Hawk-Eye: The Faith based initiative
cricketrulez
appropriate description.
1) Neither the cheerleaders on RSC nor HE Innovations can produce any
data on its accuracy and consistency under real life conditions.
2) only available information indicates that whatever testing was done
using a bowling machine. I'm sure it simulates the swing of Zak, Jimmy
Anderson and the spin and turn of Ajmal, swann etc. NOT!
3) Amateur calculations (No offense meant JZF!) based on available
tolerances doesn't paint a pretty picture.
On top of this, this stuff isn't exactly cheap.
Given all this why is this such a fanatical support for this?
In case anyone missed it, here is a great take on H-E and its
cheerleaders stand by JZF.
<<<<My cricket club is doing some testing at the local school and we
plan
to release the results soon. You will have no right to question our
results, sorry. You will also have no right to point out failings in
the regulations based on those results. >>>
jzfredricks
> 3) Amateur calculations (No offense meant JZF!) based on available
> tolerances doesn't paint a pretty picture.
None taken. I put it all up for peer review for this very reason. Mike
pointed out 2 flaws (I used 5mm instead of 2.6mm and "half a ball
width" issues), so I re-calced. It still doesn't look great for xEye.
I'm as pro-technology as anyone in this ng (I believe more so), but I
want proven, effective technology. There's no point using technology
that raises more questions than it answers. We want to reduce the
doubt, not add to it.
cricketrulez
Hope you won't haul me up for copyright violations on borrowing from
your post about the testing at your local club bit!
Mike Holmans
<jzfre...@gmail.com> tapped the keyboard and brought forth:
You haven't yet bothered to answer whether in your opinion xEye is
more or less accurate than the unaided human eye of the umpire. Saying
that you don't know because you haven't seen public testing is a cop
out.
How often have you seen a prediction from xEye that you consider
unreasonable, and how often have you seen predictions and thought
yeah, that looks about right?
I also find the theory that umpires who do the job day in day out and
say that they think xEye is good are simply brainwashed by ICC to be
in the tin hat protecting from rays area.
Your doubts, unshared by anyone who is actually relevant apart from
perhaps MS Dhoni, seem to be founded on the premise that if it can be
inaccurate by up to a certain amount, it is always inaccurate by that
maximum amount.
How about you do your maths based on a normal distribution of reading
errors and find out what tiny percentage of possible errors would get
a decision wrongly upheld/overturned, and compare that to the number
of decisions correctly changed? Remember to include your estimate of
the accuracy of the unaided human eye.
In other words, why don't you analyse the likely distribution of
circumstances in actual play rather than concentrating on limit cases
where the accuracy of the unaided human eye is likely to be least in
the first place?
Cheers,
Mike
--
Andrew Dunford
"cricketrulez" <cricketr...@gmail.com> wrote in message
news:100284ea-c43e-413b...@j15g2000yqf.googlegroups.com...
It's also worth asking why some broadcasters have switched to a rival
product.
Andrew
jzfredricks
> You haven't yet bothered to answer whether in your opinion xEye is
> more or less accurate than the unaided human eye of the umpire. Saying
> that you don't know because you haven't seen public testing is a cop
> out.
* see below
Overall I'd probably say it's more accurate.
Please don't confuse that answer with "accurate enough".
> How often have you seen a prediction from xEye that you consider
> unreasonable, and how often have you seen predictions and thought
> yeah, that looks about right?
Enough of the former to question its use. Plenty of the latter.
> I also find the theory that umpires who do the job day in day out and
> say that they think xEye is good are simply brainwashed by ICC to be
> in the tin hat protecting from rays area.
Show them the maths and explain the regulations, and they will
probably object.
> Your doubts, unshared by anyone who is actually relevant apart from
> perhaps MS Dhoni, seem to be founded on the premise that if it can be
> inaccurate by up to a certain amount, it is always inaccurate by that
> maximum amount.
You keep peddling this line, and I keep telling you that my objections
are "worst case" examples that don't happen every day.
What part of that don't you understand?
> How about you do your maths based on a normal distribution of reading
> errors and find out what tiny percentage of possible errors would get
> a decision wrongly upheld/overturned, and compare that to the number
> of decisions correctly changed? Remember to include your estimate of
> the accuracy of the unaided human eye.
There simply isn't enough buffer. xEye will incorrectly give a batsman
out when the on-field ump has (correctly) said Not Out.
We trust it down to level that is UNDER their own margin of error
(both the 5mm and the MCC's 2.6mm).
Which brainiac decided "we'll trust it to 0mm precision even though
our own testing proves we can't"...
* speaking of un-answered questions... I hope you get round to
answering mine (re how much of a ball should be hitting for the ump to
give it out)
cricketrulez
> On Mon, 20 Jun 2011 15:25:55 -0700 (PDT), jzfredricks
Mike,
Having followed your posts for a long time, this is a highly obtuse
line of argument from you.
1) till recently (by history of cricket standards) there hasn't been a
ball tracking technology leave alone a predicting technology
2) How exactly does one get at the accuracy of the umpires decisions
on LBW? By using a predictive technology? How exactly do you figure
out the human accuracy on something that wasn't measurable?
I'm sure you already know this and your pleadings sound more emotional
than anything else!
cricketrulez
This line of argument is well below your standards at RSC.
1) The ball tracking technology has been around for ~5 years in
cricket (H/E patent app was in 2001) and predictive technology in use
for even less.
2) how exactly does one measure the accuracy of human umpires
decisions so far? using predictive technology?
In order to do that, you'd need to validate the predictive technology.
wouldn't you agree?
Your postings on this topic sound emotional and not entirely
rational!
xs...@yahoo.com
i always snicker when the tv commentator says 'oh that's a very good
decision' based on some hawkeye projection.
i still think the human eye on a competent umpire is the best judge
for lbw decisions. i am willing to live with the uncertainty and
benefit
of doubts that are traditional in cricket. with edges, you can say
more
conclusively - there is no guess work involved.
besides, whinging about umpiring decisions is a pleasure no cricket
fan should be deprived of.
Bob Dubery
> I realize the title sounds harsh, but "at this point" it is a very
> appropriate description.
>
> 1) Neither the cheerleaders on RSC nor HE Innovations can produce any
> data on its accuracy and consistency under real life conditions.
exactly how the ball travelled and then compare hawkeye to that. The
trouble is having a reference trace to compare the hawkeye trace to.
If you could come up with a way of doing that then you'd be able to
validate the non-predictive component of hawkeye.
Emphatic validation of the predictive component is NOT possible. That
doesn't mean though that it's worse or as bad as the best human
umpires - and remember that there's no way at all of measuring and
calibrating their judgements of how a ball would have travelled.
> 2) only available information indicates that whatever testing was done
> using a bowling machine. I'm sure it simulates the swing of Zak, Jimmy
> Anderson and the spin and turn of Ajmal, swann etc. NOT!
>
> 3) Amateur calculations (No offense meant JZF!) based on available
> tolerances doesn't paint a pretty picture.
>
> On top of this, this stuff isn't exactly cheap.
The price will come down as the technology and it's components become
more widely used.
jzfredricks
> How exactly can you produce such data? To do that you'd have to KNOW
> exactly how the ball travelled and then compare hawkeye to that. The
> trouble is having a reference trace to compare the hawkeye trace to.
> If you could come up with a way of doing that then you'd be able to
> validate the non-predictive component of hawkeye.
We all ready know xEye operators can manually select the impact
location.
They just have to bowl a ball, have the batsman miss it, select a
pretend impact location mid-path and compare the results.
Bob Dubery
> On Jun 21, 9:14 am, Mike Holmans <m...@jackalope.demon.co.uk> wrote:
>
> > You haven't yet bothered to answer whether in your opinion xEye is
> > more or less accurate than the unaided human eye of the umpire. Saying
> > that you don't know because you haven't seen public testing is a cop
> > out.
>
> * see below
> Overall I'd probably say it's more accurate.
This is the point of UDRS - not that it's perfect but that it's
BETTER.
> Please don't confuse that answer with "accurate enough".
"more accurate" should be a good enough answer for any reasonalbe
person.
<snip>
jzfredricks
> Then the problem is???
