Isn't it time you went in?

[skypuppy 1]

Blahr says
'A good pack job is a good pack job whether the packer has 20 or 2000 jumps. That same good pack job can still produce a malfunction for no good reason and its not more likely on the 20 pack job guy than it is on the 2000 pack job guy assuming good pack jobs all around'


I don't believe this... If you inspecct your equipment before packing, pack properly and deploy stable, there is no reason for you to have a mal - you may get a pc hesitation or an off-heading, but a catastrophic failure is not in the cards without some external event (ie collision, high-speed opening, etc.). What your saying is much more likely on a BASE jump where the opening (sans slider and bag) is less structured and exits and object closeness are
always wild cards, but on a skydive where you follow the rules, you should have no more than a low(er) opening... (This is assuming you don't jump experimental equipment - like the tandem rigs produced back in the mid '80's, when people didn't necessarily understand possible design flaws - or mini risers - or slinks (incorrectly positioned) - or overstressed rapide links, etc., most of which should be caught during an inspection anyway).

As to the odds between a 30,000 jump-wonder and a 300-jump wonder, they're only the same if they use the same packer...

Similarly, you can only have a double mal if you first have a main malfunction - ergo if you don't have a mal on you're main you'll never have a double mal... This is why I pack my own...

If some old guy can do it then obviously it can't be very extreme. Otherwise he'd already be dead.
Bruce McConkey 'I thought we were gonna die, and I couldn't think of anyone

[GravityGirl 0]

You are leaving out too many variables. Jump numbers is just one variable. Gear. Gear maintenance. Conditions. Etc. Etc. Etc.

~~~~~~~~~~~~~~~~~
Peace and Blue Skies!
Bonnie ==>Gravity Gear!

[champu 1]

Quote

When they go for the world record 25,000 way there is a high probablity that one person will have a dmal.

Maybe, maybe not, but I'd say the probability of about 25000 canopy collisions is pretty high.

[skydivexxl 0]

I used to live in Vegas...

The odds are NEVER fair...

Blog Clicky

[Liemberg 0]

Quote

But mathematical law dictates that the more jumps you make, the closer you will reach the 'average' number of jumps needed for a double-mal, and the more likely it is to happen.

Dices and coins have no memory
Parachutes have no memory
Humans do.
I - myself - inspected and packed my main AND my reserve.
I sincerely hope I remembered how to do that, when I was doing it.
Actually if I didn't have some rudimentary knowledge about statistics I would worry like you do - just picture the parachute keeping track and making the odds change like in a blackjack game where after a few rounds certain cards were still in the deck and other cards had disappeared, up to a point where either blackjack or double mal were 'near inevitability'.

Besides, what you describe as a 'double mal' in terms of mathematical probability isn't a dice that is thrown or a coin that is flipped - it has steering lines that show wear, slider grommets that become dented and more important: a user that isn't always as smart as he should be. In the beginning of my skydiving career I had several freefall jumps where the opening sequence started with me falling on my back and a main-pilotchute appearing between my legs.

That must have influenced the odds, don't you think?

Luckily I don't do that anymore...
You on the other hand?

Unfamiliar DZ, unknown jumpers about and probably more people in the air at the same time than you are used to, to much to drink in the Captains Cabin - you are taking a significantly greater risk, jumping in Empuria...

"Whoever in discussion adduces authority uses not intellect but memory." - Leonardo da Vinci
A thousand words...

[ripper0289 0]

I agree, statistics aren't cumulative.

[skymick 0]

Quote

You are leaving out too many variables. Jump numbers is just one variable. Gear. Gear maintenance. Conditions. Etc. Etc. Etc.

I think you missed the point of Skylark's post, he is no specifically talking about skydiving in real life. Its a purely mathematical question. By this I mean you could describe it by saying 'if you take part in activity X 25,000 (or whatever number) times how likely is it that a certain event Y can occur?'

