Challenger to the Duckworth Lewis Method and the Jayadevan Method

[yashsr]

Challenger to the Duckworth Lewis Method and the Jayadevan Method

Hello,

I and a friend of mine have developed a method which we believe is superior to the D/L method and the Jayadevan Method. We would like for help as to whom to contact to reach the administrators.. We've already emailed ICC, MCC and they've acknowledged but they're just playing passing the parcel....sending us to one another and sometimes giving us email id of some person who just doesn't reply.

We've already submitted the method to the MCA(Mumbai Cricket Association) and Lalchand Rajput(former India manager, former Mumbai Indians coach etc.) found it superb. He's trying his best to forward it to other parties.

For those interested, a glimpse of the method can be found at:
http://www.yankmodel.com/
Let me know what you think of this method.

BTW, I'm the same 'Yash' who had a username 'yash' on this website few years ago but forgot my password and not sure which email address I had to put up....so I had to create a new username. Sorry guys!

 

[Ljp86]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Welcome back yash, good to see you back on BigCricket.

The concept you have put up is very good from the quick read I've had. Not sure how well it will go against the Duckworth/Lewis method though.

 

[Boris]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

I've read through it all and it seems pretty good. I do agree there are flaws in both systems, the the Jayadevan method being more flawed than Duckworth/Lewis, but I don't think they are quite so bothersome as suggested.

I do have one question:
You have obviously tested this method quite a lot, and during that testing have you found any flaws, similar to the ones you found in the D/L and Jayadevan methods, that prove irrational?

You have found a lot of problems with the other two methods, so showing some of the flaws (every method must have them, no matter how small) isn't going to detract from your sales pitch too much.

I'm just interested to see if the 'flaws' in your system are as dramatic as the ones you found in the other two.

 

[yashsr]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Thanks Ljp86...if you go through my Documentation, you'll see how it stands against D/L.....D/L method is in my opinion is more flawful than the Jayadevan method

Boris, this rain rule is not an exact science and thus no rule could be flawless....But what I believe is the Yank method is as good a method can get. Our objective during creating the method was to overcome all the flaws of the other 2 methods. While doing so, we ourselves had a few new flaws which we've now rectified. It took about 2-3 months for the method to get ready along with the documentation.

We'll soon be updating the Documentation....we haven't published the actual method for obvious reasons which we will in the near future...it contains numerous factors and calcullations which will I'm sure scare a layman...what've published is just the comparative analysis...We also haven't published our NRR and Mandatory overs concept which are a gem and I'm sure ICC will take them once it reaches them...

What I'm interested in how do I reach the ICC. We've emailed the MCA, BCCI, ACC, ICC and MCC. We've also talked to Lalchand Rajput who was bloody impressed by our whole documentation.

Also, just for my curiosity, does anybody have an idea whether D/L do get paid for their method? Some kind of royalty perhaps?

 

[mas cambios]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

I'll start by saying that I've only had a quick look through and as such can't really comment on the method being proposed. I'll need to have a thorough read and digest the figures, something I don't really have time for at the moment.

However, in terms of what I think you're really after, i.e. getting your proposed method to the powers that be then I think you'll be in for at best a long,hard slog and at worst a maddening and and at times disheartening journey.

The biggest hurdle will actually be in getting your method noted by the powers that be. Personally, I think you're stuck between a rock and a hard place at the moment. You need local/small leagues to take the system on but most will shun it in favour of the system sanctioned by the ICC or National governing bodies (NGBs). In order to get the ICC/NGBs to sanction your system you'll need it to be used extensively and successfully in the aforementioned leagues.

You will need to keep banging on the doors of various leagues and hope your persistence pays off. One option could be to ask if they could run your rain method alongside the one they currently use, in order to afford you greater sample data and comparisons.

Secondly, D/L are fiercely protective of not only their system but it's status in world cricket. Once and if you reach the national boards, expect very audible shouts against your system from their supporters. They will look to pick your method apart and expose any and all flaws (or at least what they perceive to be flaws). Be prepared to have data and evidence to fully back up your argument and system - ensuring that you have a plethora of 'real world' examples to draw upon.

You'll need to work on getting as many 'voices' in the game to back you - from those at local levels upwards. It would also help to have Mathematician/Statistician peers review and comment on your work. Could it be worth submitting it to a journal (sporting/maths/stats) for review? It could be one way of gaining publicity although I can't comment on the cost or time factor for doing so.

I don't wish to sound negative but I do think it will be far from easy, even to get local recognition. However, I wish you luck in doing so and we at BigCricket will be happy to help out, if we can.

 

[yashsr]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Thanks Mas for the your valuable practical oriented ideas...most helpful! Its hard to f people who'll give practical ideas on how to present the method to ICC and I'm glad some one came up!

