On Thursday, March 17, 2022 at 2:21:12 PM UTC+10, nos...@lane-farm.fsnet.co..uk wrote:
> On 15/03/2022 21:37, jack fredricks wrote:
> > On Tuesday, March 15, 2022 at 11:23:22 PM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
> >>> I inferred from this single line quote what I said - they (like me), don't like 2 identical deliveries resulting in 2 different DRS results (based purely on the on-field call).
> >> Yes, I got that. What I was wondering was whether they would agree with your solution.
> >>
> >> If the system is to be changed to one where (at least in the case of LBW), the on-field umpire's decision is completely disregarded once it is reviewed (and personally, I don't mind either way about that), then I suspect that the solution that would result in the fewest complaints would be the one where it's out if DRS shows any part of the ball hitting the wicket (if impact is in line, etc), and it's not out if it doesn't.
> >
> > I think this would result in too many dismissals, and would quickly change.
> >
> > I like the "50% of the ball" threshold. It "feels" fairest.
> Do you think there are too many dismissals already? Your proposal would
> obviously reduce the number. It could also encourage more pad play, as
> it would effectively make the wicket about 6.2 inches wide, rather than
> 9, and 1.4 inches shorter, for LBW.

I'm not sure. I think it's a great "crime" to give a batsman Out when they're Not, than to deprive a bowler of a fair wicket.
So I'd rather ere on the side of Not Out.

On Thursday, March 17, 2022 at 2:30:32 PM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
> > My comment was certainly not aimed at you.
> > For a decade or more now MH has refused to concede that umpires can be in a state of doubt w.r.t parts of the LBW decision.
> > He insists only 2 states of mind exists; umpire is "certain it was hitting" (or pitched inline etc), or "certain it was missing".
> I can't say that I recall that.

Let's see if he can be bothered correcting his error above regarding this.
There's always search, too. I can't be arsed.

On Thu, 17 Mar 2022 11:39:51 +0000, Mike Holmans <spam@jackalope.uk>
wrote:

>On Wed, 16 Mar 2022 17:17:44 -0700 (PDT), Hamish Laws
><hamish.laws@gmail.com> wrote:
>
>>On Thursday, March 17, 2022 at 10:39:59 AM UTC+11, jzfre...@gmail.com wrote:
>>> On Thursday, March 17, 2022 at 7:43:35 AM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
>>> > On 15/03/2022 21:43, jack fredricks wrote:
>>> > > On Monday, March 14, 2022 at 7:44:09 PM UTC+10, Mike Holmans wrote:
>>> > >> If the umpire gives "not out", he thinks it would have missed (or it
>>> > >> pitched or impacted in the appropriate place for a not out).
>>> > >
>>> > > This drivel again.
>>> > >
>>> > > For how long now has cricket had an umpiring foundation of "if there's doubt, not out"? 100 years? More?
>>> > >
>>> > > An umpire can give it not out if there's doubt. This doesn't mean he thinks it would have missed.
>>> > Fair point, which I must admit I forgot (maybe because I was in a hurry
>>> > when I posted).
>>> My comment was certainly not aimed at you.
>>> For a decade or more now MH has refused to concede that umpires can be in a state of doubt w.r.t parts of the LBW decision.
>>> He insists only 2 states of mind exists; umpire is "certain it was hitting" (or pitched inline etc), or "certain it was missing".
>>
>>How about you stop trying to guess other people's thoughts?
>>Because you're shit at it
>
>And his whole thesis about DRS depends on his guess about what every
>umpire thinks.....
>
>Cheers,
>
>Mike

We would have in our Christmas quiz a screen with a load of lbw
shouts. Everyone was an umpire and almost every answer was the same.
Any variations were nearly always around bat pad decisions. The clips
would have been chosen for that difficulty.
It's a procedure, a series of decisions made tick box style and if at
any stage a box doesn't get ticked you don't think any more about it
otherwise you might miss something important.
I've seen umpires getting in a fluster because they are too busy
giving lbw not out they forget the ball is still live.

There are or was a few years ago, umpire observers. They would watch
the positioning of the umpires and their 'fieldcraft' technique, how
they handed over the pitch to each other after an over and how they
deal with agressive appeals etc.

max.it

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

On Thursday, March 17, 2022 at 9:39:55 PM UTC+10, Mike Holmans wrote:
> And his whole thesis about DRS depends on his guess about what every
> umpire thinks.....

