***Official*** DRS discussion thread

[Scaly piscine]

Can't balls move in the air more after a certain while though.. Isn't that what was called swing or spin? And isn't that why it could not believe that Shane Warne delivery that bowled Strauss?

I believe the explanation was that it wasn't calibrated for that level of spin.

[weldone]

I wanted to post this for a while. I do support DRS fully, but with every incident of DRS failure my conviction is somewhat lessened (it'd be surprising if it didn't, isn't it? ). I don't think anyone here is opposing the idea of DRS itself, but only debating the modalities and admissibility of some technologies.

To that, I wanted to say that this whole issue reminds of the diagnostic tests problem in conditional probability. Let's say that 90% of umpires' decisions are right and 90% of them don't need to be reviewed. If your DRS technology can get 98% of the decisions right that sounds like a significant improvement. But what if the 90% that don't get reviewed are part of the 98% (very likely if only the marginal calls get referred)? Then out of the 10 in 100 that get reviewed, 2 are incorrectly decided by DRS. That is 20% error rate on the reviews. That is obviously not good enough to inspire any confidence!

Now if we indeed tolerate 2% error rate on reviews, we want our technology to be 99.8% accurate and not 98%. The exact numbers may differ, but that shows that we need our DRS technology to be very, very precise. With each failure that comes to light, I am not convinced that the prior accuracy is close to 99.8%. May be 97-98%, but that's not good enough.

Interesting point

[honestbharani]

I believe the explanation was that it wasn't calibrated for that level of spin.

Maybe.. but how many times have seen balls do a lot more, sometimes after passing the batsman, sometimes just when it nears the bat after pitching well short... I just have a hard time believing any software can consider that many variants and come up with definitive conclusions (that is, within 90% of success).. It is not blind faith. I work in these areas everyday and I see how often they get it wrong and how much improvements are attempted. Today everyone of these machines are controlled by human judgement. They do the automated tasks, they collect the data and they show us reports, but it is us humans who judge what they mean and come up with predictions on what would happen from there..

Hawkeye's 95% perfection?!? That's a total nonsense for a start. If you knew what you were talking about you would realise that. Is it 95% perfection if it gets 19 out of 20 right - competent umpiring level? Or is it the margin of error proportionate to the data? Or something else? Whatever it is I suggest you go and study the data yourself if you doubt it. That's what a scientist is supposed to do, not this blind faith crap.

Hawkeye doesn't need to have some sort of approximation of conscious thought towards seam, swing and so on. It can track the ball moving. If there was a 50mph crosswind it would track the ball moving and project the path. If the ball hits a crack half way down it can follow the path of the ball from where it bounced. Given the lbw rule it doesn't need to guess at the seam movement or spin if it hits the batsman on the full. If you asked Hawkeye to project the bath of the ball to the wicket-keeper then you'd have some errors and guesswork because in England the ball can swing well after it passes the batsman. But Hawkeye is dealing with something that is simply an extension of path it has tracked. The seam and spin has ALREADY happened. If the ball is a swinging full toss you can work out the swing and the lateral acceleration on the ball.

The only time guessing comes into it is when you have a gusty wind. That's when one of the variables changes - the variables themselves are shown by the ball moving and being tracked, the way it bounces, swings, seams etc. A gust of wind immediately after the ball hits the pad could cause a fractional difference to the direction. But that would be covered by margins of error anyway.

Again, you are assuming that the ball will never move more than what it moved at impact. An umpire can judge that, hawkeye cannot. And there are 3 or 4 people involved who work on Hawkeye and if and when they get it wrong, the human error there is gonna cause hell of a lot more damage than any umpire's error has. And I understand they do not stop the hawkeye's tracking at some random point and then use the predictive path to compare it on a match to match basis. Which means, all the experimentation and results provided can be of no use, if they had been so much as a 1mm displacement of one of their 6 tracking cameras, which is perfectly possible. And from reading up on the Hawkeye guy's PDF where he has shared his mail communications with Mickey Arthur and some screengrabs, it is even more obvious that they NEVER provide for exaggerated deviations at any point after pitching. They track the ball till impact and extrapolate from there.

