Retirement cruising communications

[rwidman]

I thought the reason for retiring and sailing away into the sunset was to get away from your land life. You don't really need to be attached to the Internet and phones 24/7.

That said, you won't need anything but your phone as long as you're in the USA. Check with your carrier's coverage map. I did long ago and Verizon had by far the best coverage. This may or may not have changed since then. No booster needed.

As for the rest of the trip, you have been given a lot of (conflicting) advice so you'll have to evaluate it yourself. It would be a shame to spend several thousand dollars on stuff you don't really need. So you're out of touch for a day or two. So what?


The EPIRB is a good investment though.

[barnakiel]

Originally Posted by Kenomac

(...)

When you pay for the upgraded version, you get much more data than the free services offer.

Wind, swell, water and air temp, weather routing and localized forcasts.

(...)

But like everything, if you don't pay to play..... you'll never know the difference.

Well,

I get all the data you mentioned free from zyGrib and qtVlm. I do not get water temp, true, but I get it from the sensor on my boat anyways.

Good companies let you test their product before you buy it. Every time I tried to test Predictwind I was forced into "registering" (= this is how collecting my email, phone and other data is called in places with poorly developed consumer rights protection).

I have nothing against this particular company. Still, when one sells repackaged or reprocessed data created from publicly funded and publicly available sources, one could at least let the potential user have a test ride so that we pay for something we know the value of.

Pay first then ride is like shoot first then ask. Very localised.

b.

[Kenomac]

Originally Posted by barnakiel

Well,

I get all the data you mentioned free from zyGrib and qtVlm. I do not get water temp, true, but I get it from the sensor on my boat anyways.

Good companies let you test their product before you buy it. Every time I tried to test Predictwind I was forced into "registering" (= this is how collecting my email, phone and other data is called in places with poorly developed consumer rights protection).

I have nothing against this particular company. Still, when one sells repackaged or reprocessed data created from publicly funded and publicly available sources, one could at least let the potential user have a test ride so that we pay for something we know the value of.

Pay first then ride is like shoot first then ask. Very localised.

b.

So..... you're criticizing a product you've never tried because the company asks for your email address in order to offer you a 40% discount on future purchases?

Interesting.

[barnakiel]

Originally Posted by Kenomac

So..... you're criticizing a product you've never tried because the company asks for your email address in order to offer you a 40% discount on future purchases?

Interesting.

You cannot say 'never tried'. The point is the try-out in case of Predictwind does not offer any of the extras you mentioned.

Today I got a new pair of Arnette shades. How many pairs of Arnettes, Oakleys and Mauis have I tried in the shop? All of them! (All of the shapes and modalities I liked, anyways). And we are talking of a product that goes at 50 to 200 bucks a shot.

So, Arnette got some of my hard earned freedom tokens while Predictwind got none. Because I like to try out what I am about to buy. There is no such try out in case of Predictwind: their tryout version does not offer anything above nor beyond what is available for free from alternative sources. Their paid product is a buyer's enigma.

Alas, my criticism is not directed towards any specific company, even though one can easily name a few good candidates to exemplify my reasoning. What I am actually trying to point our attention towards is that we are falling prey to industry-wide practices that do not have to be accepted by the educated consumer.

Be it for a day or an hour, what I do want is to have access to full functionality, not to some as good as anything else demo.

I am aware there are many consumer styles, so I bet Predictwind is not going to flip only because a weather router from the brims of the EU has indifferent opinion of their enigmatic product.

A 40% discount on some illusory 'future purchases' simply implies a company is overpricing its product some 66.66% at the very start. No, really, thanks.

And last but not least: you got it right: I am not giving my email address to anybody who asks for it, nor do I give them my phone number. This tough, has nothing to do with companies or products. I simply do not like my box overflown with spam and viagra adds. I am too old for that.

Let me try your product. Price it realistically, and I will buy it - if I find it does offer an extra value over what is freely available from government agencies and from free online and offline alternatives.

Ough, apologies for so lengthy a response!

+ Hugs & love,
barnakiel

[robert sailor]

Originally Posted by barnakiel

You cannot say 'never tried'. The point is the try-out in case of Predictwind does not offer any of the extras you mentioned.

Today I got a new pair of Arnette shades. How many pairs of Arnettes, Oakleys and Mauis have I tried in the shop? All of them! (All of the shapes and modalities I liked, anyways). And we are talking of a product that goes at 50 to 200 bucks a shot.

So, Arnette got some of my hard earned freedom tokens while Predictwind got none. Because I like to try out what I am about to buy. There is no such try out in case of Predictwind: their tryout version does not offer anything above nor beyond what is available for free from alternative sources. Their paid product is a buyer's enigma.

Alas, my criticism is not directed towards any specific company, even though one can easily name a few good candidates to exemplify my reasoning. What I am actually trying to point our attention towards is that we are falling prey to industry-wide practices that do not have to be accepted by the educated consumer.