You've pulled that out of your monumental ass*, as I am the only person in this thread saying that what the umpire thinks should be ignored.

* or, again, deliberately lied about my position, as you are want to do. You spiteful man.

On 17/03/2022 12:14, Mike Holmans wrote:
> On Wed, 16 Mar 2022 21:26:02 +0000, David North
> <nospam@lane-farm.fsnet.co.uk> wrote:
>
>> On 15/03/2022 15:18, Mike Holmans wrote:
>
>>> My conclusion is that deferring to the onfield umpire is as equitable
>>> as any other practical scheme one can devise, and comes without the
>>> disadvantage of reducing the onfield umpiress status as arbiters.
>>
>> I can't say that I find arguments about the umpires' status/authority
>> very persuasive. As I think I have said before, they are there for the
>> benefit of the game, not the other way round (and, btw, DRS has
>> increased the number of umpires required).
>
> I admit that the authority of the umpires' argument amounts to being
> able to say "...and it's crispy/green/tastes of saffron!" as a bonus
> and isn't of great weight. But it does make it crispy.
>>
>> One obvious advantage of not deferring to the on-field umpire is that
>> the technology is far less open to accusations of bias (although the
>> human inputs to it and interpretation of it are not immune), and cannot
>> possibly be influenced by such things as how loudly or how often the
>> fielding side appeal.
>
> If the technology were 100% accurate, that would be fair enough. But
> where the technology literally cannot give a definitive answer because
> it didn't capture the precise occurrence in question but evidence on
> either side, is that just "not out"?

I was thinking specifically of the ball-tracking aspects of LBW. I
certainly wouldn't want to rely entirely on the technology to decide low
catches, for instance.

--
David North

On 15/03/2022 21:39, jack fredricks wrote:
> On Tuesday, March 15, 2022 at 11:57:22 PM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
>> On Monday, 14 March 2022 at 21:55:37 UTC, jzfre...@gmail.com wrote:
>>> On Monday, March 14, 2022 at 8:29:44 PM UTC+10, John Hall wrote:
>>>> Even if the
>>>> technology is less than perfect, I feel that it's still more accurate
>>>> than almost all - possibly all - umpires.
>>> I believe everyone, even the luddites, think this.
>>> My beef is the *inconsistent* results.
>>> 2 identical deliveries, 2 different final DRS results.
>> Who says that the 2 deliveries were identical - ball-tracking? There's a margin of error, so maybe they weren't actually identical, and that's why the umpire gave one out and the other not out.
>
> Does it matter?
> We KNOW in theory this is possible.
> I don't think it should be possible in theory.

It is also possible in theory to have two identical deliveries, with the
on-field umpire giving the same decision for both, and ball-tracking
producing two different results.

So you are apparently saying that it is OK to have two different
outcomes from two *actually*-identical deliveries because of the margin
of error of ball-tracking, but it is not OK to have two different
outcomes from two *apparently*-identical deliveries (according to
ball-tracking) because the on-field umpire believed that they were
different.

--
David North

On Friday, March 18, 2022 at 5:35:06 PM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
> On 15/03/2022 21:39, jack fredricks wrote:
> > On Tuesday, March 15, 2022 at 11:57:22 PM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
> >> On Monday, 14 March 2022 at 21:55:37 UTC, jzfre...@gmail.com wrote:
> >>> On Monday, March 14, 2022 at 8:29:44 PM UTC+10, John Hall wrote:
> >>>> Even if the
> >>>> technology is less than perfect, I feel that it's still more accurate
> >>>> than almost all - possibly all - umpires.
> >>> I believe everyone, even the luddites, think this.
> >>> My beef is the *inconsistent* results.
> >>> 2 identical deliveries, 2 different final DRS results.
> >> Who says that the 2 deliveries were identical - ball-tracking? There's a margin of error, so maybe they weren't actually identical, and that's why the umpire gave one out and the other not out.
> >
> > Does it matter?
> > We KNOW in theory this is possible.
> > I don't think it should be possible in theory.
> It is also possible in theory to have two identical deliveries, with the
> on-field umpire giving the same decision for both, and ball-tracking
> producing two different results.
>
> So you are apparently saying that it is OK to have two different
> outcomes from two *actually*-identical deliveries because of the margin
> of error of ball-tracking,

There's absolutely nothing we can do about that. I can't foresee a time when xEye doesn't have a margin of error.
We'd also never actually know about it, unless there was a more accurate device used side-by-side.
Essentially it's ok because margin of errors will always exist. They are a reality we have to live with.