The problems are legion, but here are 4;
a) they trust it to a precision that they know it can't achieve
b) it can incorrectly give a batsman out by overruling an umpire who,
correctly, gave it not out
c) it hasn't been publicly and thoroughly tested
d) maths proves the prediction can be out by 10cm (or more)
Overall I think it's better as it won't make as many howlers as humans
do.
Andrew Dunford
<xs...@yahoo.com> wrote in message
news:fbe5cc57-9ff9-4d62...@y7g2000prk.googlegroups.com...
<snip>
> besides, whinging about umpiring decisions is a pleasure no cricket
> fan should be deprived of.
This is very true. That said, it would only be replaced by moaning about
the shortcomings of the UDRS.
Andrew
Andrew Dunford
"Bob Dubery" <mega...@gmail.com> wrote in message
news:47766120-96c5-45eb...@t38g2000prj.googlegroups.com...
> On Jun 21, 1:39 am, jzfredricks <jzfredri...@gmail.com> wrote:
>> On Jun 21, 9:14 am, Mike Holmans <m...@jackalope.demon.co.uk> wrote:
>>
>> > You haven't yet bothered to answer whether in your opinion xEye is
>> > more or less accurate than the unaided human eye of the umpire. Saying
>> > that you don't know because you haven't seen public testing is a cop
>> > out.
>>
>> * see below
>> Overall I'd probably say it's more accurate.
> Then the problem is??? You're saying it will lead to better decisions.
> This is the point of UDRS - not that it's perfect but that it's
> BETTER.
It's not necessarily about whether the available technology is 'better' than
a human umpire. The UDRS can improve the overall standard of
decision-making by correcting some errors, bearing in mind that elite
umpires get the majority of decisions correct already. And it can achieve
this without ball-tracking or Hot Spot.
Andrew
alvey
Unlike your OP...
Anyway... It's always pleasant to have one's pov confirmed. Either by the
admission of the intelligent observer or, like this issue, finding that a
blithering idiot holds the opposite pov.
'Cricketrulez'? 'Cricketdrulez' more like it.
alvey
Mike Holmans
Well, that's a reasonable end position, which is a lot better than "it
shouldn't be used at all". Of course there are some problems in
difficult cases, but that is not a reason to junk the whole thing any
more than it would be sensible to junk Duckbill-Platypus just because
you can construct some fairly artifical scenarios involving multiply-
interrupted odos in which the revised target is much too easy or much
too difficult.
Cheers,
Mike
R Bharat Rao
> On Jun 20, 11:56 pm, cricketrulez <cricketrulez1...@gmail.com> wrote:> I realize the title sounds harsh, but "at this point" it is a very
> > appropriate description.
>
> > 1) Neither the cheerleaders on RSC nor HE Innovations can produce any
> > data on its accuracy and consistency under real life conditions.
>
> How exactly can you produce such data? To do that you'd have to KNOW
> exactly how the ball travelled and then compare hawkeye to that. The
> trouble is having a reference trace to compare the hawkeye trace to.
> If you could come up with a way of doing that then you'd be able to
> validate the non-predictive component of hawkeye.
For instance, have the ball leave a spot on the pitch, and verify how
many times you actually hit it? I'm very very leery of claims that
Hawkeye gets the landing spot right based on "of course it does."
The human eye is *AMAZING* at interpolating and estimation. The
following is highly OT, so skip if you like.
A few years ago, my team worked on the problem of finding cardiac wall
motion abnormalities from cardiac ultrasound (echocardiograph or echo
for short).
Anyway, a key sub problem was to start by estimating the cardiac
wall. If you have seen a cardiac ultrasound (http://en.wikipedia.org/
wiki/Echocardiography) it looks quite fuzzy -- but it is a video. So,
I did this experiment.
I showed my kids (then 10 & 6) a few echo video & then a couple of
stills -- I then with a pencil showed them how one could draw the echo
wall on a couple of echos. Then gave them 10 each, and asked them to
draw where they thought the wall was. The results were surprising--
the 10 year old got all of them right, and the 6 year old got most of
them right -- it seemed that a couple of words of instruction about
drawing a smooth contour and ignoring drop outs & random flecks was
sufficient. And I later verified this with a few other kids -- it
seemed that 5-years was the limit (but it may have had more to do with
holding a pencil firmly and being able to execute a line with fine
motor control, than lack of visual acuity)
Why astounding? What was arguably the finest medical image processing
team in the world couldn't do this as well, even after over 10 staff
years of effort.
The valuable lesson was that just because it was easy for the human
eye to do, didn't mean that an image processing or video processing
algorithm could do it.
(How did we solve it? Used a machine learning based approach, having
expert echocardiographers contour over a 1000 different echo images,
and having our system learn how to draw the border based on image
processing primitives, coupled with tracking over the video to correct
errors. This was major research, and while the Hawkeye folks may have
done it, I haven't seen anything like it. See check out
http://www.aaai.org/Papers/IJCAI/2007/IJCAI07-082.pdf if anyone want
gory details.)
So, I'd start first with verification of the landing spot and accurate
determination of whether it intersected the pitch map...
> Emphatic validation of the predictive component is NOT possible.
It could be quite easily done as well... but lets set that aside for
now.
Bharat
Freaky
> > appropriate description.
>
> > 1) Neither the cheerleaders on RSC nor HE Innovations can produce any
> > data on its accuracy and consistency under real life conditions.
>
> How exactly can you produce such data? To do that you'd have to KNOW
> exactly how the ball travelled and then compare hawkeye to that. The
> trouble is having a reference trace to compare the hawkeye trace to.
> If you could come up with a way of doing that then you'd be able to
> validate the non-predictive component of hawkeye.
>
> Emphatic validation of the predictive component is NOT possible. That
> doesn't mean though that it's worse or as bad as the best human
> umpires - and remember that there's no way at all of measuring and
> calibrating their judgements of how a ball would have travelled.
>
I can certainly think of experiments that could test it's accuracy as
compared to a human. A simple experiment is have a human observer act
as an "umpire." Have bowlers of all sorts bowl deliveries at the
stumps and the umpire has a device that lets him click a binary choice
of hitting or not hitting. Who is more accurate - the technology or
the human? There can be more complex variants of this - you could add
a batsman who moves out of the way, or perhaps a holographic batsman.
We may not be privy to all the results, but I would think and hope
that the makers of the technology have actually done some real
validation experiments.
V
jzfredricks
> On Jun 21, 11:56 am, jzfredricks <jzfredri...@gmail.com> wrote:
>
> > On Jun 21, 6:29 pm, Bob Dubery <megap...@gmail.com> wrote:
>
> > > Then the problem is???
>
> > The problems are legion, but here are 4;
> > a) they trust it to a precision that they know it can't achieve
> > b) it can incorrectly give a batsman out by overruling an umpire who,
> > correctly, gave it not out
> > c) it hasn't been publicly and thoroughly tested
> > d) maths proves the prediction can be out by 10cm (or more)
>
> > Overall I think it's better as it won't make as many howlers as humans
> > do.
>
> Well, that's a reasonable end position, which is a lot better than "it
> shouldn't be used at all".
You snipped a bit...
It shouldn't be used at all until its accuracy has been confirmed by
thorough, public testing.. and after that the regulations are written
to suit that accuracy.
Do those things, and I'll champion xEye as blindly as you.
jzfredricks
> For instance, have the ball leave a spot on the pitch, and verify how
> many times you actually hit it? I'm very very leery of claims that
> Hawkeye gets the landing spot right based on "of course it does."
The truly saddening (and maddening) thing is the same people who
defend xEye are the ones who keep on saying "it can't be tested
properly".
If it can't be tested.. why are we using it?!?!
jzfredricks
> I can certainly think of experiments that could test it's accuracy as
> compared to a human. A simple experiment is have a human observer act
> as an "umpire." Have bowlers of all sorts bowl deliveries at the
> stumps and the umpire has a device that lets him click a binary choice
> of hitting or not hitting. Who is more accurate - the technology or
> the human? There can be more complex variants of this - you could add
> a batsman who moves out of the way, or perhaps a holographic batsman.
> We may not be privy to all the results, but I would think and hope
> that the makers of the technology have actually done some real
> validation experiments.