Let me know what you feel about the method also...how it deals with comparisons with other methods..I can assure you that D/L is probably the worst of the 3 methods...i.e. from D/L, Yank and Jayadevan which I've pointed out in the Documentation...Jayadevan fairs better than D/L but even it has its own flaws..to me, Yank method hardly has any flaws, if any...this is because we've done a lot of analysis and trial and error as well....months of hard work!

 

[Boris]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

I've just sat down and properly gone through all the examples shown on your website and I'm impressed.

I even went so far to look at the situation, without looking at the figures after it, then scratch up my own workings to come up with what I think would be the most logical, sensible and real world score that my own opinion could come up with and they were all quite close to the figures you presented. In my mind that makes a system that has common sense to back it up, one thing the D/L method doesn't have.

I'm a supporter of the D/L method, always have been and have alway argued for it, and this is the only thing I've seen to make me doubt it.

I'm not going to say just yet that it is better than the D/L method and should take over from it, but I say well done to you yash and I'm supporting your run at the big guns.

All the best to you.

 

[yashsr]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Thanks Boris...and stay tuned for more updates..

We're actually thinking of taking a copyright/patent of our method so that we can feel free in the discussing the method and how we derive those scores. Any ideas about this? And will it come under copyright or patent? And is it necessary?

 

[mas cambios]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

We're actually thinking of taking a copyright/patent of our method so that we can feel free in the discussing the method and how we derive those scores. Any ideas about this? And will it come under copyright or patent? And is it necessary?

Not sure on patent or copyright but YES, you need to sort something out.

 

[yashsr]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Not sure on patent or copyright but YES, you need to sort something out.

ok thanks

 

[Ljp86]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Okay I've had a better look at your method Yash and whilst you and your mate have done an excellent job I'm not really convinced. For the majority of the cases you have put up, the Duckworth-Lewis (D/L) and the "Yank" method seem to be reasonably smiliar which doesn't provide cricket administrators with a legitimate reason to switch methods. A few of the Duckworth-Lewis targets seem to be quite unreasonable and I will agree with you there but then there's a couple of the Yank method targets which seem unreasonable to me (examples 13, 16 and 17).

In my opinion, I think you're trying to include too many variables into your formula and I think that's producing a few target scores which are a bit awry. Perhaps if you could tweak it a bit more and focus on less variables, you might be able to produce a much more finer and more accurate product.

Overall though, I think you've done a really good job. All three methods are very good and if they were combined as one they would be much more superior than using only one of them. Another thing I'll say is that your method needs a new name, "Yank" seems a bit strange, perhaps combining your two last names would be a better option. I also agree with mas with regards to copyright, you need to start taking steps toward protecting it so your method can't be copied or stolen.

 

[yashsr]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Thanks ljp86, your comments as well as constructive criticisms are most welcome!

We've actually used linear interpolation until now but now we're using a more precise method of interpolation which should change some targets by a run...we've also made a few minor mistakes in our calculation which we'll change and post accordingly. For eg., the answer in Eg.16 will probably change from 128 to 131...

Did you go through the flaws and limiations of the other 2 methods? We're also modifying this section in our next update which should be in 2-3 days.

 

[Ljp86]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Thanks ljp86, your comments as well as constructive criticisms are most welcome!

We've actually used linear interpolation until now but now we're using a more precise method of interpolation which should change some targets by a run...we've also made a few minor mistakes in our calculation which we'll change and post accordingly. For eg., the answer in Eg.16 will probably change from 128 to 131... [?QUOTE]

Okay, fair enough. I look forward to seeing the updates.

Did you go through the flaws and limiations of the other 2 methods? We're also modifying this section in our next update which should be in 2-3 days.

Yes, I've been through those as well. Both methods have their limitations which you guys have pointed out.

I'd be very interested to see what you guys come up with when you have sorted the formulas out and are closer to providing the finished product. @|

 

[yashsr]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Yes, that should take some 2-3 days, but that won't make a major impact....target will change only for select cases and by about 2-3 runs except in 1 case where where it'll change by about 10 runs because of an error. What we'll be incorporating is a few more limitations of the other methods and modifying the existing ones...but again no major change.

If you see the method(which I obviously won't show you until it gets to the concerned parties), there is a scientific method by which all these targets are derived. Even a layman will understand the method and the factors, but it'll take time and application of mind to understand.

I'd love to share my NRR method and another method which we've evolved but it'll take time...we need to submit it to the ICC, BCCI etc. first. These 2 methods are a gem, specially the NRR method.

 

[Ljp86]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Good work Yash, I am looking forward to seeing everything once it is complete.

 

[yashsr]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Good work Yash, I am looking forward to seeing everything once it is complete.

Thanks Ljp!