BTW: His theory on it is wrong, but most people have gotten their explanations right so I wont btoher saying why its wrong

[fool 0]

well, purely mathmatically speaking then, on every jump, there is a 1 in 25000 chance of having a double mal. I don't think this is an accurate stat, but that would be the math. theoretically speaking, the more experience you have, the more likely it is that you will experience a greater amount of things (including a double mal, but you can't in any way put a number on that.

S.E.X. party #1

"Life's journey is not to arrive at the grave safely in a well preserved body, but rather to skid in sideways, totally worn out, shouting "f*#k, what a ride".

[skymick 0]

Quote

I don't think this is an accurate stat, but that would be the math

Your spot on there, in real life theer are so many variables that come into play that affect the results, so much so that you could'nt come up with an accurate figure but for a purely theoretical analysis (i.e. not taking into account external variables) then a number like that is fine.

[Fab 0]

skylark.

You base your opinion on the fact that the chance of an event will be based on the outcome of prior events in the past. That is your flaw in your reasoning. You say...if you have 20000 good jumps, the 20001 wil probably be bad. Stochastic processes doesn't work like that. True that there may be 1 fatal jump out of 25000 (average) but you shouldn;t feel worried if you approch the 25000. You have as much chance of a double mal at jump 1 then at jump 25000. The idea is that you have a chance of a mal of 1/25000 per jump....Therefor your chance to have a good jump is 24999/25000. Which is a pretty good chance that you will survive your 25000'th jump...

_______________________________________

[rehmwa 2]

Quote

When they go for the world record 25,000 way there is a high probablity that one person will have a dmal.

We have a winner for the best answer and clearest understanding above from newsstand!!! But there is an equal chance on this dive for no dmals also.

(Simple calcs below: there are also probabilities for 2 people having a dmal on this dive, 3, 4, 5, etc all the way to all participants (25K) having dmals on the same dive. They all add up to one. But the highest prob is for 0 or 1 person. Very VERY long odds, the odds or 0 or 1 'successes' approach each other) On that dive, there is 37% chance of 0 or 1 dmal, 18% chance of 2 dmals, etc. and very insignificant chance for more than 7 dmals......

Probability Density Function
Binomial with n = 25000 and p = 0.000040000
x P( X = x )
0.00 0.3679
1.00 0.3679
2.00 0.1839
3.00 0.0613
4.00 0.0153
5.00 0.0031
6.00 0.0005
7.00 0.0001
8.00 0.0000
9.00 0.0000
10.00 0.0000

thru 25,000 the probs are 0.0000 (insiginificant)

...
Driving is a one dimensional activity - a monkey can do it - being proud of your driving abilities is like being proud of being able to put on pants

[kallend 1,620]

Incorrect!

...

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

[whatever 0]

ok, ok, so I'm not a professional statistician and I expect to get it from all sides for saying this, but please take the effort to correct me specifically on where I'm going wrong, if I am.....

Please look at this purely in a statistical and mathematic manner, I'm not arguing the variables involded in mal probability, just some probability calculations.

attached please find an Excel spreadsheet with some probability calculations and a graph

the case we're looking at only has two possible outcomes for any one jump and they are:

-you do not have a double mal, therefore we assume you live

-you have a double mal, BSBD

for my calculations I have assumed the probability of a double mal is 1 in 25000 (this is just a number used for demonstarting my calculations, as some other folks have used it too)

this means the probability of not having a double mal on any one jump is 24999/25000, no matter what the jump number is on that jump...

now, what is the probability of not having a double mal for two jumps in a row?
from what I understand of statistics, the probability of two succesive events, both with known probabilities, having the desired outcomes, is the product of the probabilities of those outcomes, i.e.:

probability of not having a double mal on one jump = 24999/25000

probability of not having a double mal on the next jump = 24999/25000

(so you're looking good there...)

BUT, the probability of doing two jumps and not having a double mal on either = 24999/25000 * 24999/25000

which is still a pretty damn high number....