We can transplant hearts and put men on the moon. Testing xEye should
be easy.
xs...@yahoo.com
>
> The valuable lesson was that just because it was easy for the human
> eye to do, didn't mean that an image processing or video processing
> algorithm could do it.
the human eye in fact adapts to pitch conditions, which
is why top class batsmen can play hundreds of deliveries
in an innings. they are interpolating and anticipating the
ball's trajectory all the time. what we are not good at is
the sort of accuracy hot-spot brings - things to do with
precision - edges, run-outs etc. run-outs are hard because
the eye has to look at two different things - the bails going
off and the batsman's bat crossing the line.
Mike Holmans
>
> - Show quoted text -
Blindly is a gross misrepresentation. I look at what works practically
and has worked well to the satisfaction of all those who have had to
deal with it in real life (apart from MS Dhoni) and think it should
carry on. You want to throw it away because you have some concerns
about theoretical cases in which it is just about possible that under
some circumstances a human might by chance get it right and be
overruled, even though the range of events where this will happen is
too small to be of any major concern to anyone who isn't a total
obsessive.
Anyway, I shan't be contributing any more to this discussion.
Cheers,
Mike
jzfredricks
> Anyway, I shan't be contributing any more to this discussion.
thanks for the engagement. I both enjoyed it and learnt from it.
Vijay
> On Jun 21, 4:25 am, Bob Dubery<megap...@gmail.com> wrote:
>> On Jun 20, 11:56 pm, cricketrulez<cricketrulez1...@gmail.com> wrote:> I realize the title sounds harsh, but "at this point" it is a very
>>> appropriate description.
>>
>>> 1) Neither the cheerleaders on RSC nor HE Innovations can produce any
>>> data on its accuracy and consistency under real life conditions.
>>
>> How exactly can you produce such data? To do that you'd have to KNOW
>> exactly how the ball travelled and then compare hawkeye to that. The
>> trouble is having a reference trace to compare the hawkeye trace to.
>> If you could come up with a way of doing that then you'd be able to
>> validate the non-predictive component of hawkeye.
>
> For instance, have the ball leave a spot on the pitch, and verify how
> many times you actually hit it? I'm very very leery of claims that
> Hawkeye gets the landing spot right based on "of course it does."
>
> The human eye is *AMAZING* at interpolating and estimation. The
> following is highly OT, so skip if you like.
>
> A few years ago, my team worked on the problem of finding cardiac wall
> motion abnormalities from cardiac ultrasound (echocardiograph or echo
> for short).
>
> Anyway, a key sub problem was to start by estimating the cardiac
> wall. If you have seen a cardiac ultrasound (http://en.wikipedia.org/
> wiki/Echocardiography) it looks quite fuzzy -- but it is a video. So,
> I did this experiment.
>
> stills -- I then with a pencil showed them how one could draw the echo
> wall on a couple of echos. Then gave them 10 each, and asked them to
> draw where they thought the wall was. The results were surprising--
> the 10 year old got all of them right, and the 6 year old got most of
> them right -- it seemed that a couple of words of instruction about
> sufficient. And I later verified this with a few other kids -- it
> seemed that 5-years was the limit (but it may have had more to do with
> holding a pencil firmly and being able to execute a line with fine
> motor control, than lack of visual acuity)
>
> Why astounding? What was arguably the finest medical image processing
> team in the world couldn't do this as well, even after over 10 staff
> years of effort.
>
> The valuable lesson was that just because it was easy for the human
> eye to do, didn't mean that an image processing or video processing
> algorithm could do it.
>
> (How did we solve it? Used a machine learning based approach, having
> expert echocardiographers contour over a 1000 different echo images,
> and having our system learn how to draw the border based on image
> processing primitives, coupled with tracking over the video to correct
> errors. This was major research, and while the Hawkeye folks may have
> done it, I haven't seen anything like it. See check out
> http://www.aaai.org/Papers/IJCAI/2007/IJCAI07-082.pdf if anyone want
> gory details.)
>
> So, I'd start first with verification of the landing spot and accurate
> determination of whether it intersected the pitch map...
>
>> Emphatic validation of the predictive component is NOT possible.
>
> It could be quite easily done as well... but lets set that aside for
> now.
>
> Bharat
Bharat, thanks for sharing that!
You are right the human mind (all senses) is capable of far more
accuracy than we give it credit for. Not specific to HE or HS, assuming
that anything is better than human judgment just because the anything
happens to be technology is not gonna help.
When it comes to LBWs I am not sure about HE's accuracy but as is the
topic header, my "ok" for it based on faith...nothing more. To be really
honest I am still not able to figure how the hell did HE show Ajmal's
ball missing Sachin's leg in the SF.
On ICF (some interested in this discussion would already be aware) there
was a thread about HE using something called Kalman Filtering technique
which is used in guiding missiles or something like that.
I did raise my doubts there, not doubting HE but to understand how well
it learns the ball's course during the event...say after pitching the
amount of bounce, the turn, influence of ground rigidity, wind, ball
age, etc. I still don't know if HE takes care of all this while it
generates it's prediction.
does anyone know why ICC changed from not using prediction of HE earlier
to using it now? Was any rationale provided by ICC for the change in
heart. More importantly, did HE enlighten them with any data or any new
info that made them change heart
Freaky
I am a radiologist. There is no computer aided detection system that
has (to my knowledge) worked better than the human eyes as yet in any
exam, and not for the lack of trying. The CAD systems either turn up
too many false positives or end up missing critical disease. Our eyes
are an absolutely amazing tool. I think jzf has been trying to say
(and please correct me jz if I am wrong), that evidence showing that
HE is actually better than the human eye has not been presented yet,
and absent this evidence, it seems premature to adopt the system on
the word of a company. Just the numbers released indicate real
possible problems. This is what I have surmised from jz's arguments
and analysis here, and this does not sound unreasonable. He actually
has swayed me at least - I was for HE before his posts got me thinking
about it more, and I find myself moving towards this argument.
V
Vijay
I think the main point of cr, xss, jz, etc seems to be that just because
people "feel" it works well doesn't mean it works well. They would like
to see some solid proof before throwing the ump's eye out
jzfredricks
> I think the main point of cr, xss, jz, etc seems to be that just because
> people "feel" it works well doesn't mean it works well.
Freaky, this is my main point, and has been for a while. I want
thorough, public testing of xEye. Mainly to prove its accuracy.
During the course of this discussion I also learnt that the
regulations have shortcomings that I think should be fixed. However,
these should be based on the test results, so until they are available
I can't suggest fleshed out regulation changes. Most of that is based
on the feeling that the 3rd ump should only overturn the onfield
decision if there is CLEAR evidence.
I also developed a slightly newer position - I don't like technology
MAKING the decision. I'd prefer it to present evidence, and for a
human to decide what that evidence means. With xEye, we simply get
told "facts". There is no human judgement involved.
cricketrulez
Go thro' this thread and point out factual posts which have disproved
my original POV.
As for personal insults, it hardly matters what a convict condom waste
like you have to say.
cricketrulez
Bob,
Yes the measurements can be made as follows.
1) Capture video footage with hight frame rate cameras, the bowling
from bowlers of various types under various conditions w/o batsman
present. This way we already know what the ball did.
2) Edit the footage and feed incomplete deliveries to the to the H/E
Software and umpires and verify their prediction against the original
footage. This would a be comparison of umpires and H/E under test
conditions.
This should be done at a minimum and results made public before any
implementation of H/E in cricket.
alvey
No worries. Be glad to. With extra relish even.
However, that's not really fair is it? Shouldn't you respond to my request
for a response in that other thread that you're making a fool of yourself
in first? You know. The magic one one with different Subjects. I'm quite
keen to read your explanation.
>
> As for personal insults, it hardly matters what a convict condom waste
> like you have to say.
The bar was deathly quiet after the small man with three teeth and a broken
thong entered and spat out his ultimate insult. 'Convicts!' The bar was
quiet because it'd been closed for over an hour. The small man had made
another error. He shrugged his uneven and bony shoulders in a well-worn and
resigned way before shuffling out through the door. Pausing only to trip
over the thong that he broke upon entering.
alvey
alvey
> On Mon, 20 Jun 2011 15:25:55 -0700 (PDT), jzfredricks
> <jzfre...@gmail.com> tapped the keyboard and brought forth:
>
>>On Jun 21, 7:56�am, cricketrulez <cricketrulez1...@gmail.com> wrote:
>>> 3) Amateur calculations (No offense meant JZF!) based on available
>>> tolerances doesn't paint a pretty picture.
