Dear readers, sincere apologies for an error in the first table. In the Zimbabwe column - for the period 1930 - 1989 there should obviously be nothing since they had played no tests in this period.

Hawk-eye is the answer for this. Hawk eye removes the element of uncertainity . Earlier if the chances were 20% out , umpire will give benefit of doubt and say not out 100% of time. But now umpire will give out 20% time as he know hawk eye will support him. Also earlier umpires gave out only if they thought the ball will hit stumps on full, but hawk eye will give out even if the edge of ball is hitting the stumps , meaning even the 4 th stump and sligtly more height than stumps is acceptable. If you were to go purely by what hawk eye predicts, you will get still more lbw decisions in favour of bowlers

Very nice analysis...wonder why there is still a big difference between Host LBW% and Opposition LBW% after 1999. I think one factor you are missing here is the location where the matches are being played. You may want to see, for example, if there are more LBWs in general in India which will reflect in a high apparent Host LBW% for India. You can easily correct for this potential bias.

Since this game has absolutely no margin for error for bowlers,the leg stump rule should be cancelled,i.e.batsmen should be given out LBW even if the ball pitches outside leg-stump.That,I think,will solve a lot of LBW problems.& then,like run outs,umpires should not hesitate going to the 3rd umpire if they have doubts about the height & line of impact.I think these 2 changes would make the game fairer & easier for umpires.

It would have been more interesting to hear what Venkat feels on your statistical analysis. Does he attribute increase in LBWs to any particular factor ? One reason i could think is as you said umpires have become more lineant in applying benefit of doubt. Probably with replays they think that many of their doubts aren't correct and hence tend to favour the bowler the next time around. Are foreign teams less inclined to pad play in sub-continent than before ?

To see effect of neutral umpires, it might be more pertinent to study decade-wise LBW % and see whether there has been a rise from 90s onwards. For eg., if there was a difference between 70s and 80s and not much between 80s and 90s, then it indicates other factors at work. I would be happy to see if 90s are really the turning point as concerns LBW decisions. So some numbers about 60s,70s,80s,90s would help in this regard. This could also throw other interesting things like has % risen steadily or abruptly etc ?

Another interesting fact justifying increase in LBW decisions is that out of first 25 bowlers with highest % of LBW dismissals, the bowler to have debuted earliest is Imran in 1971. And the Top 6 are Kumble,Murali,Warne,Akram,Mcgrath, Younis. And are Bangladesh and Zimbabwe influencing this statistic heavily ?

I think umpires have in a way adjusted to the fact that now the game is heavily tilted in favor of batsman, and rightly so. With covered pitches, helmets, better quality pads and protection for every part of the body, and most importantly technologically superior bats with which it is considerably easier to hit the ball out of site. I think cricket administrators have to do something to make the contest even. Why do we have so many 50+ average batsmen in the world, almost 3 batsmen in the top 4 international teams are averaging in excess of 50 nowadays.

Hawk-eye can be erroneous at times but since no other options left we could go with hawk eye. But I could remember that Third Umpire referral system for LBW was enforced in a champions trophy[not sure] and believe me every referral was declared out, so they dropped the whole system. So in my opinion, make some solutions and before implementing it into an international series, try it out in a domestic set up. But nevertheless the doubt always goes in favour of the batsman with naked eye umpiring and goes in favour of bowlers with technologies. As simple as that.

in most cases umpires are right

Nobody says hawk-eye is a perfect solution. The LBW rule by definition requires you to guess where the ball would have gone had it not hit the batsman. Obviously, any system will be debatable. What matters is that Hawk-eye is better and more consistent than the present system (i.e.: an umpire standing 22 yards away).

If you think hawk-eye gives too many LBW decisions in favor of the bowlers, the solution is simple: reduce the size of the stumps in the hawk-eye system. Personally, however, I feel that the ICC needs to give something back to the bowlers. I say leave hawk-eye as it is, and make it official!!

Nice article, there are a few reasons for an increase in the number of LBW's.There has been an increse in matches played therefore the standard has been higher[pratice makes perfect]. The other is that batsmen do not use their feet as much therfore more chances of being hit between wickets.There are many more express blowers[each team has at least one] and they[bowlers] bowl the reverse swing at great pace, which can get you LBW or even bowled much easier.Lastly there are nutral umpires who do not have that hometown bias.

From my cricketing days I know for a fact that certain batsmen are more susceptible to an LBW because of the way they played. I think the umpires of today have a fairly good idea with reference to international batsmen and their playing styles. This "presumption" might have an effect on the how umpires decide on LBW decisions.