BTW, if there is any case if anyone would like me to do a comparison with D/L and Jayadevan with Yank, I'd be happy to oblige, specially if its a real match situation although I wouldn't mind an imaginary example...and specially if you feel that D/L hasn't done justice to one team in that match....

 

[yashsr]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Another thing I'll say is that your method needs a new name, "Yank" seems a bit strange, perhaps combining your two last names would be a better option. I also agree with mas with regards to copyright, you need to start taking steps toward protecting it so your method can't be copied or stolen.

We actually found the name 'Yank' quite cool and in fact quite good. Is there any particular reason why we shouldn't use this name or just your preference. Like does it clash with some other thing where it is called 'Yank'? We actually thought of using 'gangrun' which is a combination of our last names but we found 'yank' much better.

We're trying to do something about the copyright thing...

And I'd have 2 more questions:
1) Just for my curiosity, does anybody have an idea whether D/L do get paid for their method? Some kind of royalty perhaps? What I know Jayadevan, at least in the initial stages wasn't paid by the BCCI and I'm pretty sure he isn't even now.(BCCI switched from using D/L to Jayadevan method). He actually got paid when the ICL(Indian Cricket League) took up its method but alas, it only lasted for a year or two..
It has taken a lot of effort in making this method and its always nice to know whether there is light at the end of the tunnel(even if you're not sure of reaching the end)! But probably recognition overshadows anything monetary if any method reaches the full distance.....

2) What sort of programming language should be used to develop a calculator of the method? I assume it depends upon the method???

 

[Ljp86]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

We actually found the name 'Yank' quite cool and in fact quite good. Is there any particular reason why we shouldn't use this name or just your preference. Like does it clash with some other thing where it is called 'Yank'? We actually thought of using 'gangrun' which is a combination of our last names but we found 'yank' much better.

We're trying to do something about the copyright thing...

Yank seems a bit of a stupid name to me, I'm not too sure if people will take you seriously with it. Also, "Yank" is a slang term to describe a person who's nationality is American.

And I'd have 2 more questions:
1) Just for my curiosity, does anybody have an idea whether D/L do get paid for their method? Some kind of royalty perhaps? What I know Jayadevan, at least in the initial stages wasn't paid by the BCCI and I'm pretty sure he isn't even now.(BCCI switched from using D/L to Jayadevan method). He actually got paid when the ICL(Indian Cricket League) took up its method but alas, it only lasted for a year or two..
It has taken a lot of effort in making this method and its always nice to know whether there is light at the end of the tunnel(even if you're not sure of reaching the end)! But probably recognition overshadows anything monetary if any method reaches the full distance.....

Not sure. My guess is that the ICC would have paid them some amount of money considering they are using Duckworth and Lewis' own academic material in order to determine targets and outcomes of games. I could be wrong though.

2) What sort of programming language should be used to develop a calculator of the method? I assume it depends upon the method???

No idea, you'll have to ask someone who is good with computer programming.

 

[Ankit Gangar]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Hello guys....this is ANKIT...the joint creator of the 'Yank model'....

This 1 is the most latest comparison 4 u to evaluate........


In the Friends Provident T20 2nd semi final between Nottinghamshire(Notts) and Somerset played at Southampton on the 14th Aug 2010, Somerset batting first scored 182/5 in 20 overs. It rained in the interval and left Notts only 16 overs to bat in which the D/L which used in the match set a target of 152 runs. It rained once again in the 13 overs in which Notts were 117/4 in 13 overs. D/L target for 4 wickets at the 13 overs was 121 runs. So Somerset won by 3 runs.

16 over Target
Yank Method 161
D/L Method 152*
VJD Method 160

*D/L method gives an irrationally low target.


13 over
for 4 wkts Target
Yank Method 124
D/L Method 121
VJD Method 116

Summary
Result
Yank Method Somerset won by 6runs
D/L Method Somerset won by 3runs
VJD Method Nottinghamshire won by 2 runs*

*VJD’s deflated target is because of the sluggish fall of wickets. At this point of the match, VJD’s fall of wickets is sluggish till the fall of the 5th wicket, then there is a major impact on the target on the fall of the sixth wicket. On the other hand if the team was 117/0 the target would have not changed even by a run(the difference would be in just decimals) which is hard to fathom.

And you can just see what a major impact this flaw has made to the match. If VJD’s method would have been used, Nottinghamshire would have won the semi-final and thus qualified for the final leaving Somerset out of the tournament which would not hav been fair at all.

 

[yashsr]

Re: Challenger to the Duckworth Lewis Method and the Jayadevan Method

Thanks for your reply.

Yank seems a bit of a stupid name to me, I'm not too sure if people will take you seriously with it. Also, "Yank" is a slang term to describe a person who's nationality is American.

Ah never knew that 'slang' thing before we named it. But its probably too late as we've emailed a LOT of people about our method and changing the name now would create unnecessary confusion.