seems to be dropping off but still out of the south.
Your long range forecast seems to be in line with JMA (Japan Meteorological Agency) wind-wise and your surf report is pretty close to the other models except with more detail.
All in all, the forecast is pretty well spot on and I know it's impossible to pinpoint our local within 3 standard deviations of the mean, but a few times now the wind direction and strength have been inaccurate. Swell size and direction has been very accurate.
Love the site and not being picky, but I know you want this feedback.
Cheers for your efforts.
Great feedback Zen, just what we're after.
The model updates 4 times per day. We will put in 'next update due' times on the site shortly (one of hundreds of little tweaks to be done). This way you'll know how old the data is (which is never more than 6 hours).
Yeah I agree, wind forecasts are a little under in strength at most areas. It's just raw model output at the moment but we're going to tweak this over the coming weeks wth a few ideas I've got.
As for wind direction - hard to know without knowing exactly where you are located and exactly when it's getting it wrong (so I can correlate it to a particular kind of weather system) however it's probably due to the resolution of the model, and the local topographical effects of your region. We do see this happen in some parts of the country under particular weather patterns so it's very likely for Japan too (especially with the radical topo you have there). Again, we've got a possible solution for this but we'll have to look at slowly rolling it out to see what kind of improvement is made.
Out of interest, how did your previous source of forecast info (whoever that may have been) deal with the local winds? Most websites are pretty much using the same data sources, so I'd suspect that if we got it wrong, so would they as well. Just curious.
I generally relied on a combination of Stormsurf Japan for swell and direction, Weather-forecast.com for long range temperature and wind forecasts (9 days) locally and JMA for local conditions daily in Ibaraki. JMA updates hourly and it's useful for wind hindcasts (is that what you call it?). JMA is by far the best for reporting local winds and also good for tracking typhoons and Nth Pacific lows. JMA is also good for predicting snowfall.
And for up to the minute reports on local conditions, the old faithful look out the window and nip (no pun intended) down to the beach to see for myself never fails;)
Interesting. JMA doesn't score particularly well compared to other GCM (global circulation models, ie GFS, ECMWF, CMC, UKMET) however it would be expected that their high res modelling (ie short range wind forecasts) would be superior.
You are talking about 'forecast' winds though, and not 'observed/reported' winds?
My apologies,
I use weather-forecast.com for forecast winds and they're not bad.
JMA for observed/reported winds. I use JMA to assess local conditions, tracking, rain, storms, snow etc and hindcasting winds.
The other two sites I use mostly for forecasting .
I use yours now pretty much exclusively and just give stormsurf a cursory glance.
OK, that explains things. Weather-forecast.com uses GFS data (so their forecast winds should be very similar to ours.. if they're not I'll be interested to research why).
Observed/reported winds are much different, and we can't match that right now - however we're working on getting a live feed of them from JMA at the moment.
Sweet.
But anyway, once again, thanks for your efforts. I think Swellnet is a great site and I have no inclination to change.
I wish you and your team continued success.
MMMM yes.
Haven't really used it much since the update.
Was the "glitch free" forecast down for a week or so? That perhaps added to my reliance on Seabreeze.
New layout is freaking me out but, i'm sure i'll get used to it...
I just had a firefox update that made my bookmarks only slightly different and i hated it! For about 30mins when i realised that i didn't even notice it anymore.
Looking forward to using the new site.
(BM; I agree. Time for an Avatar update.)
Thirded.
OK, here ya go.
Haha, the similarity is striking!!
well well... the other aussie surf website just had to go and upgrade too now huh
haha yeah it's pretty sick, Nice simple layout with the report. But when you look at 'offshore swells' there's only 1 swell...
I think the next big thing should be putting a degree of certainty on forecasts. Something like error bars on the graphs in height and time
There's another Aussie surf website?
