Forecasting & Results - Analysis
#21
Ok! bplus

Maybe we have a Misunderstanding ? 

Yes! The (5) Results R1 to R5 are different in the Examples because it makes them different Result Finishers... Can you imagine 1st place came 1 then Second place came 1 Etc. ?

And yes Greyhound Races do finish in the exact Order sometimes like 1,2,3,4,5,6 The important thing is that they can finish in any Order like 3,2,1,5,6,4  or 4,6,5,1,2,3 Etc.

So back to the Forecasting Challenge I am putting forward...

F1 (Forecast-1) = R1 - 2  R2 - 5  R3 - 3  R4 - 1  R5 - 4  What Number (1 to 5) is R6 - ? 

Let's say we kept getting the same Forecasts & R6 Result then...

F1 (Forecast-1) = R1 - 2  R2 - 5  R3 - 3  R4 - 1  R5 - 4  R6 - 5 
F2 (Forecast-2) = R1 - 2  R2 - 5  R3 - 3  R4 - 1  R5 - 4  R6 - 5 
F3 (Forecast-3) = R1 - 2  R2 - 5  R3 - 3  R4 - 1  R5 - 4  R6 - 5 
F4 (Forecast-4) = R1 - 2  R2 - 5  R3 - 3  R4 - 1  R5 - 4  R6 - 5 

I am sure even a Very simple Forecasting Program would Forecast Correct the F5 - R6 Below...

F5 (Forecast-5) = R1 - 2  R2 - 5  R3 - 3  R4 - 1  R5 - 4  R6 - ? 

Now! I hope you can See why I am doing it the way I am ? Finding the Best Forecasting method is the Most Important thing! then we can move on to more Complicated Forecasting Challenges! Smile

Maybe you can Suggest One ?
Reply
#22
Sorry, for my lack of understanding as yet. I just had a new way to see the data of the tables I am looking.

The numbers are the dog 'names' not the F1 to F5 on the left side.

Each row or each column is a race? or both???

I have to study this with this new way of seeing, right now too many other things are calling me to get done. I remember many a new math course when I have been hung up by the most basic misunderstandings of what is being presented. Once I get that, it is usually smooth sailing after...
B += x
Reply
#23
No problem bplus I am also jumping the Gun a bit expecting people to understand what is in my Head! I would say I am a better doer than an ex-plainer Smile
I will get back to this later with some better clearer examples... I will write a Program where it does what I want, and the challenge is that you play against it trying to Out Forecast it while you both have the same Starting information!


Anthony.
Reply
#24
Anthony, sounds like fun!
B += x
Reply
#25
I have had a thought about what might get us on track.

We start with 5 dogs whose names are 1, 2, 3, 4, 5

We know nothing about their race record yet so we forecast the first race they run as
F1:  1 2 3 4 5   'dog #1 finish first, #2 second... who knows could happen?

The race #1 is run results were:
R1: 2 4 1 3 5  'first place to dog #2, 2nd to dog #4....

Now for next race, we have some info to make better forecast, we predict
F1: 2 4 1 3 5 ' will happen again, not likely but better guess than random

Race #2 results
R2: 4 1 2 3 5 'hey 2 right

I would table things like this:
F1: 1 2 3 4 5  R1: 2 4 1 3 5
F2: 2 4 1 3 5  R2: 4 1 2 3 5

Now we have a prediction problem, how do we forecast next race?
I would use weights system, the more recent the race the higher the weight mixed with an average of where the dog places usually.

I don't think my table system is anything like Anthony's even though my F's stand for Forecasts and my R's stand for (actual) Race or Results. In my table system, you can't do (overlap) F rows AND R columns.

So Anthony, where do, how do Actual Race results fit in (if they do)?
Can you show how you would table the scenario I have just presented?
B += x
Reply
#26
I would table things like this:
F1: 1 2 3 4 5  R1: 2 4 1 3 5
F2: 2 4 1 3 5  R2: 4 1 2 3 5

Now we have a prediction problem, how do we forecast next race? 
I would use weights system, the more recent the race the higher the weight mixed with an average of where the dog places usually.


Anthony.


OK! I just woke up to this! and one Good way is a "Progressive Score System" = P.S.S Smile


So every R1st Place Result Scores = 5  R2nd = 4 R3rd = 3 R4th = 2 R5th = 1


So the Score for R1: (2 = 5)  (4 = 4)  (1 = 3)  (3 = 2 (5 = 1)



Which would make the same Forecast as your Method of Better than Random! (Because we only have One Result) Sad

Now the Score for R2: (4 = 5) (1 = 4)  (2 = 3)  (3 = 2)  (5 = 1)


Which leads to the Forecast for F3:/R3: Using Both totals as ...


( 1 = 7) (2 = 8) (3 = 4) (4 = 9) (5 = 2)


And Now in the Correct Forecast Finish Order! ...


F3:/R3: 4 2 1 3 5  Smile
Reply
#27
Hi Anthony,

Your R3 looks darn good for Forecast 3 for Race 3, I would add to my table system like this
F1: 1 2 3 4 5 R1: 2 4 1 3 5
F2: 2 4 1 3 5 R2: 4 1 2 3 5
F3: 4 2 1 3 5

Lets say Dog #2 really hurt it's leg winning Race 1 and instead of healing, it gets worse:
Race 3 all the dogs pulled ahead 1 but #2 fell to bottom
R3: 4 1 3 5 2

So my table looks like this:
F1: 1 2 3 4 5 R1: 2 4 1 3 5
F2: 2 4 1 3 5 R2: 4 1 2 3 5
F3: 4 2 1 3 5 R3: 4 1 3 5 2

Now, will Dog #2 heal up by next race of stay injured and near bottom, will dog #4 tire of being number #1, not if he were human!