Be it for a day or an hour, what I do want is to have access to full functionality, not to some as good as anything else demo.

I am aware there are many consumer styles, so I bet Predictwind is not going to flip only because a weather router from the brims of the EU has indifferent opinion of their enigmatic product.

A 40% discount on some illusory 'future purchases' simply implies a company is overpricing its product some 66.66% at the very start. No, really, thanks.

And last but not least: you got it right: I am not giving my email address to anybody who asks for it, nor do I give them my phone number. This tough, has nothing to do with companies or products. I simply do not like my box overflown with spam and viagra adds. I am too old for that.

Let me try your product. Price it realistically, and I will buy it - if I find it does offer an extra value over what is freely available from government agencies and from free online and offline alternatives.

Ough, apologies for so lengthy a response!

+ Hugs & love,
barnakiel

Tough guy to argue with, well written response. R

[jkleins]

Originally Posted by barnakiel

There is no such try out in case of Predictwind: their tryout version does not offer anything above nor beyond what is available for free from alternative sources. Their paid product is a buyer's enigma.

I don't think this is a true statement. While I am no fan of Predictwind's forecasts they do give you there own forecasts based on the GFS and the Euro models on the free version. That is something that you cannot get for free in those other products you mentioned. The thing the paid product has that the free doesn't as best I can tell is just weather routing information. You can give them a destination and your boat info and they tell you when to leave and how best to sail it.
So it is not fair to say they are just repackaging information already available. They are giving you a distinct and different forecast that unfortunately seems to be less accurate then the original sources for the East Coast of the US.
As far as the email thing is concerned you may find the site mailinator.com of use to you. It is free so you can use it for things like this very effectively without compromising your own accounts.

Jim

[Kenomac]

Originally Posted by robert sailor

Tough guy to argue with, well written response. R

But.... I'm the one sitting here using a superior product for weather routing which only cost $11 annually.

I discovered the Predictwind product right here on CF recommended by a friend in the Med Charlie on Sulaire. Sometimes it pays huge dividends to not be so skeptical.

[Akka]

It's not easy to determine how good a weather prediction service is. Just because a service sends lots of data (say, 1km for every 2 hours it to 96 hours in the future) doesn't mean you'll get good predictions. It all depends on the weather model that service uses to generate predictions. Here are a few model features to look for:

1. Type of model. The global models are synoptic, meaning they rely on the physics of large-scale phenomena such as pressure gradients and high-altitude winds to predict the movement of existing weather systems and the development of new ones. The earth is modeled as a smooth sphere and purely local phenomena such as diurnal cycles (e.g. sea breezes during the day and land breezes at night) are ignored. Mezoscale models take regional factors such as topography and the air-sea interaction into account. Microscale models use even more detailed physics but generally are so computationally intense that they can only cover small areas.
2. Granularity. Global models have computation points as much as 200 miles apart. Microscale models may have grids as fine as hundreds of meters. Mezoscale models are somewhere in between.
3. Temporal resolution. Generally, the more computation a model requires, the greater the time between predictions. So if you want to know the weather every 2 hours you're probably not going to use a very detailed model.

Unfortunately, it's not that easy to determine the characteristics of many commercial weather models. For example, a weather app might give hourly predictions, but that doesn't mean there are model runs for each hour (I guarantee there aren't). There may be runs for every 6 hours and the forecasts between those times are (fairly sophisticated) interpolations. So gale force wind predicted for 1500 will probably arrive some time between 1200 and 1800, but the model doesn't say when. The 1500 prediction is simply the interpolation best guess.

A more dramatic issue is grid size. Just because a model produces a 1km grid of output doesn't mean it's a 1km model (and if it's free or only tens of dollars a year, I guarantee it's not a 1km model!).

So how can you tell what kind of model a given service is using? First, the service should tell you the name of the model they use; if not, don't use that service. Once you know what model they use, read the technical articles on that model and look for things like grid granularity, adiabatic wind modeling and sea/air interface physics.

Second, look at some model data for areas with major diurnal effects or topological features. Does the model show sea breezes during the day and land breezed at night? If not, it's a synaptic model and not much use near land. Does it show the venturi effect in Mountain passes?

Now for granularity: pick an area with a major feature (say, an island with a mountain on it -- I like to use Angel Island in San Francisco Bay) and see if the wind arrows bend around the island. If not, the model grid is bigger than the island's size, no matter what the service says.

Finally, the big test: download GRIB files from the model every few hours for an area prone to local phenomena until that phenomenon occurs (say, a Mistral in the Mediterranean), then look back at the predictions to see if, and how far in advance, the model predicted those winds.