Hi Ben, Stu and Craig
On Thursday a secondary S swell is indicated to come in for the Waitpinga forecast, I was curious how this isn't reflected on the graph at the top of the forecast area (ie the 5 day graph still shows SW on it's arrows for the whole day). I am still tuning in with the new graphs.
Cheers
Mysto, our surf size calculation takes into account several factors including swell size and period. The swell direction on the graph is related to the dominant swell train.
So at midnight Thursday (12am) our model has deemed the windswell more dominant over the groundswell. This is due to the larger size of the windswell compared to the groundswell (by 0.5m).
But at 6am our calculations then put the SW groundswell in the primary position even though the windswell is 0.1m larger in size. The reason for this is that a groundswell of these characteristics will produce larger waves at the beach compared to the similar sized windswell.
Stunet censoring comments and laying guidelines as to what people can comment on so as not to offend his mates overblown egos, doesn,t that contradict the point of having a forum? Why not just become Surfline.com?
No, it saves us visiting the wigs in a defamation case.
Fair enough I guess, did old mate threaten you guys with court too? starting to see a pattern here, perhaps Brutus rants were closer to the truth than it would have seemed...
Nothing has been said, but all websites set thresholds and ours is way above every other surfing website. In the nearly five years we've had comments I could count the times I've had to delete a post with my fingers.
To be fair Mike, the moderators of these forums on SN are extremely tolerant IMO.
Hi Ben
I just completed a two week comparison of your forecast V magic seaweed.
MSW was more accurate on 7 out of 14 days and yours on 3 days with 4 days being the same. Swell info was it was pretty much the same but you wind forecasts were not great with both strength and direction being out more than MSW
MSW has all current and forecast data on one page instead of your two and interestingly with no wankery about multiple swells and directions is as good or better than swellnet and definitely easier to read and understand.
I'm doing a very similar thing though I also started to incorporate Surfline (I don't use them but figured it might be worth comparing data). I plan to do it for a month, the time hasn't yet elapsed but my results show SN more correct at my beach by a ratio of approx 2:1 (over MSW).
SN undercalled a recent south swell (I'm in the lower Hunter) while MSW substantially overcalled it. This seems to be their flaw. They also don't register swell direction changes very well. We've had a mix of NE, E and S swells since I began tabulating the data and MSW are almost always off with the direction (which is contributing to their size issues).
I had to do a last minute trip to the mid-north coast (no time to surf but I saw it every day) and the pattern was similar. The place I was at has a renowned south swell magnet yet there was a NE swell running. MSW overcalled it while SN wavered somewhere between exact and undercalling it. It would've been interesting what the figures were in a S swell.
I'm fine with the display. Use any interface long enough and it becomes natural.
PS: I stopped comparing Surfline after two weeks due to inaccuracy (was also taking up too much time).
Thanks for the feedback atticus - that's great info. Just another thing re: MSW and why their system is flawed - which location do you choose for your area? If we look at the Hunter, they have a number of locations to choose from. For example today they have:
Anna Bay: 3-4ft
Avoca Beach: 4-6ft
Bar Beach: 3-5ft
Box Head: 1-2ft
Copcobana: 3-4ft
Crackneck: 4-6ft
Newcastle: 4-6ft
Redhead: 3-5ft
Shelley Beach: 3-5ft
Shoal Bay: 4-6ft
Swansea: 4-6ft
Umina Beach: 3-4ft
Wamberal: 3-4ft
I mean, the swell's out of the east yet they've got Copa at 3-4ft while Avoca is 4-6ft? And Shoal Bay is the same size as Newy? And Umina is the same size as Copa and Wamberal? Some of the forecast numbers are OK, and some of them are terrible. Refining global wave model forecast output to a beach-by-beach level is inherently flawed for a number of reasons.
The only way to get any value out of their site - and perhaps they engineered it this way - is for a user to check four or five different locations and then average them out (which means clicking on four or five pages every time - consequently generating more page impressions and therefore ad revenue).
