Some analysis on fatalities data

[Muffie 0]

The attached is my analysis of fatal incidents for the U.S. from 2004-2011. Unfortunately I don't have good canopy size/WL data yet, so it's light on that data, but baby steps.

So, (a) this isn't perfect, (b) it's just my attempt to identify some trends/issues/lessons that can be pulled from the fatalities data, (c) you may not agree with my conclusions, and (d) some of you may think the conclusions are obvious. But, I think in continuation of other discussions, we need to use the data we have and start gathering better data to identify issues and their causes that we can then potentially address. This is just a first go at doing that.

Here are the conclusions from the attached:
• The causes of fatalities do differ by experience level and also by gender
• In terms of malfunctions, jumpers with under 100 jumps are most likely to die due to a failure to execute their EPs and jumpers with 101-500 jumps are most likely to die due to a failure to complete their EPs. After that point it seems that most jumpers can handle a standard malfunction, but that a spinning malfunction may still cause a jumper to lose altitude awareness and fail to cutaway and deploy a reserve timely.
• Fatalities from HP landings both on an individual basis and involving canopy collisions appear to be most prevalent for jumpers with between 1,001 and 2,500 jumps.
• At the 2,500 jump mark health issues such as heart attacks and strokes start to appear for male jumpers, particularly those over 40 years old.

Sorry it's a Word doc, but the PDF was too large to upload.

[FlyingRadio 0]

Quote

Sorry it's a Word doc, but the PDF was too large to upload.

I fixed that for you; the attached PDF was exported from your posted word doc and is slightly smaller than the source document. (For the curious I used OpenOffice to open the original document and the included PDF export feature to generate this one.)

And the obvious disclaimer: this is Muffie's work, not mine.

--
Radio

[Muffie 0]

Awesome, thanks!

[-ftp- 0]

Nice work!

The only thing I can reccomend, is to try to not use words like "only" or anything else that people might view as a slanted or biased opinion.

Other than that, that is a really nice breakdown.

[topdocker 0]

I think the collisions should be seperate from CRW wraps. Opening and colliding or colliding low in the landing area is different then having a CRW formation funnel around you.

Granted the outcome may be the same, but the root causes are different.

Other than that, it's a very good analysis and mirrors Bryan Burke's report.

top

Jump more, post less!

[manseman 0]

I like the idea behind this, but I really wonder if there's much statistical significance for some of the analysis given such low numbers? Sure, landings are obviously more likely to kill you than getting hit by the plane but I would be very careful when attempting to draw conclusions from numbers such as 2 vs. 4 for events that happen very, very rarely.

[DarrellM5 0]

Nice work, but definitely not the thread I should be reading since my AFF training starts next weekend.

[mjosparky 3]

It is evident that you put a great deal of work into this. The one change I would like to see is Canopy Collisions on landing as a separate category. It is such a hot topic at the moment that I would like to see its stand alone numbers.

Also you mentioned 2 fatalities where the AAD’s were on but failed to fire. Do you have a link to these reports?
You also one failed to cut away due to a custom rig. If I could get a link to that please.

Good job,

Sparky

My idea of a fair fight is clubbing baby seals

[Muffie 0]

Quote

Nice work, but definitely not the thread I should be reading since my AFF training starts next weekend.

Yeah, not sure I would've wanted to know too much about skydiving fatalities before I did my first jump. I fortunately didn't really start paying attention to the incidents until I had about 30 jumps and was already hooked on skydiving so it became a matter of "I'm going to do this, what can I learn from other's mistakes or misfortunes that will give me a better chance of surviving it."

For you, as someone who is just about to do AFF, I would think of it this way: Now you know that things can go wrong from jump one and you know to pay real attention to the emergency procedures and off landing briefings you're going to get (and learn that PLF position). Now, if something does go wrong, knowing that it could will hopefully help you handle it better because you'll think "right, I knew this could happen so.." as opposed to "WTF??? Wait, what did they say - what am I supposed to do?"

Enjoy AFF! Skydiving is an amazing experience, just respect it.

[Muffie 0]

Quote

It is evident that you put a great deal of work into this. The one change I would like to see is Canopy Collisions on landing as a separate category. It is such a hot topic at the moment that I would like to see its stand alone numbers.

Also you mentioned 2 fatalities where the AAD’s were on but failed to fire. Do you have a link to these reports?
You also one failed to cut away due to a custom rig. If I could get a link to that please.

Good job,

Sparky

Sparky - I'll break down the canopy collision data and post it for you. My sub-categories under that high-level one are high collision, low collision, freefall collision, and CRW.

In the meantime:

The two AADs on but failed to fire were:
10/16/2004 James B (an Astra)
http://www.dropzone.com/cgi-bin/forum/gforum.cgi?post=1301903#1301903
3/5/2005 Susan S (set at wrong altitude)
http://www.dropzone.com/cgi-bin/forum/gforum.cgi?post=1515845#1515845

The custom rig:
7/4/2008 Keith C (custom removable deployment system)
http://www.dropzone.com/cgi-bin/forum/gforum.cgi?post=3258666#3258666

Muffie

[airtwardo 6]

So 50% of all the male fatalities in that time period, are under open parachutes...

Granted in one light, there are only two options regarding fatal jumps, parachute open or closed....but still that statistic is IMHO quite telling as to the state of the sport today.

~ If you choke a Smurf, what color does it turn? ~

[Divalent 89]

Quote

Quote

Nice work, but definitely not the thread I should be reading since my AFF training starts next weekend.

Yeah, not sure I would've wanted to know too much about skydiving fatalities before I did my first jump.