if you follow the spreadsheet down you will see that at 1000 jumps your probability of having had a double mal is just under 4% (this sounds way out of whack with observed statistics, so the real probability of a double mal is likely less than 1/25000),
move along to 10000 jumps and now the probability of having had that double mal is 33%,
at 20000 it's 55% (note that it didn't just double, as we are not adding probabilities, we are multilying them), at 50000 it's 86% and you've hit 98% percent when you've done 100000 jumps....

looking at the graph you can see that the probability of not having that double mal starts at 1 (for zero jumps) and decreases with every jump you do and at about 150000 it starts to approximate zero, this makes sense, as you are rolling a many-sided dice (think more than 25000 sides, if you can picture that ) and only one of those outcomes are bad, but, if you keep rolling that dice, no matter how low the probability of that one outcome, eventually it will happen.....

I've included a second worksheet with the same for a coin toss, head or tails and what is the probability of getting tails for every single coin toss, for x number of coin tosses in a row....

if you disagree with my calculations but cannot explain clearly why, a simple test would be to do the coin tosses and compare your observed results to my calculated probabilities...
go ahead, see how many tries it takes you to get a result of 10 tails in a row.....

sorry about the long post, couldn't do clear and short...

soon to be gone

[Hooknswoop 19]

But if you don't use your reserve in 25,000 jumps, the odds of a double malfunction are zero. Let's say reserves do malfunction once for every 25,000 deployments. That does not mean there is a 1 in 25,000 chance of having a double malfunction every jump, only on jumps where you use your reserve.

Derek

[rehmwa 2]

Derek, that's not what he means here. It's the odds of a naturally occurring double. Not for each chute. (I think 1 in 25,000 might be too low (if a main malfunctions once ever 500 jumps, then the reserve only get 50 'opportunities' to fail) and that's what likely bugging us, but that's the hypothetical odds and it's just as good as any pick).

What he's done is shown the odds of running the configuration in series. It's a good analysis. But it is only applicable when used to analyze the possibilty of the double mal BEFORE jump number 1.

Again, you can't store up luck or bad luck.

So, a little bit of work to just continue the exact same discussion...........

And note my analysis also shows an equal probability of zero malfunctions in a 25,000 jump series or, even better, the record 25000-way jump. (this is also a misuse, but a minor one and maybe that'll help someone with the 'gut feel' part of it.)

...
Driving is a one dimensional activity - a monkey can do it - being proud of your driving abilities is like being proud of being able to put on pants

[whatever 0]

Agreed, if you don't use your reserve the odds of a double mal are zero, but the fictional number I used for calculations would have that factored in, that is, the odds of having a main mal and then having a reserve mal too, are 1 in 25000.

later in my post I say that looks like too high a probability based on my calculations

if you know the probability of main mals, and the probability of reserve mals, the probability of a double mal would be the product of the two:

p(double mal) = p(main mal) x p(reserve mal)

I was using 1 in 25000 as p(double mal)

soon to be gone

[whatever 0]

Quote

that dive, there is 37% chance of 0 or 1 dmal, 18% chance of 2 dmals, etc. and very insignificant chance for more than 7 dmals......

ok, how is 25000 different jumpers doing 1 jump at the same time and there being a probability of 37% for 1 double mal,

different to:

one jumper doing 25000 jumps and getting the same probability, as per my calculations?


our calculations add up, I do not understand why you say I calculated a probability that is only valid before jump number one, before jump number one the probability of a double mal is always zero.

surely you agree that for an event with any given probability, the more you repeat it, the more likely any one possible outcome is.... (for events with outcomes that have constant probabilities for any one given occurence of that event)

as per flipping the coin, the more you do it, the more likely you are to get the result you're looking for....

how is this different to the case of the double mal's probability?

please explain

thanks

soon to be gone

[AndyMan 7]

I would love to know why you think p(double mal) is roughly 1 in 25,000.

