Predict Wind

[Paul L]

Originally Posted by s/v Jedi

In your opinion. I bet you don't do weather routing nor departure planning. If any of that includes upwind tacking then the difference is enormous. Can the algorithm decide to tack every 100nm or at 50nm or 8nm. You don't think that matters?

.......:

You are right I'm not a professional weather router. Have done 40,000 miles since we left and still going.
You are confusing the routing algorthim with the Grib information. A routing algorthm does not make its route change decisions only on actual Grib data points. Each algorithm will use whatever data point interprolation it is programmed for.
That has nothing to do with PW making up new models which are display interpolations that have no more accuracy than the original, but pretend to offer it to the viewer. The temporal error in Gribs does not lend itself to simple interpolation. You are fooling yourself if you think the accuracy of the Grib increases if you interpolate the data. This is independent of routing algorthims.

[chouliha]

The only reason we subscribe to PW is to get the European model gribs as they are not available to us when offshore via SSB. The blended forecasts for me are a waste of time and I don’t use them. I don’t use the passage planning as we are comfortable doing our own route planning. I think the app is poorly designed for the price they charge.

[mikereed100]

Originally Posted by s/v Jedi

Mike, you can enter your polar in the configuration (select your boat from a list or one close to yours) and also set a percentage of that performance. Example: the polar choosen is too fast: set the polar adjustment value to 70%

Sailing onto land is a map problem they have. Small cays/atolls/shalows are also not considered because the charts are lacking detail. But you can simply set an intermediate waypoint to steer clear. With AI you always need to coach it, never try to make it fail and most of the time it needs the help

I like to do my own routing as PW likes to send me to areas where I will be either motoring or bashing into 25 knots for days. I also think that they are using TWA instead of AWA in their calculations as well as ignoring sea state. What I do like to do is force their routes to follow my own by putting in lots of waypoints, then using the adjustment that you mention to match my own boat's speed. I can then move the little boats along the route to get a good idea of where I will be at a given point in time on the weather map which makes my weather routing easier.

I find that, even with all it's quirks, PW is very useful and that they are very responsive to questions and suggestions.

[Kenomac]

I much prefer Weather4DPro over PredictWind having used both for several years.

[zboss]

What you are seeing is perfectly normal and is basically saying that the two models you are viewing are in disagreement with each other and therefor your overall confidence in either prediction should be very low or nonexistent.

Either stay in port and wait for the models to converge for a better confidence level or look out your window and compare what you are currently seeing and match it against one of the models. Then use that model to determine what is going to work best for you. If none match then that’s a good indicator just to sit back with a book and wait for the weather to change.

[s/v Jedi]

Originally Posted by Paul L

You are right I'm not a professional weather router. Have done 40,000 miles since we left and still going.
You are confusing the routing algorthim with the Grib information. A routing algorthm does not make its route change decisions only on actual Grib data points. Each algorithm will use whatever data point interprolation it is programmed for.
That has nothing to do with PW making up new models which are display interpolations that have no more accuracy than the original, but pretend to offer it to the viewer. The temporal error in Gribs does not lend itself to simple interpolation. You are fooling yourself if you think the accuracy of the Grib increases if you interpolate the data. This is independent of routing algorthims.

Really?! I've always been confused about grib files and algorithms

Where did you read that I think the accuracy of a grib file increases with interpolation of data? You make that up

What matters is that the PW models do have added value and there is plenty of information available on what their models are and why they matter. It's just a matter of putting more time into it to read it all open minded instead of dismissing new ways before even understanding them, let alone trying them.

[GRIT]

I'm fairly new to Predict wind, and I'm using it on my B&G Zeus3.

When I ask it to predict a good route, it's sometimes got me going into the wind at 8 degrees true! I have it set to "outremer 45" polars, which doesn't show that boat sailing above 45 true.

Can anyone tell me how to setup Predict wind so it limits my pointing angle to 50 degrees true?

I tried setting it up with their "dumb polar" at 15 knots, and told it we were sailing, with the exact same issue as we had using the outremer polar. It sometimes wants me to point up to 5 degrees true.

Cheers.
Paul.

[GRIT]

I've just compared the route on the B&G with the route on the website, which shows the actual wind patterns.

I think it might be TWA off the stern? I see no TWA on my B&G Predict wind route over 89 degrees for this trip, but the forecast at predict wind shows wind on the bum for at least 10% of the trip. Do they really show angles off the stern, and off the bow, and you just have to sort of guess what's what? It's easily figured out if I'm on the internet, or have a grib overlay, but really?

Surely this can't be possible. Someone please tell me I'm reading this wrong.

Thanks.
Paul.

[s/v Jedi]

Originally Posted by GRIT

I've just compared the route on the B&G with the route on the website, which shows the actual wind patterns.

I think it might be TWA off the stern? I see no TWA on my B&G Predict wind route over 89 degrees for this trip, but the forecast at predict wind shows wind on the bum for at least 10% of the trip. Do they really show angles off the stern, and off the bow, and you just have to sort of guess what's what? It's easily figured out if I'm on the internet, or have a grib overlay, but really?