>>
>>None taken. I put it all up for peer review for this very reason. Mike
>>pointed out 2 flaws (I used 5mm instead of 2.6mm and "half a ball
>>width" issues), so I re-calced. It still doesn't look great for xEye.
>>
>>I'm as pro-technology as anyone in this ng (I believe more so), but I
>>want proven, effective technology. There's no point using technology
>>that raises more questions than it answers. We want to reduce the
>>doubt, not add to it.
>
> You haven't yet bothered to answer whether in your opinion xEye is
> more or less accurate than the unaided human eye of the umpire.
Yairsss. JZF has a bit of a habit of leaving my nasty questions alone.
> Saying
> that you don't know because you haven't seen public testing is a cop
> out.
Not to mention one of the most childish and block-headed whines ever heard
outside of parliaments. I can't think of a single company which relies
primarily on intellectual property that allows, what is it again,
"thorough, public testing" of its product. Fucked if I know how he ever
buys anything.
Anyway, this is *such* a non-argument. It's like Dawkins v that crazy-eyed
creationist.
alvey
jzfredricks
> I can't think of a single company which relies
> primarily on intellectual property that allows, what is it again,
> "thorough, public testing" of its product.
Chances are you're not thinking then.
The beauty about NOT having public, thorough testing.. for YOU.. is
you're able to continue your anti-BCCI rants with no proper
foundation. If testing was done, and xEye passed, you'd be able to
continue without looking racist. If xEye failed the tests, and BCCI
*still* said they weren't satisfied with its accuracy, then you'd be
able to continue your rants whilst looking racist.
cricketrulez
>I can't think of a single company which relies
> primarily on intellectual property that allows, what is it again,
> "thorough, public testing" of its product.
>
What a load of crap! No independent validation of promised performance
becos of IP issues?
since you have been asking questions about other people quals, what
are your quals which allows to make that statement?
the call is for validation of the technology for its intended use. Not
hijacking of H/E IP.
cricketrulez
> primarily on intellectual property that allows, what is it again,
> "thorough, public testing" of its product.
> alvey
Strong statement, so what are your qualifcations in matters like this?
It sounds like a load of crap. Regardless of IP, any vendor has to
provide performance benchmarks. Since for H/E it will fall under the
category of service rather than a product that ICC can buy and own,
The quality of the service would need to demonstrated. Has it been?
Another point, the patent application for HE apparently has been
abandoned and therefore is in public domain.
So what IP are you talking about?
jzfredricks
> Strong statement, so what are your qualifcations in matters like this?
He has no qualifications, and his "they have a right to protect their
IP" stance is, frankly, half-witted.
Asking them to demonstrate that their product performs as well as
their marketing brochure says it does is not a theft of their IP in
any way, shape or form.
If HE themselves tried the "we can't do tests, sorry, it's IP" line,
then the ICC should simply tell them to fuck off.
cricketrulez
I wonder if he will specify what IP he is talking about?
1) Patent application is dropped and therefore the idea is public
domain and any specific implementation would merely be an obvious
extension and therefore not patentable.
2) copyrights: might be applicable to the code base and the software:
how will this be compromised by validation work?
3) trademarks: once again not affected by validation work.
OTY alvey. Lets see if you can defend your words.
Bob Dubery
1) We're not talking Hawk-Eye only here. Another system is in use and
that would have to be verified too.
2) HE uses input from multiple fixed cameras, so you'd need to have
multiple sets of input data and be careful with the editing
3) Do you know what happened? Do you know exactly how much the ball
missed the stumps by? You know that it missed, but by how much?
4) You'd need to be sure of the calibration and alignment of the
cameras used to capture the test footage. This is not like calibrating
a speed trap where it's a simple question of "did the object cover
that distance in that time?"
5) All of this achieves not very much because you still haven't
answered the real question: Is HE more accurate and/or reliable than a
human umpire. And which umpire - annoyingly they don't all perform
the same and don't perform the same all the time.
The more I argue about this, the more I come round to thinking that
it's good as it is. The technology is only supposed to be there to
correct howlers. This is why technology was explored in the first
place - not to address hair's breadth decisions but for occasions when
an umpire screws up or is unsighted. If it's a question of fine
tolerances or the point of impact being too far forwards then that is
not perfect, but it's also not a howler.
And this is why UDRS is a good way to use the technology. If it were
free to use at any time players would constantly be referring just
because they could. UDRS, in effect, says "if you think a fundamental
error has been made then you can challenge the decision. If you are
right then you lose nothing. If you are wrong then you have lost a
future opportunity to challenge a decision." If the umpire is having a
bad day at the office then your challenge has a good chance of being
upheld and you will get the glaringly bad decisions changed. If you
decide to use your challenges for something other than correcting what
looks like a poor decision then you run the risk of exhausting your
"credit" of challenges. Implied in all of this is "so use it for what
it's meant to be used for."
jzfredricks
> 1) We're not talking Hawk-Eye only here. Another system is in use and
> that would have to be verified too.
True. I think cricket's right to know the technology works trumps any
IP issues (not that testing will steal the IP in the first place)
> 2) HE uses input from multiple fixed cameras, so you'd need to have
> multiple sets of input data and be careful with the editing
and?
> 3) Do you know what happened? Do you know exactly how much the ball
> missed the stumps by? You know that it missed, but by how much?
You seem to be hinting that xEye CAN'T be tested.
Do you support using technology that can't be tested?
> 4) You'd need to be sure of the calibration and alignment of the
> cameras used to capture the test footage. This is not like calibrating
> a speed trap where it's a simple question of "did the object cover
> that distance in that time?"
You seem to be hinting that xEye CAN'T be tested.
Do you support using technology that can't be tested?
> 5) All of this achieves not very much because you still haven't
> answered the real question: Is HE more accurate and/or reliable than a
> human umpire. And which umpire - annoyingly they don't all perform
> the same and don't perform the same all the time.
How can that be answered without seeing testings?
> The more I argue about this, the more I come round to thinking that
> it's good as it is. The technology is only supposed to be there to
> correct howlers.
Can you please define a "howler" for us?
> This is why technology was explored in the first
> place - not to address hair's breadth decisions but for occasions when
> an umpire screws up or is unsighted.
Then why do we trust xEye to make hair's breadth decisions?
We currently trust it to PERFECT accuracy for 2 readings - pitch
location and impact locations for when the ump has given it out.
cricketrulez
Bob,
Many valid points. All the more reason to validate any predictive
technology whether it is HE or VE. If you can't test it, no point
using it.
The experiment I suggested, while far from complete is a very good
starting point IMHO. If HE or VE is not willing carry that out and
publish the results, they should be told to take a hike.
Those tests however might only be starting point for validation.
alvey
> On Jun 22, 6:12�pm, alvey <al...@atattat.com> wrote:
>> I can't think of a single company which relies
>> primarily on intellectual property that allows, what is it again,
>> "thorough, public testing" of its product.
>
> Chances are you're not thinking then.
So could you name one or two then?
>
> The beauty about NOT having public, thorough testing.. for YOU.. is
> you're able to continue your anti-BCCI rants with no proper
> foundation. If testing was done, and xEye passed, you'd be able to
> continue without looking racist. If xEye failed the tests, and BCCI
> *still* said they weren't satisfied with its accuracy, then you'd be
> able to continue your rants whilst looking racist.
What??? Even by your perverted logical process this is bizzarre.
(Hint: Banging the "you're anti-BCCI" drum is the same to me as stamping
'I'm an Idiot!' on your forehead.)
I'm having a go at YOU for YOUR recurring, logic-defying, whine about x-Eye
having to go through "public, thorough testing".
How do you "test it"?
Why does it need to be "public" testing? (Whatever the fuck that is)
Why don't you apply this rigor to anything else?
And lastly, correct me if I'm wrong, didn't you recently *finally* (but
indirectly) answer my question put to you on what you regarded as more
accurate; humans or xEye, and YOU admitted that xEye was. Yes you did;
"Overall I'd probably say it's [AS: xEye] more accurate." jzf. That cuts
the soapbox out from under YOU I'd say.