OTOH, I did decided to do AFF *after* doing a similar detailed analysis. After looking at the breakdown, it is my hope that I can be nearly 10 times safer than average once I get farther beyond my student status; and I think this is a realistic goal. This analysis helped me see where common errors were made, and helped me ensure I was picking up the skills and training that would allow me to minimize my risk.

1. Don't do low turns. (So I spent a lot of time under canopy on my solos practicing flat turns and braked turns.)
2. Don't jump in bad winds.
3. Don't swoop.
4. Stay the hell away from people that are swooping.
5. Be vigilant under canopy for other traffic.
6. Gear checks.
7. Don't do rigging in the air. (Altitude awareness if the main is not perfect)
8. Think about everything you do, in free fall and under canopy.
9. Be an expert PLF'r. (I also don't want titanium hardware holding me up.)
10. Jump with an AAD.

(and of course, don't commit suicide or do stupid stuff like climbing out of my rig in FF)

I'm still a novice and continue to make mistakes, but when I make them, I try to recognize it an learn the lesson. Hopefully I will progressively rely more on my own skill and capabilities, and less on luck, for earning the opportunity to walk back to the hanger under my own power.

[Muffie 0]

Quote

It is evident that you put a great deal of work into this. The one change I would like to see is Canopy Collisions on landing as a separate category. It is such a hot topic at the moment that I would like to see its stand alone numbers.

Sparky,
Stand alone numbers for canopy collisions:
2-25 - - - - - - - 4
26-100 - - - - - - 0
101-250 - - - - - 1
251-500 - - - - - 2
501-1,000 - - - - 2
1,001-2,500 - - - 5
2,501-5,000 - - - 2
5,001-10,000 - - - 2
10,000+ - - - - - - 1
UNK - - - - - - - - - 2
Low - - - - - - - - - 1

I bolded the two experience levels that seem to stand out even with just a small sample size.

For the 2-25 jumps this is a bit of color: 3 incidents - 1 two students with abt 20 jumps ea. unknown canopies; 1 on a Nav 280 hit by Katana 120; 1 on a Nav 260 collided with Paratech GmbH 340 that was camo colors (2 double fatalities, 1 single fatality)

For the 1,001 to 2,500 jumps: 4 incidents - 1 two swoopers (one female) collided on Velocity 79 and Velocity 90; 1 swooper on JVX 87 coming out of turn hit by Sabre2 170 on straight approach; 1 on Spectre 190 hit by other turning to final on Pilot 190; 1 unknown canopy collided one on base and one 90 to final (3 double fatalities, 1 single fatality)

[kallend 1,623]

Quote

Quote

It is evident that you put a great deal of work into this. The one change I would like to see is Canopy Collisions on landing as a separate category. It is such a hot topic at the moment that I would like to see its stand alone numbers.

Sparky,
Stand alone numbers for canopy collisions:
2-25 - - - - - - - 4
26-100 - - - - - - 0
101-250 - - - - - 1
251-500 - - - - - 2
501-1,000 - - - - 2
1,001-2,500 - - - 5
2,501-5,000 - - - 2
5,001-10,000 - - - 2
10,000+ - - - - - - 1
UNK - - - - - - - - - 2
Low - - - - - - - - - 1

I bolded the two experience levels that seem to stand out even with just a small sample size.

For the 2-25 jumps this is a bit of color: 3 incidents - 1 two students with abt 20 jumps ea. unknown canopies; 1 on a Nav 280 hit by Katana 120; 1 on a Nav 260 collided with Paratech GmbH 340 that was camo colors (2 double fatalities, 1 single fatality)

For the 1,001 to 2,500 jumps: 4 incidents - 1 two swoopers (one female) collided on Velocity 79 and Velocity 90; 1 swooper on JVX 87 coming out of turn hit by Sabre2 170 on straight approach; 1 on Spectre 190 hit by other turning to final on Pilot 190; 1 unknown canopy collided one on base and one 90 to final (3 double fatalities, 1 single fatality)

Unequal sized bins make the comparisons somewhat suspect.

...

The only sure way to survive a canopy collision is not to have one.

[Muffie 0]

Valid point. I was trying to bucket things by experience levels more than absolute jump numbers - 2- 25 being a proxy for pre-A license, 25-100 being how long it takes before the 100 jump wonder phenom kicks in, etc. I could argue that the difference between someone's 10th and 100th jump is very different than the difference between their 1010th and 1100th. If we ever do go down the path of really digging into all the data we have then you'd get a group of subject matter experts together and debate what's the best way to dice and slice the info. And then you'd actually slice and dice it the different ways to see what brings you the most value.

Here are the raw jump numbers: 14, 15, 20, 20, 125, 400, 445, 545, 1000, 1398, 1500, 1604, 1960, 3500, 4200, 5600, 8200, and 17,000 with one "couple thousand".

If you were to bucket this into 250 jumps each you'd still see a clear issue with newer jumpers (5 fatalities under 250 jumps, 2 more between 251 and 500) and then a bit of a gap untl 1250+ (just one in the 751-1000 range and two with zero) where you then get a bit of a cluster (2 in the 1251-1500 range, 1 in the 1501-1750 range, 1 in the 1751-2000 range, and 1 "couple thousand") before you again see big gaps between each fatality (next one is in the 3,251 to 3500 range).

I think the benefit to reviews like this is that they may not be definitive but they can point you in a direction. So, to me, this says for the new jumpers (a) more focus on canopy control and/or (b) give newer jumpers a wider berth (separate landing areas, separate passes...) and for more experienced jumpers if you're going to implement a canopy control test requirement or training requirement that it should be around the 1,250 jump mark. I'd argue that that's a lot later than most people would have set the mark for the experienced jumpers.