If p(main) is 1 in 300, and p(reserve) is also 1 in 300, then p(double) is 1 in 90,000.

Given that, I think p(main) at 300 is a very low estimate, 500 is probably more accurate and p(reserve) is completely unknown, but I suspect it is somewhere in excess of 3000. By these estimates, the p(double) is 1 in 1,500,000 - which means its not a risk I'm loosing sleep over.

I just don't understand why you're wasting so much brain power on this. There simply isn't enough data to infer anything from the discussion. More importantly, there are risks in skydiving that you CAN control, I'd much prefer you to be worrying about those.

_Am

__

You put the fun in "funnel" - craichead.

[rehmwa 2]

Quote

ok, how is 25000 different jumpers doing 1 jump at the same time and there being a probability of 37% for 1 double mal,

different to:

one jumper doing 25000 jumps and getting the same probability, as per my calculations?

same - I said it was a good analysis and example of how to get to the same probs in longhand - binomial math will make it easier for you to do the same thing

our calculations add up, I do not understand why you say I calculated a probability that is only valid before jump number one, before jump number one the probability of a double mal is always zero.

no, what I meant, and sorry for the confusion, is you can't go and retroactively apply stats after the series is started - it's really what everyone has been saying the whole time

surely you agree that for an event with any given probability, the more you repeat it, the more likely any one possible outcome is.... (for events with outcomes that have constant probabilities for any one given occurence of that event)

I don't agree - it's the whole point of the thread. Each event is an independent occurrence and the likelihood depends only upon the probability of the occurrence - which you already note is "constant". There is a whole other branch of stats for dependent occurences and your spreadsheet (and my minitab run) does not apply those tools and it also isn't applicable to this example anyway

as per flipping the coin, the more you do it, the more likely you are to get the result you're looking for....

same response as above

how is this different to the case of the double mal's probability?

it's not, you are mixing up independent occurences with dependence based stats

please explain hope this helps

thanks

's ok

...
Driving is a one dimensional activity - a monkey can do it - being proud of your driving abilities is like being proud of being able to put on pants

[rehmwa 2]

Andyman - so right, luckily, this doesn't require much brainpower. Right now, I just don't want a brother skydiver or two going into stats class somewhere and completely missing the point. This is very fundamental and has moved into a stats discussion and not a skydiving discussion. In any case, Bill Dause and Don Kellner are likely no less safe today then they were a couple days ago (which for those guys is likely another 50 jumps

...
Driving is a one dimensional activity - a monkey can do it - being proud of your driving abilities is like being proud of being able to put on pants

[whatever 0]

Quote

Please look at this purely in a statistical and mathematic manner, I'm not arguing the variables involded in mal probability, just some probability calculations.

dude,
that's a quote from my post, its right near the top too......

like I said, I'm not arguing/debating/analyzing the variables, or probabilities of mals and reserve mals, I'm trying to determine if I understand probability at all...

soon to be gone

[whatever 0]

Quote

same - I said it was a good analysis and example of how to get to the same probs in longhand - binomial math will make it easier for you to do the same thing

so you agree that if I go out and do 25000 jumps the probability of me having a double mal sometime during those 25000 jumps is 37% based on the entirely fictional probabilty of 1 in 25000 for a double mal on any one of those jumps individually?

I don't agree - it's the whole point of the thread. Each event is an independent occurrence and the likelihood depends only upon the probability of the occurrence - which you already note is "constant". There is a whole other branch of stats for dependent occurences and your spreadsheet (and my minitab run) does not apply those tools and it also isn't applicable to this example anyway

my bad, I phrased that very stoopidly

I know the probability on any one jump will stay the same, regardless of how many you have done before it or are going to do after it.

If it's not applicable, please explain why, and please start with the coin toss example and show me how it's not any more likely to get one tails in 10 tosses than 1 tail in 1 toss


it's not, you are mixing up independent occurences with dependence based stats

where am I mixing it up?

