HawkEye maths, please help!!!
jzfredricks
http://www.hawkeyeinnovations.co.uk/UserFiles/File/
I assume that is a legitimate site and document.
The HawkEye Accuracy document states;
Hawk-Eye is able to deliver a system which meets the following
performance criteria:
• Pitching point accuracy under 5mm (in MCC tests it was shown to be
2.6mm)
• Interception point accuracy under 5mm (in MCC tests it was shown to
be
2.6mm)
• Prediction of where the ball passes the stumps:
o In all “normal” LBW instances under 15mm and average error of 5mm
o In “extreme” LBW instances under 25mm
An “extreme” LBW is one where there is less than 40cm of travel
between pitching
point and interception point and the batsman is hit over 2 meters from
the stumps.
The current protocol has a 45mm umpire call “margin”.
Based on that, I worked this out;
HE Pitching Y-axis position == 0mm
Actual Pitching Y-axis position == +5mm (5mm away from the predicted
position)
HE Impact Y-axis position == 0mm
Actual Impact Y-axis position == -5mm (5mm away from the predicted
position)
(Combined Y-axis error of 10mm)
HE Distance between pitching and impact == 400mm
Actual Distance between pitching and impact == 390mm (error of 2 x
5mm)
So.. we've a 10mm opposite side, and a 390mm adjacent side. This gives
an angle of 1.47degs.
1.47degs over 2500mm == 67mm.
Yet there's only a 45mm allowance/error margin?
This is an example for when the ball travels 400mm between pitching
and impact. A "non-extreme" event.
In an "extreme" event this 67mm error will blow out quite a bit more.
I *imagine* (not done the sums) that HawkEye could easily be out by
more than 10cm (and that's before taking into account the predictive
software! ie it's simply a straight line).
a few notes;
a) please check for errors. It's been ages since I did this kind of
maths
b) for every not out batsmen given out, there should be an out batsmen
given not out (as the errors can happen both ways). I just think the
former is a greater sin.
Mike Holmans
<jzfre...@gmail.com> tapped the keyboard and brought forth:
>I've downloaded the HawkEye stuff from here;
>http://www.hawkeyeinnovations.co.uk/UserFiles/File/
>I assume that is a legitimate site and document.
>
>The HawkEye Accuracy document states;
>Hawk-Eye is able to deliver a system which meets the following
>performance criteria:
>� Pitching point accuracy under 5mm (in MCC tests it was shown to be
>2.6mm)
>� Interception point accuracy under 5mm (in MCC tests it was shown to
>be
>2.6mm)
>Based on that, I worked this out;
>HE Pitching Y-axis position == 0mm
>Actual Pitching Y-axis position == +5mm (5mm away from the predicted
>position)
>HE Impact Y-axis position == 0mm
>Actual Impact Y-axis position == -5mm (5mm away from the predicted
>position)
Why, when it explicitly tells you that testing showed an accuracy of
2.6mm, do you use 5mm for your calculations? You're so insistent on
testing, yet you disregard the results when they are published. Let's
face it, you're simply against HE and aren't going to let anything
like facts get in the way.
Cheers,
Mike
--
jzfredricks
> Why, when it explicitly tells you that testing showed an accuracy of
> 2.6mm, do you use 5mm for your calculations? You're so insistent on
> testing, yet you disregard the results when they are published. Let's
> face it, you're simply against HE and aren't going to let anything
> like facts get in the way.
This is the beauty of peer review.
I'll recalc at 2.6mm.
Mind if I do it for an "extreme" event?
ps: I don't trust the MCC tests :)
jzfredricks
> 2.6mm, do you use 5mm for your calculations?
I also figured the 5mm was from HE's own internal testing. Why didn't
they just say "accuracy under 3mm", rather than 5mm?
jzfredricks
Original example recalced at 2.6mm;
Margin of error = 2.6mm
Distance between pitching and impact = 400mm - 5.2mm = 394.8mm
Opposite side = 5.2mm (2 x 2.6mm)
Angle of error = 0.75degs
Error at stumps = 33mm
Extreme event (travel distance of 200mm)
Margin of error = 2.6mm
Distance between pitching and impact = 200mm - 5.2mm = 194.8mm
Opposite side = 5.2mm (2 x 2.6mm)
Angle of error = 1.53degs
Error at stumps = 67mm
Does HE list its "official" margin of error? Is it the 5mm or the
MCC's 2.6mm?
Nice al-foiled stumps, btw. Looks very scientific.
jzfredricks
Original re-calced at 3.8mm (average of 5mm and 2.6mm)
Margin of error = 3.8mm
Distance between pitching and impact = 400mm - 7.6mm = 392.4mm
Opposite side = 7.6mm (2 x 3.8mm)
Angle of error = 1.1degs
Error at stumps = 49mm
cricketrulez
Nice work. So there is a potential error of 5 cm on any prediction
according to H/E Innovations. I shudder to think what the error in
real life conditions is. No wonder they haven't released this thing
for independent testing.
jzfredricks
> Nice work. So there is a potential error of 5 cm on any prediction
> according to H/E Innovations. I shudder to think what the error in
> real life conditions is. No wonder they haven't released this thing
> for independent testing.
I'm not 100% certain what HE's actual, or even claimed, accuracy is.
The document says both "under 5mm" and "MCC tests say 2.6mm".
I'm also not certain of my own maths, as I said it's been ages since I
did this stuff, and I'm going from memory.
That example is for a 40cm travel distance before impact, with the
ball pitching at 2.5m.
If the ball travels less, then the margin of error is greater. As the
travel distance approaches 0, the error distance approaches infinity.
cricketrulez
What is the frame rate of these cameras they use? apparently it needs
3 measurements after pitching to have a prayer.
On a typical LBW appeal which happens withe the ball pitching anywhere
from 1 to 2 m from the batsman the time for the ball to hit the pad at
140km/h is about .025 sec to 0.5 sec. If the frame rate is 60 fps,
you'll get 3 frames at 2 m pitching length anything less.....
The more and more we dig into this, H/E is beginning to sound like
snake oil!
jzfredricks
> What is the frame rate of these cameras they use?
I think I read 106fps, and Andrew mentioned ~220fps
alvey
Did you know that Ashocker's been dropped twice from the EU panel?
alvey
jzfredricks
> Did you know that Ashocker's been dropped twice from the EU panel?
How many time has Taufel been dropped? Or Dar?
Andrew Dunford
"jzfredricks" <jzfre...@gmail.com> wrote in message
news:61da15c0-e531-408a...@r27g2000prr.googlegroups.com...
> On Jun 17, 5:29 pm, cricketrulez <cricketrulez1...@gmail.com> wrote:
>> What is the frame rate of these cameras they use?
>
> I think I read 106fps, and Andrew mentioned ~220fps
The most recent figure I could find for Hawk Eye is 106fps. Virtual Eye
operates at a variable rate between 60-230fps according to brightness of
natural light and speed of bowler.
Andrew
sdavmor
3rd time's the charm?
--
Cheers, SDM -- a 21st Century Schizoid Man
Systems Theory project website: http://systemstheory.net
find us on MySpace, GarageBand, Reverb Nation, Last FM, CDBaby
free MP3s of Systems Theory, Mike Dickson & Greg Amov music at
http://mikedickson.org.uk
Eugene
> travel distance approaches 0, the error distance approaches infinity.
Same for a human umpire, the closer the ball pitches to the point of
impact on the pad the harder it is to determine how much deviation
there is off the pitch.
jzfredricks
So true. I know it, you know it, hopefully umpires know it.
2 wrongs don't make a right, and if technology is going to overrule
human umps it should be held to a higher standard.