faq

Hydroscope differentiates itself from other data analysis tools in two main ways :

1. a data enrichment step that add weather and ship context to your data to get as full a picture as possible of your sailing sessions
2. state of the art, machine learning grade, mathematical tools to analyse large dataset and to generate the performance predictions you need measure and improve a yacht’s speed.

A detailed comparison with other existing tools can be found here.

Hydroscope works because it creates a complete picture of your sailing sessions, with an accurate description of the weather and the state your ship is in at any given point in time. With this data Hydroscope is able to understand what factors make your ship go fast or slow with a lot more detail. Being more accurate, means Hydroscope can help you make better informed decisions on your performance, much like a doctor using an MRI can take better decisions than one equipped with an X-ray machine.

Performance Prediction Error (PPE) is a measure of how well a Performance Prediction Tool \(P\) (such as a polar) matches the measured data \(D\) of a yacht. The prediction tool can predict the speed of the yacht from the data \(s_P(D)\) which can be compared to the actual speed of the boat measured in the data \(s_D(D)\). The PPE is then computed as the standard deviation of the prediction error :

Monitoring this quantity helps to rank Performance Prediction Tools, pick the most accurate and more importantly know how much to trust their predictions.

Hydroscope provides results in a variety of ways :

1. answers to specific questions you might ask yourself
2. result of numerical speed tests, helping you decide the best configuration/trim in a given condition which can be used to update your tuning guide.
3. aggregated data such as polars and wave tables which can be inputted into any routing software.
4. reports meant to help you track your progress, identify weak or unexplored regions in the ‘sailing enveloppe’
5. datasets you can explore by yourself or send to other parties

Hydroscope requires your inputs in two steps of its methodology :

1. to set up the ship model which requires :

   • drawings / 3D geometry of the yacht (to place the different parts within)
   • description of the systems and various control mechanisms
   • description of the different parts on board
   • sail plan description

2. in the data logging step which is required to save each sailing session and requires :

   • data describing the track of the yacht during the session + any other sensor logs you might have
   • the ship’s manifest as it left the dock (who was on board, how much fuel / fresh water in the tanks, what sails where on board?)
   • the ship’s briefing describing how you intended to trim the yacht during the session (what sails for each point of sail, appendage settings etc.)
   • deviations from the ship’s briefing on a segment per segment basis

For each of these steps the Hydroscope team will coordinate with your team to make these exchanges as quick and efficient as possible.

Hydroscope accepts almost all open formats both for the ship modelling and data logging steps.

A detailed list of supported formats in the data logging step is available here.

If your preferred format is not listed feel free to contact the team to find out if it can be supported.

Depending on the nature of the results of the analysis Hydroscope can export to a variety of formats :

• tabulated data :
  • spreadsheets : .csv, .xlsx, .xls, .ods
  • compressed datasets : .parquet
  • human-readable : .csv, .json
• polars : .pol
• text based results/reports :
  • email
  • slideshows : .ppt, .pdf, .html (interactive plots)
  • reports : .word, .pdf

If your preferred format is not listed feel free to contact the team to find out if it can be supported.

Unfortunately, Graphana and other web-services offer very poor download options, and unless you manually fill in the required ship context they don’t provide the data required by Hydroscope. For this reason Graphana isn’t supported, but if your data is already available in Graphana it won’t require any reformatting to be inputted to Hydroscope.

Yes, Hydroscope keeps developing new methods to analyse your data. Some are still experimental at this stage and therefore not listed on the dedicated page, but can still be offered to clients willing to accept potential errors in the analysis. The current experimental services are :

• timeseries analysis : Hydroscope with it’s advanced mathematical tools can extract insightful information from your temporal data (notably in the case of waves)
• local effects analysis : using your data and the weather context Hydroscope can identify areas where your yacht might have been impacted by local effects. This can help to analyse races.
• race analysis : by tracking lost time with our specific in-house tools, Hydroscope can evaluate how much faster you could have completed a race, and which decisions cost you the most.
• sailing session briefings : because it has access to a comprehensive weather context before you set sail and has developed the tools to extract the relevant information for a skipper Hydroscope can provide detailed briefings before you go sailing, examining the weather situation (including local effects) and how to best set up the boat for the expected weather, and what race strategy to adopt.
• sailing session debriefings : having access to a detailed dataset of your navigations and races, Hydroscope can assess your performance, as well as to monitor your strategy during races with our race analysis tools

Get in touch if you would like to know more about these tools.

Yes, Hydroscope can accept simulator data as input, if it is sufficiently annotated, allowing interesting comparisons between your data and your simulator results.