So YOU'RE advocating that cricket should use a method that 99.9% recurring
of the world, YOU included, believes is inferior. Weird hey?
And another reason I recently added for xEye's mandatory use was the fact
that it would significantly reduce the odds of an umpire being bullied or
worse, bought. Or would you care to "argue" this as well?
alvey
alvey
> On Jun 23, 10:53�am, cricketrulez <cricketrulez1...@gmail.com> wrote:
>> Strong statement, so what are your qualifcations in matters like this?
>
> He has no qualifications,
Sorry? Exactly what qualifications does one need?
Oh, and what did you say that yours were in respect to addressing Hawkeye?
> and his "they have a right to protect their
> IP" stance is, frankly, half-witted.
Please explain the existence of legions of attack lawyers employed by
corporations who sniff the world for breaches.
>
> Asking them to demonstrate that their product performs as well as
> their marketing brochure says it does is not a theft of their IP in
> any way, shape or form.
Absolutely. But what has this to do with "public" testing?
>
> If HE themselves tried the "we can't do tests, sorry, it's IP" line,
> then the ICC should simply tell them to fuck off.
An extrapolation of the last irrelevancy. Ignore.
alvey
jzfredricks
> How do you "test it"?
I'll ask you the same question I asked Bob (btw, Bob, you seem to have
run off)
Can xEye be tested?
If not, do you support using technology that can't be tested?
I'll respond to the rest some time today - I've gota run to the damned
thing called work :(
jzfredricks
> On Wed, 22 Jun 2011 18:34:19 -0700 (PDT), jzfredricks wrote:
> > On Jun 23, 10:53 am, cricketrulez <cricketrulez1...@gmail.com> wrote:
> >> Strong statement, so what are your qualifcations in matters like this?
>
> > He has no qualifications,
>
> Sorry? Exactly what qualifications does one need?
A reasonable amount of intelligence.
> Oh, and what did you say that yours were in respect to addressing Hawkeye?
A reasonable amount of intelligence.
> > and his "they have a right to protect their
> > IP" stance is, frankly, half-witted.
>
> Please explain the existence of legions of attack lawyers employed by
> corporations who sniff the world for breaches.
Come on mate, you don't strike me as being completely moronic. You can
do better than this.
If the testing does reveal some IP (and I don't think it will) that
doesn't give xEye competitors a right to use that IP.
The hoards of lawyers will, sadly, still be able to keep their jobs.
> > Asking them to demonstrate that their product performs as well as
> > their marketing brochure says it does is not a theft of their IP in
> > any way, shape or form.
>
> Absolutely. But what has this to do with "public" testing?
Thorough, public testing is the only way to remove all doubt about the
efficacy of the system.
Having said that, I could settle for thorough, indepenent testing. So
long as the 3rd party was reputable *and* the testing methods (and
full results) were made public.
I don't consider the MCC independent enough, and I don't think their
testing was good enough (I base this on the photos I've seen in the
accuracy reports which I *assume* were taken during MCC testing).
jzfredricks
> > On Jun 22, 6:12 pm, alvey <al...@atattat.com> wrote:
> >> I can't think of a single company which relies
> >> primarily on intellectual property that allows, what is it again,
> >> "thorough, public testing" of its product.
>
> > Chances are you're not thinking then.
>
> So could you name one or two then?
[http://www.nocsae.org/MediaKit/2011/NOCSAE%20News%20Release%20on
%20Data%20Request%2001%2024%2011%20FINAL.pdf]
That's for NFL equipment - helmets.
The list of companies that make the helmets would be quite large. More
than the "one or two" you've requested. Do you want me to provide a
list of companies that make NFL helmets?
alvey
> On Jun 24, 7:59�am, alvey <al...@atattat.com> wrote:
>> How do you "test it"? �
>
> I'll ask you the same question I asked Bob (btw, Bob, you seem to have
> run off)
>
> Can xEye be tested?
> If not, do you support using technology that can't be tested?
Hang on.
Let's discover what happened to the "public" bit of the goalposts before
moving on to your second question.
as
Mike Holmans
<jzfre...@gmail.com> tapped the keyboard and brought forth:
>On Jun 24, 7:59�am, alvey <al...@atattat.com> wrote:
>> On Wed, 22 Jun 2011 15:18:15 -0700 (PDT), jzfredricks wrote:
>> > On Jun 22, 6:12 pm, alvey <al...@atattat.com> wrote:
>> >> I can't think of a single company which relies
>> >> primarily on intellectual property
^^^^^^^^^^^^^^^^^^^^^
>> > Chances are you're not thinking then.
>>
>> So could you name one or two then?
>That's for NFL equipment - helmets.
^^^^^^^
Spot the similarity. I certainly can't.
Cheers,
Mike
--
jzfredricks
I've answered that. Over to you.
alvey
> On Jun 24, 8:13�am, alvey <al...@atattat.com> wrote:
>> On Wed, 22 Jun 2011 18:34:19 -0700 (PDT), jzfredricks wrote:
>>> On Jun 23, 10:53 am, cricketrulez <cricketrulez1...@gmail.com> wrote:
>>>> Strong statement, so what are your qualifcations in matters like this?
>>
>>> He has no qualifications,
>>
>> Sorry? Exactly what qualifications does one need?
>
> A reasonable amount of intelligence.
So that'd be No Qualifications Required then. Thanks for that.
>
>> Oh, and what did you say that yours were in respect to addressing Hawkeye?
>
> A reasonable amount of intelligence.
So that'd be 'I've got no relevant quals' then. Thanks for that. Job done.
>
>>> and his "they have a right to protect their
>>> IP" stance is, frankly, half-witted.
>>
>> Please explain the existence of legions of attack lawyers employed by
>> corporations who sniff the world for breaches.
>
> Come on mate, you don't strike me as being completely moronic. You can
> do better than this.
> If the testing does reveal some IP (and I don't think it will) that
> doesn't give xEye competitors a right to use that IP.
> The hoards of lawyers will, sadly, still be able to keep their jobs.
>
>>> Asking them to demonstrate that their product performs as well as
>>> their marketing brochure says it does is not a theft of their IP in
>>> any way, shape or form.
>>
>> Absolutely. But what has this to do with "public" testing?
>
> Thorough, public testing is the only way to remove all doubt about the
> efficacy of the system.
> Having said that, I could settle for thorough, indepenent testing. So
> long as the 3rd party was reputable *and* the testing methods (and
> full results) were made public.
> I don't consider the MCC independent enough, and I don't think their
> testing was good enough (I base this on the photos I've seen in the
> accuracy reports which I *assume* were taken during MCC testing).
And all this is "...the damned thing called work" hey.
as
jzfredricks
> Spot the similarity. I certainly can't.
You think helmet manufacturers rely LESS on IP than xEye?
I'd say the helmet manufacturers' IP lawyer budget would be 1000x that
of Hawkeyes.
Or is the emphasis on "primarily"?
alvey
> On Jun 24, 7:59�am, alvey <al...@atattat.com> wrote:
>> On Wed, 22 Jun 2011 15:18:15 -0700 (PDT), jzfredricks wrote:
>>> On Jun 22, 6:12 pm, alvey <al...@atattat.com> wrote:
>>>> I can't think of a single company which relies
>>>> primarily on intellectual property that allows, what is it again,
>>>> "thorough, public testing" of its product.
>>
>>> Chances are you're not thinking then.
>>
>> So could you name one or two then?
>
> [http://www.nocsae.org/MediaKit/2011/NOCSAE%20News%20Release%20on
> %20Data%20Request%2001%2024%2011%20FINAL.pdf]
Here's one that's more relevant!
www.tinyurl.com
Get back to me with a TinyURL. All I can get out of the above mad woman's
is something from a high school.
alvey
Ahh. You mean this; "Having said that, I could settle for thorough,
indepenent testing."
Good stuff! We're making progress now!
So you're dropping the ludicrous "public" from your mantra then. Yes?
alvey
jzfredricks
No. I'd much prefer public, and will continue to "demand" it.
Your turn to answer my question (re can it be tested?)
alvey
Save the TinyURL, although please use it in future, I found the link. And
there's no similarity between the two cases that I can find.
From your link I quote from 5.1, 'Summary of Test Method' [AS: On helmet.
And this is the doc 'Standard Test Method...']