Despite being a scientist I remain very skeptical of Hawkeye's accuracy. It seems as though in many instances a lot of information about late swing is completely unknown and Hawkeye does not provide error bars to put its predictions in some sort of context. I am often appalled by what Hawkeye will give, it strikes me as overly optimistic. Even interpreting its results with some conservatism has I think given umpires the confidence to give decisions that previously might have been regarded as marginal.

how accurate is hawkeye? my question is does hawkeye consider the bounce of the pitch becoz diff pitches offer diff bounce? if it does then it means hawkeye has to be modified before every game considering the pitch report!

further i think explaining the the low LBW percentage in aus and sa by using the bounce factor is erronous. the ball tends to swing more as in aus and sa than the subcont. so shouldnt that balance out the bounce factor. also for spinners the bounce factor in deciding LBW is invariably useless becoz more often than not the ball strikes well below the knee-roll.

i am in agreement with Tony-out of 100 Hawk-eye decisions,i am OK with just under 50%.Whenever there is late deviation or movement,seamer or spinner,Hawk-eye does not pick it up.It also tends to minimise bounce showing balls hitting when they would have passed over.If we let Hawk-eye decide,no innings will cross 150 runs

Has anybody measured the accuracy of hawkeye?

I had presumed that had been done and was seen as reasonably accurate. To use it on TV without any sense of the reliability of Hawkeye does seem to be the height of foolishness.

Types of bowlers would have an impact on these stats as well. Definitely in the post 1990 era, Wasim and Waqar got lots of lbws because of their bowling style and because Pakistan's fielding was atrocious.

Also bowlers from home teams are better adapted to get lbws in their conditions. Warne in Australia, Murali in Sri Lanka and Kumble/Harbhajan in India would have got more lbws as compared to touring teams.

I do not understand why it is such a big problem to gauge the accuracy of the Hawk-Eye. A large number of balls in any given match go through to the keeper without hitting anything. The trajectory of these deliveries can be compared to those predicted by the hawk-eye and a standard deviation can be calculated. This gives the accuracy of the Hawk eye. Any decisions which are within the margin of error should not be given out when using the hawk eye.
In addition, to reduce this margin of error, a few deliveries can be bowled prior to the start of the game to tune the hawk eye to the pitch and adjust the parameters to the bounce of the wicket.
When there is a will there is a way.

If the law was different pre-1970 surely this data should be put in a different category? Also I'm not sure when it was (I seem to think in the early part of the 20th century) but previous to this be ball had to pitch in line with the stumps, i.e. couldn't pitch outside off, to be out LBW.

LBW decisions are always debatable, and will be so in future too. We are burdening the poor Unpires with so may parameters to consider in a split second ( and before that he has to look at the bowlers foot to see whether it is No Ball too) and expecting him to make the right decision too all the time. My Suggestion is rather outrageous. Scrap LBW rule altogether. And instead make catches from pads legal, whether the ball hit the bat or not. The advantage is the batsman will think not twice but several times before using the pads as a line of defense. And the umpires will be unburdened of this thankless job of adjudicating LBW decisions and the game will become more interesting.

Cricket is fast turning into an batsman game, bowlers only job is to bowl. if there is nothing in favour of bowlers in future we will see robots to bowl

I have no faith in Hawkeye, as it never shows a change in trajectory if there is an inside edge off the bat. So how do I know where it starts predicting the trajectory from? Probably only 100% accurate at saying where the ball pitched, nothing more.

I would trust Hawk-eye more if what we saw on TV had error bars on it. The more frames that the computer has after the ball pitches must increase the accuracy. Similarly, the further the ball has to travel after it strikes the pad, the less accurate will be the prediction.

Scientists and Engineers know that every measurement has an associated error. If we are going to see the ball flicking the top of leg on Hawke-eye, then let's also see the associated error bar with that estimation. Only if > 95% sure that the ball would have struck the stumps should Hawke-eye be used to give a batsman out.

I am concerned though about late swing. Hawke-eye doesn't seem to deal with this well. The physics of swing is still a matter for research, so it stands to reason that it may not be modelled well.

I was pleased to read this, "He will make the same decision whether the batsman is a No.1 or 2 or whether he is 9, 10 or jack."

I have become convinced that the tail enders are given out erroneously much more often that the top order batsman. Often it seems that the umpire is of the opinion that the umpire is umpiring the batsman rather than the position of the ball, pad and stumps.