This exactly highlights the overarching problem with their system, and justifies why we've taken a completely different approach - providing regional forecasts (ie within a 50km coastal region) and then displayed the surf conditions relative to 'open beaches', 'south facing beaches' or 'north-east facing beaches'. Sure, it means less page impressions than our competitors but we're confident that the user experience is much better.
One other thing to point out (if you weren't already aware): MSW use 'face feet' while we use 'Surfers Feet'. Roughly, this means their 4-6ft forecast is the same as our 2-3ft forecast.
This may explain why they often overcall large south swells (which is very noticeable in the southern states). Although they seem to have a threshold that somewhat smooths the top end of significant swell events (just a hunch).
As per Ben's last comment, keep in mind that MSW forecast have to halfed because they talk in face feet, not surfsers feet. Ie Thei 6ft is really our 3ft.
A couple of other quick observations from around the country:
Bells Beach - observed surf report is 2-3ft today (Surfers Feet) with NW winds
Swellnet forecast was 2-3ft (accurate) with NW winds (accurate)
MSW forecast was 2-4ft (Face Feet = 1-2ft Surfers Feet) - which is an undercall, with NW winds (accurate)
Phillip Island - observed surf report is 2-3ft today (Surfers Feet) with N'ly winds (nearest weather station reporting W/NW)
Swellnet forecast was 3ft (accurate) with NW winds (reasonably accurate)
MSW forecast was 2-3ft (Face Feet = 1-1.5ft Surfers Feet) which is a significant undercall, with NW winds (reasonably accurate)
Mornington Peninsula - observed surf report is 3-4ft today with moderate W/NW winds
Swellnet forecast was 3-4ft (accurate) with W/NW winds (accurate)
MSW forecast was 3-4ft (Face Feet = 1-2ft Surfers Feet), which is an significant undercall, with NW winds (reasonably accurate)
So - a quick assessment here is that MSW is undercalling Vic locations by about half.
Thanks for the feedback pointy. What location are you at? Hard for me to respond without knowing which coast or location you're at (because there are flaws in the MSW model which are dependent on which location you're at).
I've been assessing our model against the surf conditions nationally each morning since we went live and it's been very accurate (for example, every dawn report that's coming through this morning is exactly as the model forecast it).
Also just to clarify, MSW do have multiple swell train information like we do. However their interpretation and display is a little different. I completely understand that some people will prefer their disply to ours, but the feedback we've had so far is that we're on the right track.
Also, I'd be more than happy to do a real time, public assessment of our surf forecasting system vs any other forecasting website (ie in a forum thread). This is the only way to really assess the accuracy with any degree of transparency. If anyone's interested, let me know.
Hi Ben
I took into account he face feet V surfers feet call when comparing.
I'm at Thirroul, MSW have a forecast for Austimer (next beach north).
interesting to note that the days where SN highlighted different sizes for diff facing beaches you were wrong. Thirroul misses south swell but gets NE swell. It did not get the extra foot on a day where N facing beaches should have got it but did get the extra foot on a day S facing beaches should have got it. MSW got it right both of those days.
But as I said, swell wise there is not much in it but wind wise SN accuracy was poor. I set my work/surf schedule (when I can be flexible with work) more to get the right wind than swell as I don't have time before work to drive around much and so surf my local (needs W but can live with SW - NW). Nov 15th to 19th SN had the wind direction wrong everyday and while MSW missed two of those days as well were better on the other two.
e.g. one day you predicted S winds and MSW said SW. I wouldn't get up for an early with S winds bit will for SW and got good waves that day.
I can live with the swell forecast on SN but the wind forecast needs tweaking
PS SN is a better website as it has the odd article of interest (not enough though) and I would prefer to use SN. And yes I will probably get used to the layout with time but you need me coming back regularly for that.