Hydroscope has taken steps to make its tools as robust and bug-free as possible. However, humans being humans, mistakes can always occur. If you believe you have spotted a mistake, or disagree with our conclusions, an investigation will be launched to analyse the source of the matter. If it is found that a mistake by Hydroscope is responsible for an issue, Hydroscope will :

• correct the mistake
• deliver the correct result

Yes logging data recorded before joining Hydroscope is encouraged. It’s a great way to accumulate the critical mass of data necessary to make data analysis insightful. However, it is best if there are good records of the yacht state during these sailing sessions (what sails, what crew, what bunkering, etc…) to keep a consistent data quality throughout the database.

Depending on your location, Hydroscope can access a variety of models. Get in touch to know which data would be downloaded in your case.

If you have access to privately owned weather data (cloudrun models for instance) it can be added to your Hydroscope analysis.

By all means yes! This enables Hydroscope to download weather predictions and measurements for your sailing sessions, rather than weather re-analysis which gives us a chance to get a more comprehensive picture of the weather situation.

The Hydroscope team doesn’t include meteorologists at this time. However, as people who have :

• studied the fundamental weather theory
• evolved in the offshore sailing industry
• processed large amounts of weather data
• sailed we feel comfortable handling meteorological data.

Besides, Hydroscope only compiles weather data generated by meteorologists, it does not generate it.

No, Hydroscope can process data from strain gauges and optical fibres like any other sensor output, but Hydroscope doesn’t have the required tools to give answers regarding the structure of the boat.

Not yet, but these are tools that are currently being developed. The project is to integrate these tools within the ship model so to make the comparison of VPP results and measured data seamless, opening the way for rigorous model validation.

Just contact us through your preferred channel with a list of modifications, and we’ll implement them. Make sure you do this rapidly after any modification of the yacht, to make sure the dataset remains as accurate as possible.

Hydroscope is a tool built to analyse the performance sailing vessels. The minimal requirement is that the track of the yacht is recorded. As long as this is the case any of the following boats can be processed in Hydroscope :

• offshore sailing (mini 6.50, figaro 3, class 40, ocean 50, IMOCA 60, ULTIM)
• dinghy and olympic classes (470, 49er, ILCA, Nacra, IQFOil, kitefoil)
• IRC (Club OD30, JPKs, X-Yacht, etc…)
• WASP (Wind Assisted Propulsion) ships
• cruising vessels : catamarans, etc…

We’re not really sure what that means. Here are some answers to help you make up your mind.

• Does Hydroscope use the latest mathematical/statistical tools from machine learning -> Yes.
• Does Hydroscope generate large amounts of data (as in Big Data)? -> No. Hydroscope can produce roughly 100[Mb] of compressed data per 24h sailing session. But this is a ridiculously small amount compared to what is usually referred to as Big Data.
• Does Hydroscope train neural networks based on this data? -> No.
• Does Hydroscope require large computers to run it’s analysis? -> No. Hydroscope outsources some of it’s computing to cloud services to avoid having to purchase new computers. Our biggest calculations can run on any modern computer with a recent GPU. Hydroscope was developed on 11 years old MacBook Pro.
• Does Hydroscope use an LLM to analyse your data? -> No.
• Does Hydroscope use LLMs for code generation? -> Yes, for a few well-defined and clearly bounded tasks, using libraries we are unfamiliar with or which lack documentation our team sometimes resolves to using LLMs to set up a working prototype rapidly. These small scripts are then integrated into the complete Hydroscope code manually, because the structure of this code has been completely designed by humans.

Yes, Hydroscope aims to be as transparent as possible with its customers. However, these details are the result of an important R&D effort on Hydroscope’s part and cannot be shared without certain precautions preventing them from getting into the hands of our competition.

If you wish to have access to this information you will have to sign a Non Disclosure Agreement (NDA), preventing you from disclosing or discussing this information to people outside said (NDA).

Yes, Hydroscope will answer any of your sollicitations within 1 (french) business day.

Hydroscope can be reached on any freely available platforms :

• email
• phone / SMS
• whatsapp
• telegram
• signal
• slack
• discord
• others…

If you would like us to join your internal private networks we can do that.

Your data is stored on our servers. You will be given an access link to the directory where all the data you have submitted is stored as well as all the analysis results.

To log a sailing session, contact Hydroscope through your preferred communication channel, to :

• send your data
• set up a video-conference (or in person meeting) so that we can fill in the ship context

This can be planned in advance of the sailing session.

To request an analysis of your data just contact Hydroscope through your preferred communication channel with a description of your query. If it is simple and clearly understood by our team we’ll first acknowledge the request, and then we’ll send the results your way as soon as they are available. If your request is more complex or not clearly understood by our team we’ll schedule a call with you to clarify the request.

Yes Hydroscope enjoys meeting its clients in person. But being based in Brest in Brittany means that for most clients, in person meetings requires significant travel times and costs. If you are willing to cover these costs we’d be happy to either greet you in Brest or come to your location. Otherwise, if we happen to travel near your location we’ll make sure to warn you in advance to try and set up a meeting.