> but it is not OK to have two different
> outcomes from two *apparently*-identical deliveries (according to
> ball-tracking) because the on-field umpire believed that they were
> different.

No need for the "apparently". My real life example was not for apparently similar deliveries.
But yes. If there were two apparently identical deliveries, according to ball tracking, they should result in the same final outcome, and I am not "ok" that there are 2 different outcomes.

On Friday, March 18, 2022 at 4:51:16 PM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
> I was thinking specifically of the ball-tracking aspects of LBW. I
> certainly wouldn't want to rely entirely on the technology to decide low
> catches, for instance.

I would. If such a technology existed and it was thoroughly and publicly tested.
I don't consider today's TV replays such a technology.

On Saturday, March 19, 2022 at 3:31:49 AM UTC+10, Mike Holmans wrote:
> The present system works very well at answering the question it's
> supposed to answer. The objection being made is essentially that the
> review system is asking the wrong question.
>
> If someone wants to argue for a different question, as jzf clearly
> does, that's their right.

I don't believe today's DRS "works very well at answering" that question. I think it's pretty good, but can be much better. Keep in mind I am this ng's biggest DRS fan - I want, eventually, *all* appeals automatically reviewed by DRS. NB appeals!
I also think your summary of the question is wrong. The question is way more nuanced than that.

Today's question is not just "Was the umpire definitely wrong?"
It's "Was the umpire definitely wrong, and was the relevant player willing to ask that question given numerous limitations eg; 15 second window to consult and decide (and to be certain the umpire isn't going to initiate an Umpire review if possible), are they confident technology will definitively show the evidence they know to be true (eg will snicko show a noise they did/didn't hear), is their wicket valuable enough to risk a limited resource on (eg a #7 not using last review as his #3 captain is at the other end), furthermore was the player even allowed to ask the question (ie did they have reviews remaining)".

I don't think we can simply ignore confirmed umpiring mistakes outside of DRS and say "well, they're outside of DRS so they're not a DRS issue". To me that's like saying "the hospital was full, so the patient died on street.... as they didn't die in hospital it's not a hospital issue".

There was an LBW in T2 Aus vs Pakistan. A top order Pakistan batsman was given Out, and they didn't refer it. Later, HotSpot footage showed the ball had hit the glove before the pad.
If it was reviewed, we'd be saying "thank goodness for DRS, another howler overturned".
So... was it a howler or not? Is it only a howler if the player knows it to be a howler? This top order batsman obviously didn't feel the ball hit his glove. Is that his fault? Glove too thick? Hands too insensitive?
Does the definition of howler depend on each player's skill?

Howlers not referred to DRS are still howlers.

And yes, I think elite cricket would be better if a different question was asked. Mainly as I think "definitely" is redundant, or it gives "wiggle room" to allow for unjust decisions.

"Was the umpire wrong?" is all we need to ask. To me though, that's the same as asking "was it the correct decision under the Laws of Cricket?".

When DRS technology (eg TV replays, HotSpot, Snicko, Balltracking) shows, a mere 30 seconds later, that an umpiring mistake was "allowed to go through to the keeper", I think that's a huge mistake for cricket.
We have the ability to remove more howlers, but we're not.

As for "preserving the on-field umpire's decision when there's doubt"... it seems to me there's a hypocrisy here.

Take this example I've given many times;
OFU thinks "impact inline, missing stumps"
DRS shows "impact Ump's call, hitting stumps".

Natural justice screams to me that this should be Out.
The umpire gave it Not Out because they thought it was missing the stumps, but DRS proved them wrong.
However, the DRS result will remain Not Out, due to the "Ump's call" w.r.t Impact (which the OFU also though was Out).