[marc71178]

Can't balls move in the air more after a certain while though.. Isn't that what was called swing or spin? And isn't that why it could not believe that Shane Warne delivery that bowled Strauss?

Yes, and obviously the technology has not been updated in the last 6 and a half years at all

[marc71178]

Interesting point

And ultimately still a load of ****.

98% is still better than 90% ffs.

[8ankitj]

And ultimately still a load of ****.

98% is still better than 90% ffs.

Given a decision by DRS, there is 20% chance that it was incorrect!

If that doesn't drive home the point, consider a 90% vs 95% situation. 95% is still better than 90%?? What of the fact that only 50% of the decisions by the technology are accurate? You could just toss the coin rather than use the 95% accurate technology.

[GingerFurball]

Technology doesn't make mistakes ffs. And if the accuracy was as low as 50%, then it wouldn't be getting used.

You're building a straw man argument here.

edit: and ironically, it will be through the use of technology that you declare a decision to be incorrect.

[honestbharani]

Yes, and obviously the technology has not been updated in the last 6 and a half years at all

or maybe you just don't read what is posted and just pull out the rolleyes..

[honestbharani]

Technology doesn't make mistakes ffs. And if the accuracy was as low as 50%, then it wouldn't be getting used.

You're building a straw man argument here.

edit: and ironically, it will be through the use of technology that you declare a decision to be incorrect.

Just wrong on so many levels I am not even gonna bother..

[wellAlbidarned]

Again, you are assuming that the ball will never move more than what it moved at impact. An umpire can judge that, hawkeye cannot. And there are 3 or 4 people involved who work on Hawkeye and if and when they get it wrong, the human error there is gonna cause hell of a lot more damage than any umpire's error has. And I understand they do not stop the hawkeye's tracking at some random point and then use the predictive path to compare it on a match to match basis. Which means, all the experimentation and results provided can be of no use, if they had been so much as a 1mm displacement of one of their 6 tracking cameras, which is perfectly possible. And from reading up on the Hawkeye guy's PDF where he has shared his mail communications with Mickey Arthur and some screengrabs, it is even more obvious that they NEVER provide for exaggerated deviations at any point after pitching. They track the ball till impact and extrapolate from there.

The LBW law states that an umpire is supposed to assume that the ball will travel straight on from where the ball strikes the batsman. If he's trying to predict the swing or spin then that's poor umpiring.

[honestbharani]

The LBW law states that an umpire is supposed to assume that the ball will travel straight on from where the ball strikes the batsman. If he's trying to predict the swing or spin then that's poor umpiring.

Really? That is what the LBW law says?

Law 36: Leg before wicket (LBW). If the ball hits the batsman without first hitting the bat, but would have hit the wicket if the batsman was not there, and the ball does not pitch on the leg side of the wicket, the batsman will be out. However, if the ball strikes the batsman outside the line of the off-stump, and the batsman was attempting to play a stroke, he is not out.

Strangely enough, nothing mentioned about umpires having to assume the ball would continue straight on...

[weldone]

A genuine question here: What if a full toss from a spinner hits a batsman on the front-foot by the way? Assuming the ball was supposed to pitch before reaching the wicket, how is the DRS going to review the decision?

[8ankitj]

2.5m rule comes into play?

[wellAlbidarned]

Really? That is what the LBW law says?





Strangely enough, nothing mentioned about umpires having to assume the ball would continue straight on...

2. Interception of the ball
(a) In assessing points (c), (d) and (e) in 1 above, only the first interception is to be considered.
(b) In assessing point (e) in 1 above, it is to be assumed that the path of the ball before interception would have continued after interception, irrespective of whether the ball might have pitched subsequently or not.

z

[wellAlbidarned]

A genuine question here: What if a full toss from a spinner hits a batsman on the front-foot by the way? Assuming the ball was supposed to pitch before reaching the wicket, how is the DRS going to review the decision?

The DRS assumes that the ball would travel straight on from it's initial path with no deviation as the LBW law states. There's absolutely no chance of the ball bouncing over the stumps, if that's what you mean.