hope this helps

it does a bit, but I still don't see the difference between the double mal stats problem and flipping a magic coin that has one side far more likely to land up than the other...

for any one flip the odds are the same...

but if you keep flipping, wanting to always get the more probable outcome, sooner or later you'll get the less probable outcome, it's just a matter of time...

soon to be gone

• Quote

[benny 0]

Hmm, well, we've all thoroughly ripped the lark's original question to shreds and gone on to some rather interesting analyses. I have another thing to thow in here. If we look at it from a purely gaming point of view, the odds of double mal are much higher. a probabity that an event will occur in its purest theoretical sense is simply that event divided by the total number of possible outcomes (in the case of a coin 1:2). now, the way i see it when jumping out of a plane, the possible out comes are: no mal, single mal, double mal. So, 1:3. now those kinds of odds would keep me grounded. I guess my point is the theory of probabilities doesn't really apply to this situation or much of any other real world scenarios, unless of course you're in insurance.

Never go to a DZ strip show.

[rehmwa 2]

YOU ARE STARTING TO GET IT - this is my last for a bit. Time to go on vacation.

Quote

it's not, you are mixing up independent occurences with dependence based stats
where am I mixing it up?
---- dependence based probability (analogy) is like picking colored marbles out of a bag of 10 white and 10 black. If you don't replace the marble, then you change the probabilities on each draw - that is dependent on previous outcomes. If you draw 10 black in a row (unlikely) and don't put each marble back each time, then the 11th draw is 100% probable you will get a white. Do you see it? But if you put the marble back each time, then you are 50/50 every draw or 'independent of previous outcomes'. In the jumping example, your are effectively replacing the marble each time (jump)

If it's not applicable, please explain why, and please start with the coin toss example and show me how it's not any more likely to get one tails in 10 tosses than 1 tail in 1 toss

------Ok, this one is simple. We are talking about getting tails 'on the tenth toss'. We are NOT talking about getting one tail 'in ten tosses'. We can't do stats on that since we already made 9 tosses.

{does it help} it does a bit, but I still don't see the difference between the double mal stats problem and flipping a magic coin that has one side far more likely to land up than the other...

for any one flip the odds are the same...

but if you keep flipping, wanting to always get the more probable outcome, sooner or later you'll get the less probable outcome, it's just a matter of time...

------ flipping a coin does not have a 'more probable' outcome. it's always 50/50 and always independent of previous outcomes. If you have a 'majic' coin, that's completely different - see Ivan.

The key use of 'inferential' statistics is to 'predict' outcomes. Predict means you do all the calculations ahead of time based on probabilities. You can't apply the stats based on time = zero when you already live at time = later

Try considering the probabilities as the bag of marbles - 24,999 blue and 1 turd brown. But you have to remember to replace the marble back in the bag each time before you reach into the bag also to stir up the marbles really well each time. This is as good an analogy as I have.

This example is 'pick and replace', not 'pick and set aside'.

Oh, the 24000-way world record is no different, each individual has his own bag of blue and turd brown marbles. They are not all drawing from one bag.

It's fun to write with smart people, even boring stuff like stats.

merry Christmas

...
Driving is a one dimensional activity - a monkey can do it - being proud of your driving abilities is like being proud of being able to put on pants

[rehmwa 2]

Very funny and very cynical.

And very false.

Stats does apply to the real world in a much more real way than theoretical physics. Because there is variation in everything and we should account for it and most don't. Unforetuneatly those that don't are the builders and engineers, etc. Worst case analysis is the most expensive way to build things, and it's the most common way to design things because people don't get it.

Try reading the book, "How to Tell the Liars from the Statisticians". It's really boring but the first few examples are enough to get it.

Off the soap box. Merry Christmas

...
Driving is a one dimensional activity - a monkey can do it - being proud of your driving abilities is like being proud of being able to put on pants