Predict Wind

[boatman61]

Okay you knowledgeable App folks..
Just downloaded Predict Wind to my laptop and checking the N Wales area for next week a split screen pops up..
Now what confuses me is, one screen shows the wind coming from a Westerly direction yet the other shows the winds for the same day, time and position coming from an Easterly direction..
Is this just forecasters playing safe with this is what we think but then again it could be the opposite..

[Kijit]

Predict Wind lets you see different models side by side, so you are probably seeing two different predictions. When I use Predict Wind, I try to look at all the models side by side to see which have the best agreement. I bet that as the actual time gets closer, you will see those two start to converge.

[dgmultimedia]

Originally Posted by boatman61

Okay you knowledgeable App folks..

Just downloaded Predict Wind to my laptop and checking the N Wales area for next week a split screen pops up..

Now what confuses me is, one screen shows the wind coming from a Westerly direction yet the other shows the winds for the same day, time and position coming from an Easterly direction..

Is this just forecasters playing safe with this is what we think but then again it could be the opposite..

Zoom OUT until you can see the bigger picture - you will probably find the two different models show the low that is affecting the airflow in that region is located in a significantly different position hence opposite flow.

[robert sailor]

Predict wind has an algorithm that it uses for both the American and European grib models so on their site you can see GFS, Euro and the PWE and PEG the latter two are PW presentation with the algorithm working. In the end nothing beats a real meteorologists interpretation of these files. Like others have said as you get real close to the day they all tend to agree but what else would you expect.

[boatman61]

Thought I would give it a go as so many owners rave about it.. used it on my phone (no split screen) the last trip to Corfu and it was not much better than my other sites like PW..
Last 30hrs of the trip it was as wrong as the others..

[Kijit]

I would just listen to the weather forecast, then I started using SailFlow, then I tried PredictWind. Now I really don't ever know what is going on. I find it is like the old saying about watches. A guy with one watch knows what time it is. A guy with two is never sure...

[boatman61]

Originally Posted by Kijit

I would just listen to the weather forecast, then I started using SailFlow, then I tried PredictWind. Now I really don't ever know what is going on. I find it is like the old saying about watches. A guy with one watch knows what time it is. A guy with two is never sure...

[gaucho]

I find PredictWind etremely confusing. I use it, but it took me quite a while to understand it's ... nuances. This is my very humble opinion:

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

I believe PW needed a product to actually "separate it from the rest" and they tghought their own "propietary" models or interpretations based on GFS and ECMWF (hence Predict Wind GFS, Predict Wind ECMWF) is just ... balooney. I feel they are forcing something that ... what's the point? Are you really smarter than the models spit out by super computers? If it were a forecast that would be fifferent. But it is a propietary (hence nobody understand what ... it is) re-modeling of GFS and ECMWF.

So stick to the two big ones and use them as comparison to each other to figure out how close or far the model might be

- Use PW Offshore app, forget about the PW app. It's useless.

I only use offshore now. Why might you ask? Because it works, even if you get a glimpse of internet on some slow wifi it will let you select what to download to speed things up. Because the data stays with you even without internet. Because it doesn't need a "buoy" info location. You just get GFS/ECMWF offshore of a grid square. That's what you need. Microclimate modeling ... is far from reliable.

PW Offshore is a WHOLE new set of learning. It truly is a bad interface. But the functionality is fantastic.

[Paul L]

Originally Posted by gaucho

I find PredictWind etremely confusing. I use it, but it took me quite a while to understand it's ... nuances. This is my very humble opinion:

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

I believe PW needed a product to actually "separate it from the rest" and they tghought their own "propietary" models or interpretations based on GFS and ECMWF (hence Predict Wind GFS, Predict Wind ECMWF) is just ... balooney. I feel they are forcing something that ... what's the point? Are you really smarter than the models spit out by super computers? If it were a forecast that would be fifferent. But it is a propietary (hence nobody understand what ... it is) re-modeling of GFS and ECMWF.

So stick to the two big ones and use them as comparison to each other to figure out how close or far the model might be

- Use PW Offshore app, forget about the PW app. It's useless.