Nipun at May 10, 2009 6:32 PM - allowing LBW for a ball pitching outside leg stump would completely and utterly ruin the game. Captains could set a 9-0 leg side field and bowl negatively at leg stump and outside waiting for the batsman to miss a ball that would hit the stumps. That's the reason the rule stands, and it will never be removed.

this gives structure to "facts" previously known but not formalised eg how many times (if any) has Miandad been given LBW in Pakistan by non neutral umpires?

You note in your introduction that there was a rule change increasing the chances of LBW around 1970 (I haven't checked that date). In that case, it would be interesting to compare LBW incidence after 1970 (e.g. 1970 - 1989) with more recent rates. To some extent, that would also help control for changes in bowling style since 1930, which as others have pointed out are part of the explanation for increased incidence.

this gives structure to "facts" previously known but not formalised eg how many times (if any) has Miandad been given LBW in Pakistan by non neutral umpires?

In sub-continents, the low bounce and turn would result in more LBW decisions. Moreover, spinners bowl more than pace bowlers do and the probability of LBW is much higher than any other dismissal. Nothing is best than the human eye which picks up the real motion. Best umpires are "best" becuase they pick ip up cleanly by the human eye.

I am also not a big supporter of Hawk-Eye and I also don't like commentators dissecting an umpire's decision after seeing 6-7 replays from different angles in slow motion. At the time of criticizing the umpire, they tend to forget that umpire had on a split second to decide. Umpiring is the toughest thing to do on cricket field and people generally don't think much about umpires work.

Has the increase in LBW's had a significant impact on the number of "Bowled". In the 80's, off spinners couldn't buy an LBW but got plenty of bowled. Now spinners regularly get LBW's, making lfe a little fairer for them.

I guess you need to take into account the amount of shots today's players play across the line when compared to 1970's or 1960's. I rarely saw Gavaskar or Allan Border or Boycott play across the line.

I used to be in favor of umpires giving the benefit of the doubt to the batsmen at all times, because thats how the game has been played. However, in recent times, all the technological improvements favor the batsmen, with improved bats, better protection, and batter friendly wickets. The game is unknowingly favoring the batsmen, and if you were to include LBW benefit of the doubts to them as well, there is really no chance left for the bowlers. I say make hawkeye mandatory for each game and allow the onfield umpire to consult with the third umpire for every decision. It may be controversial, but atleast it'll give the bowlers something to counter the increased favoritsm the batsmen have enjoyed throughout the history of the game.

Indeed a nice dissection of LBW. Also the scrutiny of umpiring decision has increased many folds since the 90s.
Few of thoughts on Hawk eye :-
I do not believe Hawk eye is the way to go for LBW decisions. First it does not take the swing of the ball bowled along with wind speed (Always bowling against the wind & with the wind behind are two separate things). Second if bowler is bowling his first over & if bowler is bowling his 23 rd over, the stamina will be different & so will be the bounce, which bowler would extract from the pitch. Third it does not consider the height from which ball is bowled. It measures it as standard. The release of the ball for Ishant Sharma would be so different than that of say Amit Mishra.

i'm afraid that making catches off the pads legal dismissals will only encourage bowlers to neglect the stumps and the edges and just aim at batsmen a la jardine. will that law include thigh pads and forearm pads? what about the helmet? batsmen will spend their days ducking and covering. mind you, fielders will be more on their toes.
i've always thought the lbw restriction on balls pitched outside leg was unfair to leg spinners and lefty offspinners. make it legal to dismiss batsmen who pad up and offer no shot to legside balls, as long as they're stuck in line, same as for offside balls.
oh, and get rid of hawkeye altogether. save it for commentators and technogeeks. let the umpires do their jobs. if they have to be perfect, then the players have to be perfect too, and we know they aren't.

isnt this a factor that batsmen have started playing across the line and against the spin far more since the 1980's?
Also i notice that hosts are still getting favored hugely in the sub-continent. Since the umpires are neutral, perhaps it is just that guests misjudge the bounce here (it is lower), and get struck on the pad more in line with the stumps. We need a country-wise breakup to be able to judge.
Finally: if LBW's increased as a percentage, clearly at least one other mode of dismissal has DECREASED correspondingly. Any number of factors can then be cited. Maybe bats are heavier and grounds are smaller, so less catches in the deep. Maybe the current cricket balls dont swing as much - fewer teams tend to have 5 slips and 2 gully sort of fields set up.
Maybe tailend batsmen have gotten better, thanks to helmets they dont run away towards squareleg - and so the number of "bowled"s has decreased? Need more data!

