Flight Scoring

Where Is The Relevance?

Now, let’s say that my score of 15 was in basketball. Even here, there are lot of other factors to take into account to assess my offensive ability. A way to provide more meaningful statistics would be to increase the sample size and take my points tally from many games. However, even this wouldn’t tell the whole story as there are so many other parts to offense and to the sport as a whole. For example, we haven’t accounted for any assists or steals, and personal fouls haven’t counted against me in any way.

There are so many separate parts of basketball, which contribute to an individual’s overall performance, and it’s the same for flight safety. However, a key difference to point out here is the amount of data available from one flight compared to one basketball game. We can count the number of blocks a player makes, or the point tally, and we can offer a ratio of complete to incomplete passes. Whereas data coming from aircraft in 2020 is so wide-ranging that, in equivalent terms, we would be measuring the peak height reached for each block, the angle of hand and wrist, and the force exerted for each steal, or the angular velocity of the ball as it contacts the board for each lay-up. In flight data monitoring each measurement or calculation is designed to improve safety or efficiency. How long did the aircraft take to de-rotate on landing? How far was the aircraft above the target height? How hard was the landing? Because of the amount of individual measures taken it is not possible for a person to review the performance of every flight for each measure. This is where event monitoring comes in.

Aviation Significance

An airline will set limits on certain measurements, such as wanting an alert (event) to trigger if a landing was harder than a given g-force. This serves multiple purposes, not least allowing appropriate attention to be afforded to the extreme exceedances. Any modern Flight Data Analysis system with flight data monitoring allows airlines to pick what they want to monitor and the thresholds at which they want their alerts to trigger. For an individual airline, this system works very well. With good thresholds they won’t suffer from ‘nuisance’ alerts and will know that when an alert triggers, it needs attention. A good flight data monitoring system will also allow the airline to dig into the data surrounding that alert, similar to reviewing game footage around points conceded to find the weak spots.

Conclusion

So back to my basketball - I have a few friends and each of us plays for a different team. The points we concede are not identical, we are at different teams, and different teams have different weaknesses and strengths. Where one team may concede a lot from lost possession, another might be particularly poor at defending from range. Likewise in aviation, one airline may trigger a lot of taxi speed events whereas another may have issues with automation mismanagement. Imagine we are primarily interested in comparing our bad basketball games, so have agreed to count our number of bad games per month. At the end of the season, we could get together and compile our statistics, but there are problems with this approach, primarily, what counts as a bad game? If we all define ‘bad’ differently, comparing our ‘bad’ game count is meaningless, while I might be happy with my performance in a game, one of my friends might not.

Can this be solved? Yes, to a point. Basketball leagues use different measures of player performance, based on equations taking into account steals, points scored, free throws made etc. Using these ‘end scores’ for an individual player for an individual match, and putting the statistics together over the course of a season, a player’s overall ability can be approximated. The same can be done with flight data, albeit taking into account many more variables. One airline’s event count or event rate with another isn’t a valid comparison as different events and thresholds might be in place. However, the aggregated key point values from each flight ARE directly comparable.

In some leagues this is referred to as an ‘efficiency rating’. A player’s efficiency rating is calculated from an equation taking into account multiple factors of the game, with different parts afforded different weightings. These efficiency ratings are then averaged out over a season. A similar thing can be done with flight data. An equation, albeit more complex, can be used to approximate an individual pilot’s, or a whole airline’s performance. Breaking this down into separate phases, it may turn out that some crew members or operators are performing abnormally well or poorly. This type of approach can also highlight risks at airports that might have ‘flown under the radar’ until now. An advanced flight data monitoring programme will already be taking hundreds of measures per flight the debate is in determining which measures to use.