Surely this can't be possible. Someone please tell me I'm reading this wrong.

Thanks.
Paul.

Hey Paul how are you

It may be angle off the stern but also it may be that wind speed went below a setting where it starts motoring.

I only use the app

[GRIT]

Originally Posted by s/v Jedi

Hey Paul how are you

It may be angle off the stern but also it may be that wind speed went below a setting where it starts motoring.

I only use the app

Hi. I'm well, thanks. How are you? It's early, are you on passage?

Thanks for your quick reply Jedi.

I have it set so that it does not assume motoring at any windspeed. It's puzzling; something must not be set right on my plotter.

I don't think I'm going to be relying on this program much. If I still have to do all the work myself, it's not adding much value. It might be nice as a "second opinion" on my route choices, but there are two of us aboard, so second opinions are easy to find.

Cheers.
Paul.

[BigBeakie]

I had to speak with PredictWind in NZ about another matter, but FWIW I raised this issue of the PW weather routing giving upwind sailing angles that cannot be achieved.

With the proviso that for any specific instance or example they would have to see all the relevant data for that instance ( boat polars if available, date/time/latlong, GRIBs etc.) for review to give a specific answer.

The tech thought that probably what is going on is that people are expecting PW to give them a predicted route that is akin to a "roadmap", that is, a series of course segments that are more accurate than is possible. Basically if PW gives a course heading that is 15 degrees TWA,and your boat can't beat at 15 degrees (of course) then it is saying, "sail to windward as best you can".

It is an indicative heading, not a carved in stone reality.

[wingssail]

Originally Posted by gaucho

... use GFS and ECMWF ... forget about PWG and PWE, they are useless...

We usually sail in a large bay and in the nearby offshore area just beyond its mouth. These areas are dominated by the daily thermals and night time offshore flow (katabatic winds) as well as the overall gradient winds.

The GFS and ECMWF models do a fair job for the big picture (although often give quite different predictions) but for in the bay and the nearby offshore waters, by ignoring the katabatic winds, they are nearly useless for our purposes.

But the PWE and PWG apparently modify the global models to account for the local conditions, at least in our area. They show the local winds that the global models leave out. We find PWE and PWG are very accurate in that they predict the strength, direction, and timing of the thermals quite nicely.

Usually I find PWE is most accurate. If it says we'll have a 12kt thermal from 242deg arriving at 11:00AM we can generally count on it. If it says the wind will start from 210deg and gradually shift to 242deg, that is good enough so that we can play that shift in a race and gain a tactical advantage. It's good enough that we can select sails based on the PW models.

They are not 100% but by playing the odds we get pretty good results.

For passage planning we rely on the normal PW global models and GRIB files (GFS usually). Once away from the local areas the PWE and PWG don't add much. However, on the 350 mile crossing of the Sea of Cortez I have successfully played wind shifts predicted two days out by GFS and ECMWF.

For the West Coast of the United States we can select NOAA’s The High-Resolution Rapid Refresh (HRRR). The HRRR is a NOAA real-time 3-km resolution, hourly updated, cloud-resolving, convection-allowing atmospheric model, initialized by 3km grids with 3km radar assimilation. This was first introduced to us by Stan Honey and we found what he said about it was true: GRIB files based on HRRR are remarkably accurate.

[AllenRbrts]

Originally Posted by GRIT

Surely this can't be possible. Someone please tell me I'm reading this wrong.

Thanks.
Paul.

Try turning off motoring. You have asked the program to find the fastest way to get between two points, motoring might be the solution if it enabled and has parameters that enable that solution. I usually have motoring set for boat speed less than 4 kts and motoring speed is 5kts.

[s/v Jedi]

Originally Posted by BigBeakie

I had to speak with PredictWind in NZ about another matter, but FWIW I raised this issue of the PW weather routing giving upwind sailing angles that cannot be achieved.

With the proviso that for any specific instance or example they would have to see all the relevant data for that instance ( boat polars if available, date/time/latlong, GRIBs etc.) for review to give a specific answer.

The tech thought that probably what is going on is that people are expecting PW to give them a predicted route that is akin to a "roadmap", that is, a series of course segments that are more accurate than is possible. Basically if PW gives a course heading that is 15 degrees TWA,and your boat can't beat at 15 degrees (of course) then it is saying, "sail to windward as best you can".

It is an indicative heading, not a carved in stone reality.

I’m not sure on what they told you but I am sure that I never had anything like that happen with PredictWind. I’ve been using it in the West Caribbean, Bahamas and US east coast. Maybe it acts different for other meteo areas?

[donradcliffe]

I got the same routing results while using the offshore predict wind this summer coming back from Hawaii. As I remember, it had me at 15 degrees TWA in 12 knots of wind 1000 miles from land. I can't remember whether turning off the motoring option helped, but it seemed like a programming bug at the time.