I don't support Hawk-Eye. Once Kumble had bowled a batsman out, the ball clipping the top of the stumps. Hawk-Eye showed that the ball would have missed the stumps by an inch. On-field umpires are the best judges of these decisions.

Hawk-eye is the solution. It is consistent for both teams. It doesn't depend on a particular type of appealing or the mood the umpire is in at the time of decision or perhaps a bit of bias also and off course no benefit of the doubt. It is repeatedly said in the media - it all evens out. Well it doesn't, because the timing of the right or wrong decision has a lot to do with the outcome of a particular match and also the series. Hawk-eye it is for me.

Hawkeye for me seem to take into account the bounciness of the pitch. How many times have we seen hawkeye analysis for the balls pitched on the same length but bounced differetly. The trajectory doesn't change a lot for a short distance between the pad and the stump.

Nice article - I've always been fascinated by the LBW law, something that is extremely difficult to explain to Americans, since there is nothing comparable to it in common sports over here. As a mathematician, I expect HawkEye to have basic triangulation functionality, regardless of the height of the cameras. As long as the spin is gauged accurately, the technology will correctly predict the flight to within an inch.

Also note that Sunny Gavaskar is listed twice on the infrequently dismissed list.

I played first grade cricket here in Australia for over 20 years as an opening fast bowler. I received and I also had turned down many debateable LBW decisions. Likewise I had suspect "catches" awarded and dead set clangers declined. Never once did I criticise nor abuse the umpire. (Batsmen however, are a different story!)

The umpire is the boss and his decision is final. It should be accepted, regardless of whether the bowler or the batsman disagrees with it. That's the fundamental rule of our game (and most other games) and it seems to be forgotten.

The last thing I want is for the game I've played and loved for so many years governed by a computer and played by automatons. Players are allowed to make mistakes and so should the umpires.

Of course I agree that Test umpiring should be of the highest calibre and every effort should be made to ensure as few mistakes as possible are made. But let's not make things too precise and roll with the punches huh?

I'm amazed by the comments complaining that 'Hawk-eye doesn't taking into account swing/height of delivery/tightness of the bowler's underwear'. Whenever I see it used on TV, I can see the path of the ball from delivery in 3 dimensions (including height), where it pitches, and how the ball deviated off the pitch and through the air to the point it reaches the batsman. At that point, the colour of the line changes to indicate that the rest is predicted. So the data on late swing is included before it reaches the batsman, and would be incorporated into the prediction of where the ball will go. Not to say that Hawk-Eye is 100% accurate, but if calibrated correctly and provided the appropriate data, it should be more accurate than a human with parallax error standing 22 yards away.

Like Warnesie I too have played cricket at a high level, high enough to warrant county standard umpiring. I have also played in club matches where the batting side have provided the umpires from their team. Much as I agree with the the idea of preserving the traditions of cricket, any tradition which cannot change with the times will fade into obscurity. Even in a country which is as tradition bound as my own, those traditions must be suited to the times. Look at the Royal Family. They must change in order to preserve those traditions. I cannot agree that there is anything against the spirit of cricket in asking for a second opinion. Umpires have always been allowed to do it. Giving the players the opportunity to say "Sorry ump but I got/didn't get an edge on that one and I think someone else should have a look at it", is well within the "Spirit of the Game" and in no way diminishes or demeans it if done properly. Cricket has Laws not Rules. As such appeals should be heard

There seems to be a misunderstanding of the way a ballistics path projector (like Hawke-eye) would work.

All you need to determine where a projectile will end up (assuming no air resistance) is two positions and a time between them. That's just high school maths. So, if you have two frames AFTER the ball has pitched, then that is all you need for that crude analysis. Your only error would be how well the positions were determined from the two frames... and wind resistance. Oh, the spin also matters – that's how a spinner gets dip, with overspin.

Now, with three frames you should be able to do a better job of even the wind resistance because you know how the path of the ball is deviating from the 2-frame projected parabola and you can fit a curve to it.

How tired the bowler is, the state of the pitch have no affect whatsoever on how good hawke-eye's projection will be.

Late swing is hard because you are predicting a big movement from a small one, so small errors accumulate.

IAAP

Thank you, terrific article.

Could they not use stump cameras for operating Hawk Eye, instead of cameras mounted high up? That should increase the accuracy of Hawk Eye considerably. I am not an engineer, so I don't know if this is a practical solution or not. If it is, then maybe the rule could be amended for the third umpire to automatically call the on-field umpire for each lbw appeal and weigh in on it. This would even out the game a little towards the bowlers.

many umpires these days don't go by the "batsman gets the benefit of the doubt" idea which is WRITTEN INTO THE LAWS OF THE GAME.