On 18/03/2022 09:18, jack fredricks wrote:
> On Friday, March 18, 2022 at 4:51:16 PM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
>> I was thinking specifically of the ball-tracking aspects of LBW. I
>> certainly wouldn't want to rely entirely on the technology to decide low
>> catches, for instance.
>
> I would. If such a technology existed and it was thoroughly and publicly tested.
> I don't consider today's TV replays such a technology.

Agreed. I meant to specify _current_ technology.

--
David North

On Fri, 18 Mar 2022 21:33:02 +0000, David North
<nospam@lane-farm.fsnet.co.uk> wrote:

>On 18/03/2022 09:18, jack fredricks wrote:
>> On Friday, March 18, 2022 at 4:51:16 PM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
>>> I was thinking specifically of the ball-tracking aspects of LBW. I
>>> certainly wouldn't want to rely entirely on the technology to decide low
>>> catches, for instance.
>>
>> I would. If such a technology existed and it was thoroughly and publicly tested.
>> I don't consider today's TV replays such a technology.
>
>Agreed. I meant to specify _current_ technology.

Current technology is available to pinpoint accuracy for probably any
decision that is visible. It needs to be adapted to purpose and would
no doubt be better than the current tv DRS system, but the questions
I would have are.
Does cricket need this to happen?
Who is paying for it?

max.it.

--
This email has been checked for viruses by Avast antivirus software.
https://www.avast.com/antivirus

On 18/03/2022 20:21, jack fredricks wrote:
> On Saturday, March 19, 2022 at 3:31:49 AM UTC+10, Mike Holmans wrote:
>> The present system works very well at answering the question it's
>> supposed to answer. The objection being made is essentially that the
>> review system is asking the wrong question.
>>
>> If someone wants to argue for a different question, as jzf clearly
>> does, that's their right.
>
> I don't believe today's DRS "works very well at answering" that question. I think it's pretty good, but can be much better. Keep in mind I am this ng's biggest DRS fan - I want, eventually, *all* appeals automatically reviewed by DRS. NB appeals!

I can't see any reason to go any further than allowing all appeal
decisions to be challenged. There doesn't seem much point in reviewing
decisions that everyone accepts (apart from checking for a no-ball,
which happens anyway).

--
David North

On Saturday, March 19, 2022 at 7:46:51 AM UTC+10, max.it wrote:
> Does cricket need this to happen?

Why did we get the first iteration of DRS? Because the most basic of technology, TV replays, showed time and time again that umpires made bad mistakes.

At the end of the day, elite cricket is entertainment played for the benefit of the fans.
And fans were sick of these mistakes. So yes, elite cricket needs DRS.

I believe fans will also be sick of today's flaws in DRS (non referral of bad mistakes, limited referrals, ump's call), and eventually there will be changes to DRS to remove/reduce these.

I've not spent much time thinking about new DRS technologies.
I think we could automate front-foot no-ball calls.

As for existing technologies. They can all be improved.

> Who is paying for it?

ICC should.

On Fri, 18 Mar 2022 22:37:06 +0000, David North
<nospam@lane-farm.fsnet.co.uk> wrote:

>On 18/03/2022 20:21, jack fredricks wrote:
>> On Saturday, March 19, 2022 at 3:31:49 AM UTC+10, Mike Holmans wrote:
>>> The present system works very well at answering the question it's
>>> supposed to answer. The objection being made is essentially that the
>>> review system is asking the wrong question.
>>>
>>> If someone wants to argue for a different question, as jzf clearly
>>> does, that's their right.
>>
>> I don't believe today's DRS "works very well at answering" that question. I think it's pretty good, but can be much better. Keep in mind I am this ng's biggest DRS fan - I want, eventually, *all* appeals automatically reviewed by DRS. NB appeals!
>
>I can't see any reason to go any further than allowing all appeal
>decisions to be challenged. There doesn't seem much point in reviewing
>decisions that everyone accepts (apart from checking for a no-ball,
>which happens anyway).

As I suspected, jzf's ideal is that televised cricket should have no
umpires and be officiated from the control room. He may well wish to
deny it, but it's the logical consequence of his extreme view.

That's not a sport I would enjoy as much as the one we have.

I'm not against the odd beneficial tweak here and there, but having
spent my working life in the bowels of computers, I don't like the
machines taking over. I would need a LOT of persuasion that the
current system isn't basically right, and nothing I've seen so far
comes anywhere near convincing me that there's even an issue.