I only use offshore now. Why might you ask? Because it works, even if you get a glimpse of internet on some slow wifi it will let you select what to download to speed things up. Because the data stays with you even without internet. Because it doesn't need a "buoy" info location. You just get GFS/ECMWF offshore of a grid square. That's what you need. Microclimate modeling ... is far from reliable.

PW Offshore is a WHOLE new set of learning. It truly is a bad interface. But the functionality is fantastic.

Agreed, the two Predictwind models based on GFS and Euro models are just marketing BS. They add nothing to understanding potential passage weather while adding to user confusion.
When using the models its important to check the dates on them, as the GFS is updated more frequently than the Euro.

[boatman61]

Cheers guys..

[s/v Jedi]

If you use PW just for the bigger picture weather then the PW models do not add anything. If you want more localized weather or if you use PW for routing / departure planning, then the PW models add resolution. Models are digital but the weather is not, so resolution makes a lot of difference for deciding where to tack etc.

With human interpretation we do this in our mind (10 knots there, 20 knots here so halfway in between it'll be 15 knots or so) and the PW models allow their algorithms access to that same data.

Even if all they do is interpolation then it already helps for routing decisions. Just look at PWE as the hi-res Euro model and PWG as the hi-res GFS model.



Here in the Caribbean all the models look pretty good.

[mikereed100]

I have been using PW Offshore for the past year and really like it as it is simple and fast to use and I can get both GFS and ECMWF models. As others have pointed out, at 7-10 days the models can have a lot of variability, but tend to agree once they get to 2-3 days out. If they agree further out, that just increases my confidence in the prediction. I like that I can get pressure, wave and rain information as well.

I'm not sure what to think of the PW models as I have no idea what they are doing to tweak the data provided by NOAA and the Europeans and they are not saying. I have noticed some increased resolution, as Jedi points out, especially inshore, but they sometimes show some strange phenomena offshore, so I take them with a grain of salt.

The passage planning function is fairly useless as it seems to have an exaggerated idea of what my boat is capable of. On a recent passage from Ecuador to Easter Island it wanted me to take a tack southeast to Macchu Pichu. My boat sails well, but mountain ranges are still a challenge.

[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

[Paul L]

Originally Posted by s/v Jedi

If you use PW just for the bigger picture weather then the PW models do not add anything. If you want more localized weather or if you use PW for routing / departure planning, then the PW models add resolution. Models are digital but the weather is not, so resolution makes a lot of difference for deciding where to tack etc.

With human interpretation we do this in our mind (10 knots there, 20 knots here so halfway in between it'll be 15 knots or so) and the PW models allow their algorithms access to that same data.

Even if all they do is interpolation then it already helps for routing decisions. Just look at PWE as the hi-res Euro model and PWG as the hi-res GFS model.



Here in the Caribbean all the models look pretty good.

So averaging two points from a GFS model somehow gives you higher resolution? The two original points are a prediction of the future with a lot of potential error. Displaying them in higher resolution does not remove any error, it just makes it look pretty. The errors in direction and speed might be smaler, while the time of arrival is often significantly in error.
From a practical passage usage point of view we often don't even bother to download the highest GRIB resolution available, as the bigger picture is all that is needed to make passage decisions.

[s/v Jedi]

Originally Posted by Paul L

So averaging two points from a GFS model somehow gives you higher resolution? The two original points are a prediction of the future with a lot of potential error. Displaying them in higher resolution does not remove any error, it just makes it look pretty. The errors in direction and speed might be smaler, while the time of arrival is often significantly in error.
From a practical passage usage point of view we often don't even bother to download the highest GRIB resolution available, as the bigger picture is all that is needed to make passage decisions.

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?

I meet cruisers who have never seen a VMG read-out nor a lay lines display. They don't need it because they got there, right? I fully agree with that and I could do that if I want but instead I like to use those tools because I am proficient in their use and know it can easily take 10% or more off the passage time. Most will agree that getting there sailing faster, making better tacks, anticipating wind shifts etc. is a good thing. In order to do all that, you need a resolution better than 100nm