IMO, these umpires should simply be chucked out. all umpires make mistkaes. they -
a) give batsman out when they were not out
b) give batsman not out when they were out

if you are following the RULES, then the number of b) should be much higher than a). i'd suggest that the type of errors umpires make be broken down along these two categories, and if b) type errors aren't a lot higher than a) type, the umpire be put aside.

Bucknor and Bowden are two of the worst umpires I've ever seen.

trouble with Hawk-eye is that, if implemented, it'll change the way the game is played. Batsman routinely go down the wicket to spinners and pad the ball away, knowing that they are safe from LBW. Hawk-eye might well give those out.
having said that, perfect or not, Hawk-eye probably already makes far fewer errors than human umpires do.

An excellent article. Having seen so many dubious lbw dismissals (the worst being Gilchrist twice been given out lbw for 0 and 0 during a Test in India)two neutral umpires were a necessity.

So Hawkeye is not 100% reliable. But if the degree of accuracy is close to 100% then a simple solution would be to only allow lbw dismissals when no part of the ball is above bail height or outside off/leg stump.

I think there are a couple of factors that contribute to these stats. Firstly, fewer tests were played in the sub-continent pre-1989 (certainly in the 30's, 40's and 50's)which is where LBW decisions are more frequent. Secondly, the period 1930-35 had an LBW law where the ball had to pitch in line with the stumps, which made LBWs much harder to get. Furthermore, the period 1935-70 had an LBW law where a batsman could not be out to a ball pitching outside off stump unless he did not offer a shot. Again, this made LBWs harder to get than under the present law. These are the main factors pre and post 1989, not the umpires.

the article,as we wrote it,has a lot of triggers on umpiring competency,consistency,predictability.Propensity of current batsmen to fall LBW more than earlier top batsmen,neutral umpires taking away home advantage in the sub-continent etc.It is sad to see it being reduced to largely a debate on Hawk-Eye

Shafeen,
Could you please provide the law number and clause in the laws of cricket, where it is stated "the batsman gets the benefit of the doubt". In my capacity as an umpire and as a trainer of new umpires, I have read and studied the laws extensively. I will be very surprised if you can prove that this benefit of the doubt notion is actually documented within the laws as you so loudly shout.

That is all.

I do agree that Hawk-eye is not 100% accurate. But I think we need to use it in the referral system. In the second test between south africa and australia, sa's first innings, it took 10 minutes to confirm Amla was lbw after he didn't accept the umpires decision - ven though it was plumb.

Spending 10 minutes looking at side on relays to argue whether the ball is clipping the bails or not is stupid. Hawwk-eye might back a small misjudgement, but a) he shouldn't have let the ball hit his pad in the first place, and b) it makes the game quicker.

I am all for a 3 minute rule. If the 3rd umpire can't mae up his mind within 3 minutes the original decision stands.

In support of Dave's comment on 11 May: Pre 1937 the ball had to pitch wicket to wicket for LBW to be given, a completely biased favour to batsmen compared to thereafter. So Bradman's data should not be included.

it's a very good analysis.Thanks too both of you.
I do not want to go by the Hawk-eye option as I like the presence of human aliment in cricket as that also gives life to the game. In a game there is scope for the batsman, ballers,fielders so why not for Umpaires. Bringing to mouch technology may make it machenical.

This is a nice, thought provoking article. However, though you have presented a lot of data, raised some questions, and presented Venkat's views on umpiring and LBW decisions (though not his opinion on the reason for more such decisions going in favour of the bowler), you have not offered your own opinions on the matter. For that reason I find the article incomplete.

My opinion, which come only from watching cricket on tv and reading articles such as yours, is that batsmen of the present era are less patient (look at the decrease in the number of match saving innings) & willing to take more risks to dominate the bowling. There is a greater tendency to play across the line to open up the leg side and in that process batsmen are more vulnerable to the LBW dismissal. Added to this is the fact that throughout the world, the pitches are reputed to have slowed down, which probably means that they have a lower bounce than they did 30 years ago and earlier.

I have two queries regarding matters you raise here:

(a) Has there been any talk of modifying the `pitching outside the leg stump' rule when batsmen employ the `reverse sweep'?

(b) Have tests been conducted or statistical data obtained concerning the snick-o-meter's ability to distinguish between the noises made by (a) a snick, and (b) the bat rubbing against pad or ground?

Hi Raghu uncle, Nice article...I loved the analysis and summary.

from,Jagan (netherlands)