Cheers,

Mike

On Saturday, March 19, 2022 at 8:37:08 AM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
> On 18/03/2022 20:21, jack fredricks wrote:
> > On Saturday, March 19, 2022 at 3:31:49 AM UTC+10, Mike Holmans wrote:
> >> The present system works very well at answering the question it's
> >> supposed to answer. The objection being made is essentially that the
> >> review system is asking the wrong question.
> >>
> >> If someone wants to argue for a different question, as jzf clearly
> >> does, that's their right.
> >
> > I don't believe today's DRS "works very well at answering" that question. I think it's pretty good, but can be much better. Keep in mind I am this ng's biggest DRS fan - I want, eventually, *all* appeals automatically reviewed by DRS. NB appeals!
> I can't see any reason to go any further than allowing all appeal
> decisions to be challenged. There doesn't seem much point in reviewing
> decisions that everyone accepts (apart from checking for a no-ball,
> which happens anyway).

I don't recall saying I wanted to go "further" than reviewing all appeals.
What else could be reviewed?
Leg-byes when the batsman hit it? (and visa versa)
Short runs? Are they checked?
Letting 3rd ump decide on intent to play a shot?

I personally wouldn't mind anything being checked, so long as the game flow isn't interrupted too much.
In my previous posts re wanting all appeals DRSed, I did mention that for this to happen they have to be sped up. Plus some other things around time wasting. I've not bothered going into those details here.

On Saturday, March 19, 2022 at 8:52:24 AM UTC+10, Mike Holmans wrote:
> As I suspected, jzf's ideal is that televised cricket should have no
> umpires and be officiated from the control room. He may well wish to
> deny it, but it's the logical consequence of his extreme view.

How dare I want to increase the number of umpiring mistakes fixed by technology from "some" to "more".
My views aren't extreme.

And yes, I would be "happy" if there were no on-field umpires.
The only purpose they serve in the DRS world is to make decisions on intent, and I despise intent-based cricket Laws. Sadly I can't think of a way to get rid of them eg "was the batsman playing a shot" I think adds good balance to the LBW laws.

> That's not a sport I would enjoy as much as the one we have.

Many said that about any DRS.
Too bad.

On Fri, 18 Mar 2022 22:37:06 +0000, David North
<nospam@lane-farm.fsnet.co.uk> wrote:

>On 18/03/2022 20:21, jack fredricks wrote:
>> On Saturday, March 19, 2022 at 3:31:49 AM UTC+10, Mike Holmans wrote:
>>> The present system works very well at answering the question it's
>>> supposed to answer. The objection being made is essentially that the
>>> review system is asking the wrong question.
>>>
>>> If someone wants to argue for a different question, as jzf clearly
>>> does, that's their right.
>>
>> I don't believe today's DRS "works very well at answering" that question. I think it's pretty good, but can be much better. Keep in mind I am this ng's biggest DRS fan - I want, eventually, *all* appeals automatically reviewed by DRS.

It works very well at answering the question "Was the umpire
definitely wrong beyond all reasonable doubt?" If you deny that, then
you are unable to understand English.

>I can't see any reason to go any further than allowing all appeal
>decisions to be challenged.

If you do that, how do you propose to fit three T20 games into a
single day as currently happens on the Blast Finals Day when each
innings lasts three hours?

I realise that jzf only cares about the Ashes and just witters
aimlessly the rest of the time, but those of us who watch cricket on
TV just about every day may not appreciate the inordinate effort he
wants to have put in to achieve precision to his exacting
specifications. And I doubt very much whether those who are in the
stands will appreciate the endless hanging around rather than getting
on with the game, particularly with the shorter forms. The long form
is already far too slow, but dear old jzf wants to subject us to vast
amounts more examination of replays. Perhaps we should go down to 75
overs a day to accommodate him.

In the nightmare world of remote-controlled cricket a la jzf, we will
be bored rigid with 15-minute overs and the game will die a thoroughly
deserved death.

It ain't broke, and it don't need fixin'.