I here Stu has moved to this neck of the woods. It would be interesting to get his feedback
Thanks for that extra feedback pointy. The wind is a bit of an issue, and we have plans to get higher resolution modelling around certain regions in to help this.
In the meantime though I find if the wind is under 10 kts or so and from the S or N, local effects will swinging it SW or NW etc. And if it says onshore at say 5kt then you'll probably see a light morning land breeze instead of that onshore.
SE swells also need to be improved upon across the East Coast and we've also got fixes in the works for this.
Yeah that's great feedback - thanks Pointy.
As per Craig's comments, we're working on an improvement in our wind forecast. However, all websites - Swellnet, MSW and Surfline - get their wind forecasts from the same source so theoretically there shouldn't be too much of a difference bwteeen 'em.
I'll keep a close eye on our Coal Coast forecast to see if I can pick up on any trends.
Also.. just checked the MSW forecast for your region (seems they only have three locations between Sydney and the 'Gong). They had:
Coalcliff Reef (?) - 4-6ft with N'ly winds
Austinmer Beach - 3-4ft with N'ly winds
Wollongong - 4-6ft with N'ly winds
Our forecast was 2ft at open beaches, reaching 2-3ft at NE facing beaches, with N/NE winds.
Surf report came in at 2-3ft and the Bellambi AWS is reporting winds between the N/NE and N.
So it seems their Austi report is an undercall while the CC and Gong forecasts are about right, as was our forecast. Both websites got the wind forecast bang on.
I don't want to ramble on about this Pointy, but I thought it was also worth checking the current forecast data for next Tuesday. We've got an interesting system developing in the Tasman Sea that's looking to kick up a pretty sizeable S/SE swell.
As per the images below (screengrabbed a few minutes ago), you can see Swellnet's forecast for the Coast Coast sees Tuesday increasing from 3-5ft to 6ft+ during the day.
However Magic Seaweed has 2-3ft surf in the morning, building to 3-4ft around lunch and then 4-7ft by the evening.
So using surfers feet, MSW's forecast is 1-1.5ft building to 1.5-2ft then 2-3ft+. That's around half of our projected size range. Will be interesting to see what happens.
actually, all over called it. Reality (at Thirroul) was 1-2Ft (by your terms) with the 2's being rare. (if the forecast was 1ft I would call it correct)
That was at 5:15am. I'm working from home and will wander down at lunch time and let you know what it's doing then.
Well, at least there's some consistency across the sites!
lunchtime update is 2-3ft at Thirroul so swell is rising
Better get yourself up to Stanny then Pointy. Lot more swell hitting there at the moment!
too much work to do unfortunately.......
G'Day Pointy,
Yeah, I've moved to the area and I've been watching the forecast and reports pretty closely, as you could imagine. I've been seeing a few errors with our model though none of the magnitude that you mention; a north swell two weeks ago came in a bit under, as did last weekends SE swell (though that had accompanying SE winds so I surfed at more sheltered spots anyway).
I've been watching MSW for a few years - again, as you'd expect, it's my job - and they miss so many days I only pay attention for professional reasons. Of course I'm gonna be biased but I do try and apply some objective detachment.
Here's an example: This morning MSW has 4-6 foot at my old local, Cronulla. It's also calling the NE winds 'cross/offshore'. Yet the surf is 1-2 foot and the NE winds are only crosshore at Wanda, everywhere else is onshore.
More about winds: I've lived in four different places over the last six years, all within Sydney/Wollongong, and understand that micro-effects can alter the wind - especially morning land breezes. The Northern Beaches is a classic example of that, the morning landbreeze (offshore) will always blow longer at the N Beaches than at Cronulla.
For this reason, I take any wind forecasst with a pinch of salt. The only foolproof method is checking the wind obs (current wind observations) and Swellnet has them on the forecast page.
I miss having a written forecast, all there is now is a swell height chart :(
We've still got the written forecasts JB, you can find them here: https://www.swellnet.com/forums/swellnet-forecast-notes