Tutorial 3: Learning From A Recorded Session

Tutorial 3: Learning From A Recorded Session

Posted: Feb 2, 2018
Categories: Tutorials

Prerequisites: tutorials Getting Started and Recording a Session.

So we have gathered all this data, now what?

There are 3 situations that we used to design our workflows:

  1. On the boat. You have finished the race, and preparing for the next one. Usually there is not much time so the goal here is to get an idea about the general performance. This will help you with some points to focus on during the next race. The tools commonly used are the Race Report, the BSP% and VMG% indicators, the Navigator, data cells, the right bar and (to a lesser extent) tacks, jibes and speed segments.

  2. On the dock. After the last race of the day; in the harbour or back home. The work mostly consists of taking (accepting) speed segments, tacks and jibes, i.e. selecting the speed segments, tacks and jibes from the session that you think are relevant and could be used to optimise the polar. At the end of the race day it is often easier to determine relevance while the race is still fresh in one's mind.

  3. After the event, a.k.a. polar polishing workflow. Takes place a couple of times per racing season, usually at home. The work consists of evaluating previously taken speed segments, tacks and jibes and choosing whether to apply them to their respective polars or not.

    Often, speed segments, tacks and jibes that match the long-term trend are applied while anything that stands-out too much is regarded incidental and is discarded.

1. On the boat

As mentioned before, in this workflow we are mostly interested in speed performance, that is, BSP% , VMG% and speed segments. We approach the data top-down, that is, starting at the Race Report we work our way down towards data cell level.

In the Events Screen, please select the session that was just recorded to open it.

Race Report

  1. in the left bar, press the button to go to the Race Report screen.

  2. firstly, the rating gives a general indication of how well the race went. It is based on the average VMG% of the upwind and downwind legs and the average BSP% of the reaching legs. An 8 means that you managed to equal the polar performance.

  3. inspect the upwind, downwind and reaching sections and see if one of the sections show a performance that is significantly lower than the others. It is often best to focus on the weakest link of the last session and to try to improve that part in the next session. Note that the weakest section is not per se the section with the lowest VMG% value. You also need to factor in the time that was spent in that section. Thus, the VMG% value, weighted by the time spent in the section determines the impact of the section.

  4. it might be interesting to know if your performance increases or decreases with more wind. If such a trend is visible then it can likely be attributed to boat configuration, e.g. sail trim or mast settings. You can find this out by looking at the TWS row of the specific section: if the high performance column (> average VMG% + 5%) shows a TWS value that is significantly higher than the average VMG% and the time spent there is vast then you can conclude that performance increases with more wind. If the TWS value is significantly lower then your performance increases with less wind. The opposite reasoning holds for the low performance column (< average VMG% - 5%).

  5. the VMG% row of each section can be expanded to uncover the corresponding BSP and TWA values. These values can be significant. For instance, they might state that the low performance VMG% of the upwind section is the result of a high TWA since the BSP values of all columns are approximately identical, meaning you maintained good boat speed but sailed too low.

Session Screen

Let's continue our investigation at the next level of detail: the legs. Please exit the Race Report screen and look at the Navigator. Let's assume we want to focus on the upwind part of the race.

  1. press the UPWIND tab of the Navigator to open it. The upwind tab lists all the upwind legs of the race.

  2. first, you might want to quickly scan each leg's VMG% and TWS values to see if there is a relation between these numbers, possibly implying a boat configuration issue.

  3. next, look for legs that represent the majority of the VMG% values.

  4. for every interesting leg found in the previous point, examine its BSP% performance and VMG% performance. The easiest way to do this is by checking the color-coded map track and color-coded timeline. Specifically, in parts where the VMG% is low it can be worthwhile to check the BSP%: if the BSP% is good then this means that the angle (TWA) is bad. If the BSP% is bad then the low VMG% can be attributed to the combination of a bad angle and bad boat speed.

  5. select each interesting leg and drag the playhead to go through time, from the start of the leg to the end. While slowly moving through time, keep an eye on the values displayed by the data cells and the boat state displayed in the right bar. Try to explain to yourself why certain values are the way they are, also keeping in mind the race as you experienced it. Look for correlations between quantities, e.g. TWS v.s. BSP. This process of slowly replaying parts of the race, while witnessing the interrelation of quantities and while reasoning about the situation might reveal causes for the good or bad performance at hand.

  6. continue by having a quick glance at your maneuvers. Press the TACKS tab of the Navigator to get a list of all the session's tacks. Again, look for the common denominators. E.g. if the TWS is fairly constant and 4 out of 5 tacks show a meters-lost value of approximately 11m then either of these 4 tacks is representative and therefore interesting. You can select a tack to reveal its info popup on the timeline. Here you can see how your tack holds up compared to the target, i.e. the tack polar. Press the button to enter Ghost View, where your boat is set against a ghost boat representing the target, i.e. the speed, tack or jibe polar. This allows you to do a side-by-side comparison of both boats for every moment in the speed segment, tack or jibe. When you've finished processing all the tacks continue with the jibes.

  7. if a clear cause for bad performance has materialized from the data then try to "fix" these issues before the start of the next race session, e.g. by making the relevant changes to the boat configuration. If no clear cause was found then at least try to make an hypothesis based on your educated guess and make the proper boat configuration changes or battle plan changes to be able to test this hypothesis during your next race session. Later on, use the data of your next session to validate your hypotheses.