"A headgear is positioned on a headform and then dropped in order to
achieve accepted impact velocity within 3% of the theoretical free fall
velocity for that drop height."
http://tinyurl.com/666fzkq
How is this relevant to IP?
alvey
alvey
Your choice. It'll make you look breathtakingly unworldly, but it's
entirely your choice.
>
> Your turn to answer my question (re can it be tested?)
I have no idea. (And only fractionally more interest). But I don't see why
not. I assume that it's based on sets of algorithms that *qualified* people
could analyse & report on.
alvey
jzfredricks
I actually don't think it IS relevant to IP, just as I don't think
testing of xEye has anything to do with IP. The "IP will be stolen"
thing is YOUR absurd position.
The similarity is this;
Both xEye and helmet manufacturers have genuine IP protection needs,
but both should have their product thoroughly and publicly tested.
This testing will not devalue their IP, nor expose it to more abuse/
theft than they're already exposed to.
IP is simply NOT an issue at this stage, and it's only you who keeps
on bringing it up.
Having said that... I wouldn't mind discussing how deeply their
techology (specifically their software) is looked at. This would
remove any tampering fears.
jzfredricks
> > Your turn to answer my question (re can it be tested?)
>
> I have no idea. (And only fractionally more interest). But I don't see why
> not. I assume that it's based on sets of algorithms that *qualified* people
> could analyse & report on.
You've half answered my question.
Do you support using technology that CAN'T be tested?
There are 2 main elements to xEye;
a) the actually readings of the position of the ball
b) the prediction software
I mainly want A tested, as everything else stems from it. I have more
faith in B than A, and if I had to pick between them I'd want A tested.
alvey
> On Jun 24, 10:41�am, alvey <al...@atattat.com> wrote:
>>> Your turn to answer my question (re can it be tested?)
>>
>> I have no idea. (And only fractionally more interest). But I don't see why
>> not. I assume that it's based on sets of algorithms that *qualified* people
>> could analyse & report on.
>
> You've half answered my question.
> Do you support using technology that CAN'T be tested?
That would depend entirely on what it's used for. If it was a fuel
injection system then 'No'. However, HE isn't quite as life-threatening as
an untested fuel injection system and even you admit that it's better than
Ashocka de Taufel, so 'Yes', I'm perfectly happy to see an "untested" HE
let loose on the open cricket field.
>
> There are 2 main elements to xEye;
> a) the actually readings of the position of the ball
> b) the prediction software
>
> I mainly want A tested, as everything else stems from it. I have more
> faith in B than A, and if I had to pick between them I'd want A tested.
I don't care.
as
jzfredricks
so 'Yes', I'm perfectly happy to see an "untested" HE
> let loose on the open cricket field.
This strikes me as strange, especially considering your "the officials
are influenced by the BCCI" stance.
Only through testing can we remove doubts about the efficacy of the
system, including tampering etc.
war...@bbgsports.com
The rumour I have heard this week and it has been confirmed by two
different sources, is that Tendulkar has been down at Hawkeye's office
in the UK looking at the tracking system. Don't know anything more
than that.
Warren
alvey
> On Jun 24, 11:03�am, alvey <al...@atattat.com> wrote:
> so 'Yes', I'm perfectly happy to see an "untested" HE
>> let loose on the open cricket field.
>
> This strikes me as strange, especially considering your "the officials
> are influenced by the BCCI" stance.
I'm not surprised that you get struck. It appears to have happened quite
often.
Again, what would you pick as the easier to "tamper" with;
a) An umpire
b) An xEye product.
> Only through testing can we remove doubts about the efficacy of the
> system, including tampering etc.
Yawn.
(Perhaps it's time that you set it to music?)
alvey
cricketrulez
Warren, Could you please shed some light on the topic of validation of
the predictive aspect of HE/VE?
Available public information indicates that HE testing was done using
a bowling machine. Is this correct? could you comment on potential
limitations of using a bowling machine?
cricketrulez
alvey, Here are the issues you have evaded so far
1) Validation of HE: Has it been done?
2) IP issues with testing/validation of HE: What kinda of IP? patents?
copyrights? trademarks? trade secrets?
3) You dabbled a bit on the WADA whereabout topic : Does WADA
compromise its rules depending on who they are dealing with?
jzfredricks
wrote:
> The rumour I have heard this week and it has been confirmed by two
> different sources, is that Tendulkar has been down at Hawkeye's office
> in the UK looking at the tracking system. Don't know anything more
> than that.
I hope their office has a cricket field!
Warren, what's your thoughts on the testing of HotSpot? Specifically;
a) would you object to independent testing of HS by a reputable 3rd
party?
b) would you object to the testing methods and results being made
public?
c) would this kind of testing expose your intellectual property to
theft or some kind of abuse?
d) has the ICC demanded such testing?
cheers
cricketrulez
Warren,
I have a question regarding cost. The publicly available cost for HS
is about $10K US per day of cricket. Will you be able to confirm that
this number within an order of magnitude of the correct amount.
war...@bbgsports.com
Its been very interesting to read everyones comments in relation to
the ball tracking.
I try very hard not to make any comments about competitors systems as
I think it is unprofessional.
What I will tell you is this....when Hawkeye emerged on the scene
around 2000-01 I was very impressed with the product from a television
perspective, it was innovative, fresh and eye catching. The one
element that struck me as a concern was....would people believe
it....emphatically. I wasn't sure as I felt there was too many
unknowns.
This convinced me to look for something a bit more believable. A year
or so later I saw some infrared vision and thought people would
believe infrared vision more than a Hawkeye type of product and hence
we headed down this route. I guess my initial reactions were probably
correct.
Consequently, I haven't paid too much attention to how the ball
tracking systems work as its not my field of expertise. About 12
months ago the Ch9 guys asked me to help out the Virtual Eye people as
they wished to switch from Hawkeye to Virtual Eye (aka Eagle Eye). I
have found the Kiwi's very upfront about the limitations of their ball
tracking. It was their idea to try and merge Virtual Eye and Hot Spot
into Hot Track to try and improve their accuracy for LBW prediction.
They also upped their frame rates from 120fps to 230fps to also help
the situation.
Cheers
Warren
war...@bbgsports.com
I was contacted originally by the MCC about the umpire technology and
they informed me that they were doing some testing of technology for
the ICC. They mentioned Hawkeye, Virtual Eye and Hot Spot.
I know that Hawkeye and Virtual Eye have both been tested by the MCC
in the field. Up-to-now they haven't requested that they test Hot
Spot, but I have a feeling that they have visited us in the BSkyB
truck a couple of years ago and watched us operate the system. Will
need to check with my guys on this one as I wasn't there at the time.
This was 2-3 years ago.
As far as 3rd party testing is concerned, I think we need to progress
a little further down the track and see if the ICC mandate the system
next week at their AGM in Hong Kong. If we can progress through this
process it maybe possible. Any public testing will disclose many
intellectual property concepts that most likely wouldn't be
commercially prudent to disclose.
My gut feeling is that the DRS process is a highly political process
between the Cricket boards....there are still many stages to navigate
through before it gets accepted across the board. Who pays for it will
be a big issue if be get over the current ICC/BCCI impasse.
Cheers
Warren
war...@bbgsports.com
Currently US $10k per day is our rate for a 4 camera setup (as per Ch9
and BSkyB).
The rate for a two camera setup is US $6k as per SkyNZ.
I would assume the norm outside of UK/Aus will be a two side-on camera
setup.
Just putting the finishing touches on the two new cameras and they
will be shipped out to the UK in a week or so. I am very happy with
the improvements and hope you guys will be too.
First day at Lords we will run the new cameras side-by-side to the old
cameras and compare. Hopefully if its a big difference then we will
switch over from day two or three onwards.
Warren
jzfredricks
wrote:
> Just putting the finishing touches on the two new cameras and they
> will be shipped out to the UK in a week or so. I am very happy with
> the improvements and hope you guys will be too.
Looking forward to it!
> First day at Lords we will run the new cameras side-by-side to the old
> cameras and compare. Hopefully if its a big difference then we will
> switch over from day two or three onwards.
No chance of seeing it in the SL ODIs?
ps: Lord's (apostrophe). You don't want to upset the fuddyduddies!
cricketrulez
Thanks Warren.