One last caveat: weather is chaotic, meaning that very small changes in the values of input parameters can result in very large changes in the weather. The finer the model, the more closely it conforms to real weather physics, so the more chaotic its predictions become. Weather modelers say their model "diverges", meaning it no longer can be reconciled with the actual weather in the analysis phase of the model runs. In that case the model must be reset using current data and the output of a synoptic model. For a day or two, the reset model will essentially agree with the far less precise synoptic model. To guard against model divergence, you should continually monitor the weather service's predictions as compared to the observed weather at your location and avoid using that service when and shortly after those two diverge.

Good luck!

[conachair]

Originally Posted by Kenomac

But.... I'm the one sitting here using a superior product for weather routing which only cost $11 annually.

I discovered the Predictwind product right here on CF recommended by a friend in the Med Charlie on Sulaire. Sometimes it pays huge dividends to not be so skeptical.

Predictwind superior?

I've just signed up for another year but mainly because their data display is very good and it doesn't cost much.

But their own 2 models are still on probation, they seem very erratic more than a day or 2 out and seem to be on the high side as well.
Ass said in another post, getting a handle on accuracy is very difficult and a lot of work even just to get a rough idea. GPS does the best in the areas where I've spent some time cutting and pasting forecasts against actuals.
If you have any data to share it would be interesting to have a look.

Forecast




Actuals.

[barnakiel]

Originally Posted by jkleins

(...) forecasts based on the GFS and the Euro models on the free version. That is something that you cannot get for free in those other products you mentioned.

The thing the paid product has that the free doesn't as best I can tell is just weather routing information.

You can give them a destination and your boat info and they tell you when to leave and how best to sail it.

So it is not fair to say they are just repackaging information already available. They are giving you a distinct and different forecast that unfortunately seems to be less accurate then the original sources for the East Coast of the US.

(...)

GFS is available from many free sources, e.g. zyGrib (as well as qtVlm, Windfinder, Windguru, etc., etc.)

What 'Euro models' do you mean. ECMWF?

If so then it is incorporated by most EU met agencies in their state models: e.g. Met Office, yr.no, Aemet, to name just three. It is also freely available online for your own use. Read here:

Accessing forecasts | ECMWF
Public, WMO and ACMAD Datasets | ECMWF
Member States | ECMWF

Weather routing is available free from qtVlm.

Of course, any forecast may differ from any other forecast. But just being different does not add overall value. It may add niche value.

Let's just hope Predictwind will simply let us test its full functionality so that each and every party can tell for themselves how well the product fills each user's specific needs.

b.

[jkleins]

Originally Posted by barnakiel

GFS is available from many free sources, e.g. zyGrib (as well as qtVlm, Windfinder, Windguru, etc., etc.)

What 'Euro models' do you mean. ECMWF?

If so then it is incorporated by most EU met agencies in their state models: e.g. Met Office, yr.no, Aemet, to name just three. It is also freely available online for your own use. Read here:

Accessing forecasts | ECMWF
Public, WMO and ACMAD Datasets | ECMWF
Member States | ECMWF

Weather routing is available free from qtVlm.

Of course, any forecast may differ from any other forecast. But just being different does not add overall value. It may add niche value.

Let's just hope Predictwind will simply let us test its full functionality so that each and every party can tell for themselves how well the product fills each user's specific needs.

b.

But the actual predictions that they make from this available info is what they are selling. It is not fair to say you cannot test it when it is readily available in the free version. You may like or not like the predictions but they are unique and are basically the product. The routing info that you pay for is based on these predictions and just is customized to your boat.

Jim

[Paul L]

Originally Posted by jkleins

But the actual predictions that they make from this available info is what they are selling. It is not fair to say you cannot test it when it is readily available in the free version. You may like or not like the predictions but they are unique and are basically the product. The routing info that you pay for is based on these predictions and just is customized to your boat.

Jim

So PW is blendng two models to come up with their display. They are using a proprietary method to do this. Anyone know what they are actually doing? Averaging, averaging removing extremes, guessing at best model for location/time???

If the models are close to each other there accuracy is usually pretty decent, as they diverege the accuracy of both models is likely lower. In practice I'd rather compare multiple models when planning while still ashore and use a single model that has been most predictive over the last few weeks while offshore.

[Kenomac]

Using Predictwind, I can choose between several different models and accuracy down to 1 square km. We also use a radar model and pressure map, not just grib files. Total cost to me is $11 for all of 2017.... a bargain.

[Kenomac]

We also use Weather4DPro, Pocket gribs and Buoyweather, but now PredictWind is our goto app for accuracy and ease of use.

Our PC also downloads gribs and weather information overlays onto Nobeltec Odyssey from a weather service. We also have a weatherfax onboard which never gets used.

[Paul L]

Originally Posted by Kenomac

Using Predictwind, I can choose between several different models and accuracy down to 1 square km. We also use a radar model and pressure map, not just grib files. Total cost to me is $11 for all of 2017.... a bargain.

That's resolution, not accuracy. There's a big difference.

PW has good displays of available weather data. It just isn't some magic forecasting system that beats the accuracy of the base data.