2. On the dock

As said before, this workflow is about taking speed segments, tacks and jibes, in other words, selecting those speed segments, tacks and jibes from the session that are representative of your performance and therefore are candidates for application to their respective polars.

Schools of thought

Before we begin taking speed segments, tacks and jibes, it is important to mention here that there are two "schools of thought" with respect to relevance:

  1. the one that believes that every representative speed segment, tack and jibe - good or bad - is relevant and should be taken and applied in the long-term. This approach will result in polars that reflect your average performance.

  2. the one that believes that only the representative speed segments, tacks and jibes that are roughly equal or better than the polar should be taken and applied in the long term. This approach results in polars that reflect your best performance.

School 1 has the benefit of more samples and thus is more reliable, especially in the short-term. Another benefit may be the fact that it is often easier to improve on the average than on the maximum (i.e. school 2) which may be better for morale.

Calibration issues

You should constantly be wary of boat sensor calibration issues. If your sensors are not calibrated properly and send misaligned data to annalisa then annalisa simply cannot generate sensible output. Proper calibration is the responsibility of the boat owner.

The following might indicate a calibration issue:

  • a bow dependency. E.g. if the VMG% color indication on the track is mainly orange on port-end side and blue on starboard-end side then this probably indicates a wind sensor misalignment. Also, AWA (Apparent Wind Angle) should be symmetrical around 0 degrees when comparing port-end sailing to starboard-end sailing. If this is not the case then the wind sensor might be the culprit. You can easily check this with a graph data cell.

  • BSP values that are significantly different from SOG values in stretches without any current. This might indicate a boat speed sensor issue.

Speed segments

Ok, now let's start with processing the speed segments. In the Events Screen select the first session of the day.

The idea is that we go through all the legs of the session and for each leg evaluate its speed segments. The easiest way to do this is to first select the UPWIND tab of the Navigator and then work through all the upwind legs. Then we repeat this for all the downwind legs and finally, all the reaching legs. Selecting a speed segment will show its info popup on the timeline. Once again, look for the common denominator speed segments: these are the ones to take.

Sometimes you run into a stretch of track that you believe to be relevant (based on hands-on experience) but where annalisa has not auto-placed a speed segment. In that case you can manually place a so called manual speed segment, as follows:

  1. in the right bar, press the button. The manual speed segment will be inserted at the playhead position.

Tacks and jibes

The workflow for taking tacks and jibes is identical to the speed segments workflow. You can use the Navigator's TACK and JIBES tabs to step through all the tacks and jibes. If you select a tack or jibe then the info popup on the timeline offers detailed information.

Note that if the info pop up displays a flashing button then there are warnings to consider. Press the button to view these warnings. Here are some situations for which warnings will be given:

  • if the TWS (True Wind Speed) before the tacking point is significantly higher than its value after the tacking point then the BSP (boat speed) after the tacking point is in large part caused by the boat's momentum, not by human effort.

  • if the meters-lost value is positive. This would imply that the maneuver did not have a negative impact on your speed, which is odd. If this occurs then it is likely the result of external factors like wind shear.

Please be mindful of these warnings before taking the speed segment, tack or jibe.

3. After the event (long term)

As mentioned, the work here consists of evaluating previously taken speed segments, tacks and jibes and choosing whether to apply them to their respective polars or not.

This workflow takes place solely in the Polar Screen. To get there: in the Events Screen, press the button in the top-left corner of the screen.

Speed polar and crossover

All taken speed segments end up on the timeline covering the bottom part of the screen. The speed segments are sorted chronologically and segmented per event. The timeline displays time horizontally and % of the initial VMG vertically.

The idea is to go through all the unaccepted speed segments and consider them for application in the speed polar.

If you select a speed segment both the polar chart, the timeline and the polar table will give a preview of the impact of application. The polar chart and the timeline indicate this by means of a white line extending from the TWS line that is influenced most. The polar table shows by color which numbers will be influenced, where green indicates a value increase and red indicates a value decrease.

When you select a candidate speed segment the polar chart will show all applied speed segments that affect the corresponding TWS line. Use this to determine if the candidate's performance is confirmed by speed segments that are already applied.

In other words, if the candidate speed segment becomes part of a cloud of previously applied speed segments then you should probably apply it. If it is an outlier then you should not yet apply but wait for similar speed segments to emerge in the future and apply it then, or, if nothing similar emerges, untake the speed segment altogether.

Be aware that an applied speed segment might also influence the sail crossover: if the speed segment's BSP% is higher than the crossover's value for the corresponding TWS and TWA then the sails that were used during the speed segment replace any sails currently set in the crossover for that TWS and TWA. So essentially the crossover reflects your best performance.

The timeline graph gives a nice view of the development of your performance over time. In the ideal world the TWS lines would keep on rising or at least would never drop (except perhaps for an initial drop resulting from an imported polar that was too optimistic).

Tacks and jibes

The workflow for tacks and jibes is identical to the workflow for speed segments.

The following heuristic might be helpful: generally you would expect tacks or jibes in mid range wind speeds (8 kn to 16 kn) to have lower meters-lost values than tacks or jibes in the low range (< 8 kn) and high range wind speeds ( > 16 kn).

That concludes this tutorial!

© annalisa B.V.