But do you realize how much trouble your are with the HE evangelists
on RSC ;-). You have become an undercover agent of BCCI now!
Regards,
CR
war...@bbgsports.com
Cameras are sitting right next to me. They normally take 8-10 days
before they are delivered to Sky in the UK.
I will watch my spelling next time.
jzfredricks
wrote:
> Any public testing will disclose many
> intellectual property concepts that most likely wouldn't be
> commercially prudent to disclose.
Like what? :)
Is the IP related to the accuracy of HotSpot and/or technical details,
or would it be other stuff (ie user interfaces of software,
algorithms, setup etc).
If the latter, I imagine you have IP disclosure issues at the games,
too, no? Is only HotSpot staff allowed near your systems at the games?
war...@bbgsports.com
Actually I am not really pro-or-anti ball tracking. I was just giving
the background on how hawkeye has affected what we have done in the
past several years.
I have never said that HS is perfect and therefore can't expect the
same of the ball trackers as I don't think there is ever going to be
100% in DRS. The right people have to make the right decisions about
whether the technology will be more of a plus for the game or more of
a negative.
If they decide Yes then they should have some very thorough testing
and publish the tolerances including HS and ball tracking.
Warren
war...@bbgsports.com
Definitely software interfaces, algorithms, configuration and general
operation.
I am lucky that I have a bunch of guys that have worked for me for
between 7-15 years so we have a very tight group.
Warren
alvey
> Any public testing will disclose many
> intellectual property concepts that most likely wouldn't be
> commercially prudent to disclose.
Praise be to the Lord Dennis!
Bless you Warren. May all your spots remain hot for much longer than the
average.
alvey
alvey
Oh FFS. A grammar snit? Well done.
alvey
> On Jun 23, 6:03�pm, alvey <al...@atattat.com> wrote:
>> On Thu, 23 Jun 2011 17:52:38 -0700 (PDT), jzfredricks wrote:
>>> On Jun 24, 10:41�am, alvey <al...@atattat.com> wrote:
>>>>> Your turn to answer my question (re can it be tested?)
>>
>>>> I have no idea. (And only fractionally more interest). But I don't see why
>>>> not. I assume that it's based on sets of algorithms that *qualified* people
>>>> could analyse & report on.
>>
>>> You've half answered my question.
>>> Do you support using technology that CAN'T be tested?
>>
>> That would depend entirely on what it's used for. If it was a fuel
>> injection system then 'No'. However, HE isn't quite as life-threatening as
>> an untested fuel injection system and even you admit that it's better than
>> Ashocka de Taufel, so 'Yes', I'm perfectly happy to see an "untested" HE
>> let loose on the open cricket field.
>>
>>
>>
>>> There are 2 main elements to xEye;
>>> a) the actually readings of the position of the ball
>>> b) the prediction software
>>
>>> I mainly want A tested, as everything else stems from it. I have more
>>> faith in B than A, and if I had to pick between them I'd want A tested.
>>
>> I don't care.
>>
>> as
>
>
> alvey, Here are the issues you have evaded so far
"evaded"? I think not.
>
> 1) Validation of HE: Has it been done?
Don't care. It doesn't matter to me.
> 2) IP issues with testing/validation of HE: What kinda of IP? patents?
> copyrights? trademarks? trade secrets?
Don't care. (And Warren has crunched this clueless tangent).
> 3) You dabbled a bit on the WADA whereabout topic : Does WADA
> compromise its rules depending on who they are dealing with?
I don't recall saying anything about the WADA whereabouts clause.
Hope this clarifies things & helps you along. (The edge is ===> )
alvey
cricketrulez
Hell, you are the one to talk! want me compile your off topic focus of
language errors.
jzfredricks
wrote:
> Definitely software interfaces, algorithms, configuration and general
> operation.
I don't think any of that would need to be exposed in testing.
None of that is related to the accuracy of the HotSpot results. That
is more related to the How.
Would you consider the technical limitations of HotSpot to be IP? More
to the point, the *published* technical limitations. (ie you might say
"can detect to 0.05 of a degree" where the real value might be 0.01.
I would understand you not wanting to run tests that expose the real
limitations, but I think you should run tests that verify any
published limitations.
Mike Holmans
<jzfre...@gmail.com> tapped the keyboard and brought forth:
>On Jun 24, 4:54�pm, "war...@bbgsports.com" <war...@bbgsports.com>
>wrote:
>> Definitely software interfaces, algorithms, configuration and general
>> operation.
>
>I don't think any of that would need to be exposed in testing.
Really?
You have a system which has to be set up, and the only people
qualified to set it up are those who work for the supplier. You then
have to have it operated in order to test it, and the only people
qualified to operate it are those who work for the supplier. Allowing
anyone else to do either would mean compromising commercially
confidential intellectual property.
You keep ranting about independent 3rd party testing. How can you get
independent 3rd party testing when the supplier has to set up and
operate the equipment?
Cheers,
Mike
--
jzfredricks
> On Fri, 24 Jun 2011 00:54:58 -0700 (PDT), jzfredricks
I've been over this.
With xEye, the accuracy tests are more important than seeing the
predictive elements (ie software). Same goes for HotSpot.
I've also mentioned tampering. This is a big issue we've only touched
on.
If tampering is an issue, then I again say that cricket's right to
know it's tamper-proof far outweighs the technology provider right to
protect their IP.
It's like voting machines - they should only use open source
software.
if (HomeTeam == "England")
{
SlightlyAdjustThePitchLocationTowardsInLine();
}
^^^ if THAT code is in the xEye software, do you think we have a right
to know?
jzfredricks
> if (HomeTeam == "England")
> {
> SlightlyAdjustThePitchLocationTowardsInLine();
>
> }
>
> ^^^ if THAT code is in the xEye software, do you think we have a right
> to know?
If you do actually answer this question, I'm expecting you to say "no,
no one has a right to see their source code as it's their IP".
alvey
"...your off topic focus of language errors."????
But whatever. If you're compiling a list of errr, something, then please
take your time and do it properly. I'd recommend a couple of years as an
absolute minimum.
eod
alvey
> On Jun 24, 4:54�pm, "war...@bbgsports.com" <war...@bbgsports.com>
> wrote:
>> Definitely software interfaces, algorithms, configuration and general
>> operation.
>
> I don't think any of that would need to be exposed in testing.
Please stop. You are making a fool of yourself to the extent that it's
embarassing.
alvey
You're gibbering now. Turn the pc off. Go on a holiday. A longgg holiday.
jzfredricks
You do realise the owner of HotSpot just said this;
Vijay
Thanks for that, Warren. I am sure most of us here would look forward to
the new system Hot Track.
As it stands today, I assume ball tracking and prediction technology is
not something that you would be right up there when it comes to
believing its results as a fan? I haven't referred to any competitors
here so you can give us an indication of where you stand :-)
Vijay
Good luck for that! And we will look forward to seeing the new tech
cricketrulez
> On Fri, 24 Jun 2011 00:54:58 -0700 (PDT), jzfredricks
Mike,
Guess you decided to come to his thread after all. Few points
1) This thread is about HE and not HS.
2) We (meaning myself, JZF, Bharat and BCCI etc) are the not the only
ones who have reservations about HE's performance, reliability and
accuracy.
3) Independent "validation" can still be carried out even if the
vendor does all the set up and operation. If you are not sure how,
I'll be happy to elaborate.
4) the technology/IP behind HS is the hardware and results are an IR
image of what happened, not what might have happened. The technology
of HE is the software which predicts what might have happened. I don't
need to look at the code base to or the algorithm to measure it
performance.
5) If independent validation will compromise IP of HE then there is no
way it should be used.
cricketrulez
> > alvey, Here are the issues you have evaded so far
>
> "evaded"? I think not.
> > 1) Validation of HE: Has it been done?
>
So what are doing in these topic which you don't care about? Trolling?
> > 2) IP issues with testing/validation of HE: What kinda of IP? patents?
> > copyrights? trademarks? trade secrets?
>
> Don't care. (And Warren has crunched this clueless tangent).
You brought the tangent in. Warren was talking about HS a highly
specialized hardware software combo, which does not predict what might
have happened. HE on the other hand is a software which makes
predictions on what might have happened using run of the mill video
cameras.