Cheers,

Mike

On 18/03/2022 22:53, jack fredricks wrote:
> On Saturday, March 19, 2022 at 8:37:08 AM UTC+10, nos...@lane-farm.fsnet.co.uk wrote:
>> On 18/03/2022 20:21, jack fredricks wrote:
>>> On Saturday, March 19, 2022 at 3:31:49 AM UTC+10, Mike Holmans wrote:
>>>> The present system works very well at answering the question it's
>>>> supposed to answer. The objection being made is essentially that the
>>>> review system is asking the wrong question.
>>>>
>>>> If someone wants to argue for a different question, as jzf clearly
>>>> does, that's their right.
>>>
>>> I don't believe today's DRS "works very well at answering" that question. I think it's pretty good, but can be much better. Keep in mind I am this ng's biggest DRS fan - I want, eventually, *all* appeals automatically reviewed by DRS. NB appeals!
>> I can't see any reason to go any further than allowing all appeal
>> decisions to be challenged. There doesn't seem much point in reviewing
>> decisions that everyone accepts (apart from checking for a no-ball,
>> which happens anyway).
>
> I don't recall saying I wanted to go "further" than reviewing all appeals.

That's not what I said. If the umpire gives a batter out and he doesn't
want to review it because he knows he's out, what's the point?

--
David North

On Saturday, March 19, 2022 at 3:50:37 PM UTC+10, Mike Holmans wrote:
> On Fri, 18 Mar 2022 22:37:06 +0000, David North
> <nos...@lane-farm.fsnet.co.uk> wrote:
>
> >On 18/03/2022 20:21, jack fredricks wrote:
> >> On Saturday, March 19, 2022 at 3:31:49 AM UTC+10, Mike Holmans wrote:
> >>> The present system works very well at answering the question it's
> >>> supposed to answer. The objection being made is essentially that the
> >>> review system is asking the wrong question.
> >>>
> >>> If someone wants to argue for a different question, as jzf clearly
> >>> does, that's their right.
> >>
> >> I don't believe today's DRS "works very well at answering" that question. I think it's pretty good, but can be much better. Keep in mind I am this ng's biggest DRS fan - I want, eventually, *all* appeals automatically reviewed by DRS.
> It works very well at answering the question "Was the umpire
> definitely wrong beyond all reasonable doubt?" If you deny that, then
> you are unable to understand English.

But I do deny that!

Take my example given a few posts ago re the LBW;
OFU thinks - impact inline, missing stumps
DRS shows - ump's call, hitting stumps

In my mind this shows, beyond reasonable doubt, that the umpire's original call was wrong.
However, for *simplicity*, DRS does NOT change the result.

DRS also has a MASSIVE limitation which means that question isn't always asked and answered. DRS relies on *players* to ask the question.

Take the other example I mentioned;
Pakistani top order batsman given LBW, not referred, replays before next ball show it hit the glove first.

DRS was capable of fixing this mistake, but did not attempt to do so.

As for the rest of you post... snipped...
I don't think more DRS would make cricket boring. Nor would the lack of umpires (btw, first time I've heard that umpires are part of the entertainment package).
I actually find the DRS reviews quite interesting. A little bit of technological suspense.

On Saturday, March 19, 2022 at 6:33:45 PM UTC+10, nos...@lane-farm.fsnet.co..uk wrote:
> > I don't recall saying I wanted to go "further" than reviewing all appeals.
> That's not what I said. If the umpire gives a batter out and he doesn't
> want to review it because he knows he's out, what's the point?

The obvious stepping stone to "refer all appeals" is "refer all Outs" (as this MIGHT aid the batting team more, debatable as umps might give more wickets, the fielding team should be given additional reviews).

All Outs. Even the bloody obvious ones eg caught at long on.
These obvious dismissals would take very little time to review. Check for no ball. Did batsman hit it to long on? Yes. Was the catch clean? Yes.
This could all be done before the batsman left the field.
There are zero excuses not to review.
Sometimes a batsman THINKS they're out, but they're not.
I recall one of the Akmal brothers being given out Caught Behind once and replays showed it missed the bat by almost a foot. Akmal walked. DRS was available. It was such an egregious decision that I've always wondered if both the umpire and the batsman were involved in match fixing.
Then last week the Pakistani top order batsman didn't review after being given LBW. It was plumb, but the ball brushed his glove first.

Good Test cricket is such a finely balanced game, that mistakes like this can change the entire outcome. And they are umpiring mistakes that are so easily avoided.
People will say "it's the batsman's fault for not referring".. but I think that's manifestly unfair.