Apple to to Oranges between technology of HS and HE
> > 3) You dabbled a bit on the WADA whereabout topic : Does WADA
> > compromise its rules depending on who they are dealing with?
>
> I don't recall saying anything about the WADA whereabouts clause.
As usual, you did poke you head in that thread about which you didn't
know a whole lot and didn't really care
> Hope this clarifies things & helps you along. (The edge is ===> )
>
> alvey
Lets see
A professional expresses the same reservation about HE as me, JZF etc
have done
Who do you think the edge is with?
who do you think is the crazy eyed creationist? HINT: It is most
probably the guy call for implementation of uproven technology becos
he hates BCCI.
alvey
This has naught to do with your original denigration of my point about IP.
Let's recap your exchange on that subject with Warren:
Blockhead: Would this kind of testing expose your intellectual property to
theft or some kind of abuse?
Warren: "Any public testing will disclose many intellectual property
concepts that most likely wouldn't be commercially prudent to disclose."
Blockhead: Like what?
Warren: Definitely software interfaces, algorithms, configuration and
general operation.
In a cricket analogy... You're standing at the crease long after being
caught head-high at mid-on.
alvey
In Bne, post-scripting; Don't bother to respond to my request for a company
which allows "thorough, public testing" of it's product. This game is over.
alvey
> On Jun 24, 12:39�am, alvey <al...@atattat.com> wrote:
>
>>> alvey, Here are the issues you have evaded so far
>>
>> "evaded"? I think not.
>>> 1) Validation of HE: Has it been done?
>>
>
> So what are doing in these topic which you don't care about? Trolling?
Why would one troll in these issues? Or "these topic" even? There's nothing
worth catching.
>
>>> 2) IP issues with testing/validation of HE: What kinda of IP? patents?
>>> copyrights? trademarks? trade secrets?
>>
>> Don't care. (And Warren has crunched this clueless tangent).
>
> You brought the tangent in. Warren was talking about HS a highly
> specialized hardware software combo, which does not predict what might
> have happened. HE on the other hand is a software which makes
> predictions on what might have happened using run of the mill video
> cameras.
>
> Apple to to Oranges between technology of HS and HE
Errr. I bought the point in about Blockhead's braying on about "public
testing" and how no company that I could think of does that because of IP
issues (among other reasons). I believe that Warren's posts emphatically
agreed with me.
>
>>> 3) You dabbled a bit on the WADA whereabout topic : Does WADA
>>> compromise its rules depending on who they are dealing with?
>>
>> I don't recall saying anything about the WADA whereabouts clause.
>
> As usual, you did poke you head in that thread
Indeed, but you made specific reference to the 'wherabouts' clause. A
subject that I;
a) have no recollection of commenting on.
b) have no recollection of a 'WADA whereabout' Subject.
c) have no interest in.
d) can't see where the fuck it fits into *this* Subject.
> about which you didn't
> know a whole lot and didn't really care
Am beginning to yawn now.
>
>> Hope this clarifies things & helps you along. (The edge is ===> )
>>
>> alvey
>
> Lets see
>
> A professional expresses the same reservation about HE as me, JZF etc
> have done
You think that!!! Crikey! Your English comprehension is far worse than I
thought.
>
> Who do you think the edge is with?
lol. (Someone who cares could enlighten you)
> who do you think is the crazy eyed creationist? HINT: It is most
> probably the guy call for implementation of uproven technology becos
> he hates BCCI.
Yep. It's logical that you're a BCCI apologist.
And while dropping HINTS is the flavour du jour... I'll only be responding
to anything of yours in future that's not rubbish. Life's too short etc etc
as
jzfredricks
> This has naught to do with your original denigration of my point about IP.
I'll ask you the same question I asked Mike. Let's see if you copy him
and run away...
Is it ok if xEye software has this code in it?
if (BowlingTeam == "India")
{
GiveItOutIfCloseAndNoOneWillNoticeWeCheated();
NoOfExpectedPaymentsFromBCCI++;
}
Mike Holmans
I do not propose to follow you down this lunatic road. When Alvey
calls your whining block-headed, it's an insult to blocks. Your lines
of argument have now become as pathetic as the twats who went on and
on about Obama being a foreigner.
Cheers,
Mike
--
jzfredricks
> On Fri, 24 Jun 2011 02:30:08 -0700 (PDT), jzfredricks
So, you don't have a worthwhile response to a worthwhile question.
That code COULD exists, and you know it.
I'll leave the question open Mike, but I have little faith that you'll
answer it.
alvey
> Your lines
> of argument have now become as pathetic as the twats who went on and
> on about Obama being a foreigner.
Obama *is* a foreigner. Not just to my fellow Australians, but to his
chosen profession. He must be from a different place, or possibly planet,
to those creatures that are currently slithering to the front of American
politics. They're scary dumb.
alvey
Bob Dubery
> On Jun 23, 1:46 pm, Bob Dubery <megap...@gmail.com> wrote:
>
> > 1) We're not talking Hawk-Eye only here. Another system is in use and
> > that would have to be verified too.
>
> True. I think cricket's right to know the technology works trumps any
> IP issues (not that testing will steal the IP in the first place)
>
> > 2) HE uses input from multiple fixed cameras, so you'd need to have
> > multiple sets of input data and be careful with the editing
>
> and?
>
> > 3) Do you know what happened? Do you know exactly how much the ball
> > missed the stumps by? You know that it missed, but by how much?
>
> You seem to be hinting that xEye CAN'T be tested.
> Do you support using technology that can't be tested?
I'm saying it can't be verified to the nth degree. Interestingly
though we are seeing refinements of the xEye's use EG the condition
that surprised Dhoni in the CWC - if the point of contact is more than
was it 2.5 meters in front of the stumps then xEye is not able to give
a sufficiently reliable prediction of what would have happened.
It would be interesting to know who is imposing these restrictions and
why. They may be arbitrary. They may have come about because of
benchmarking. They may have come about because cleverer buggers than
me have looked at the method used and figured out the margin of error
and how it changes with distance to travel.
>
> > 4) You'd need to be sure of the calibration and alignment of the
> > cameras used to capture the test footage. This is not like calibrating
> > a speed trap where it's a simple question of "did the object cover
> > that distance in that time?"
>
> You seem to be hinting that xEye CAN'T be tested.
> Do you support using technology that can't be tested?
>
> > 5) All of this achieves not very much because you still haven't
> > answered the real question: Is HE more accurate and/or reliable than a
> > human umpire. And which umpire - annoyingly they don't all perform
> > the same and don't perform the same all the time.
>
> How can that be answered without seeing testings?
How do you test the humans? Wire electrodes into their brain and feed
images in and ask them say if the ball would have hit the stumps or
not.
>
> > The more I argue about this, the more I come round to thinking that
> > it's good as it is. The technology is only supposed to be there to
> > correct howlers.
>
> Can you please define a "howler" for us?
That's a better question than perhaps you think. What's important, I
think, is that if a team thinks the umpire is making bad decisions
they can call for the review. So UDRS would have come in handy in
situations like that England v SA match with a clearly incompetent
umpire who'd just got off a plane and got bullied (by both sides) into
a rash of bad decisions.
I'm just not convinced that the teams use it for that. I think they
may be using it to try to get an extra edge against a key batsman or
in desperation to keep the 10th wicket pair going a little longer.
Which is their right - it's a resource that they have and they have to
manage the use of it.
>
> > This is why technology was explored in the first
> > place - not to address hair's breadth decisions but for occasions when
> > an umpire screws up or is unsighted.
>
> Then why do we trust xEye to make hair's breadth decisions?
We do?
> We currently trust it to PERFECT accuracy for 2 readings - pitch
> location and impact locations for when the ump has given it out.
I'm not sure that it's being presented as unfailingly accurate.
Also the locations that you mention are a matter of RECORDING, not
prediction. They should be easier to do to a high degree of accuracy.
Those would be fairly easy to test by comparing the video to the xEye
trace of WHAT happened from the time the ball left the bowler's hand
to the point of impact. Easy peasy stuff that.
When Dhoni was told that xEye wasn't relied on if the point of impact
was more than 2.5 (or whatever) meters ahead of the stumps that didn't
mean that the system hadn't tracked the ball up to point of impact, it
meant that there was an increased probability in the predictive part
of the system.