500 for 2, 4, 4
400 for 4, 1, 1
300 for 1, 2, 3
200 for 3, 3, 5
100 for 5, 5, 2


dog 1 = 400 + 400 + 300 = 1100/3 = 366.7
dog 2 = 500 + 300 + 100 = 900 / 3 = 300
dog 3 = 300 + 200 + 200 = 700/3 = 233.3
dog 4 = 500 + 500 + 400 = 1400/3 = 466.7
dog 5 = 200 + 100 + 100 = 400/3 = 133.3
check sum .................................... = 1500 total points for one race

F4 by straight averages = 4 1 2 3 5 moves 2 back up, not bad for a guess

What if we double weight of last race? 4 1 3 5 2
dog 1 = 400 + 400 + 300 + 400 = 1500/4 = 375 2nd
dog 2 = 500 + 300 + 100 + 100 = 1000/4 = 250 3|4 ?
dog 3 = 300 + 200 + 200 + 300 = 1000/4 = 250 3|4 ?
dog 4 = 500 + 500 + 400 + 500 = 1900/4 = 475 first
dog 5 = 200 + 100 + 100 + 200 = 600/4 = 150 5th
check sum ................................. 6000 = 4 race total points

My 4th forecast is 4 1 2 3 5 given the 3rd race was 4 1 3 5 2
B += x
Reply
#28
Greyhound Cracker for Windows Free

http://rc.rawinfopages.com/greyhoundcracker.html

The Program can be used to Pick (5) or More/Less Random Greyhounds...For Fantasy Tests! Smile and used as a Greyhound Race Forecaster.

About Roland Waddilove and RAW Computing...
[Image: me1.jpg]RAW Computing is the website of Roland Waddilove, a self-employed freelance writer and occasional programmer. I write about PCs and peripherals, Apple Macs, Windows, OS X and Linux, the internet, hints, tips and tweaks, software, programming, problems and solutions, tutorials, articles and more. Almost everything on this website has been designed, created and written by myself.
I currently use Windows XP, Vista, 7 and 8 on various PCs, OS X Lion on Apple MacBooks, and even occasionally use Ubuntu Linux. I also have an iPhone and iPad, and an Android tablet and Android phone.
I started out in the early 1980s with a Sinclair ZX-81, which was followed by a Jupiter Ace and then an Acorn Electron. In the mid 1980s I was the editor of Electron User and then in the late 80s I was the editor of Atari ST User and associate editor of Computing With The Amstrad CPC. In the early 1990s I was editor of PC Today and in the late 1990s I was editor of PC Home. From 1999 on I have been a freelance writer.
Contact me if you can offer any work! My email address is at the bottom of the page
About me
I got my first computer, a Sinclair ZX-81, in 1982. Back in those days, you pretty much had to be a programmer to do anything with your computer. You either typed in program listing from magazines (which took about two weeks and when you ran it, it said "Syntax Error", or the RAM pack wobbled and you lost the whole lot), or you created your own programs. Typing in listing was tedious, but a side effect was that you learned how computer programs were constructed, so after w while I wrote a few BASIC programs of my own and then learnt Z80 machine code. Compilers didn't exist in then. You had to write out the assembly language on paper, look up the op-codes in a big book, write them down in hexadecimal and then type in the hex on the computer.
After the ZX-81, I got a Jupiter Ace. Now this was an unusual computer because it didn't have BASIC as most other home computers did and it ran Forth instead, which is a completely different programming language. I spent a year or two writing my own programs in Forth because there wasn't any software for thr Jupiter Ace. I also wrote some Z80 machine code programs on it too. Strangely, writing programs in Forth on the Ace was fun, but then I've always been interested in things that are slightly out of the ordinary.
The next computer was an Acorn Electron (a cheaper version of the popular BBC Micro), and I wrote lots of programs in BBC BASIC and 6502 machine code. Then I got an Atari ST. I didn't write a lot for it, but dabbled in C, 6800 machine code and BASIC. At the same time, I also had an Amstrad CPC-464, another Z80-based home computer and wrote a lot for that too.
Finally, I got a PC. Programming a PC isn't as much fun as those early home computers and the processor was said to be designed by someone with a grudge against programmers. It's not at all friendly and despite all the previous machine code and Basic programming, I never wrote any programs for the PC for a long time. This is partly because in its early days, it was purely a business computer and couldn't do all the fun stuff that modern PCs are capable of. It wasn't until Microsoft Visual Basic 3 was released that I took up programming again.
Programming has always been a hobby because you had to be a programmer in the early days of home computers, but it's not the day job though, that's writing for magazines. Many are long gone, like Electron User, Atari ST User, Computing with the Amstrad CPC, and PC Today, but I currently write for several UK magazines on various topics.
I wasn't always a computer geek!
Check this out: An unexpectedly strong hydrogen bond: ab initio calculations and spectroscopic studies of amide-fluoride systems. You'll find the article about a third of the way down the page and I'm listed as one of the contributors. It is from around 1978-79 when I was at university. Click the First Page link to see a pdf of the first page of this riveting article.Actually it doesn't start till well over half way down the page, so you only get to read the first paragraph. Don't ask me what it's all about, that was a long time ago.

[attachment=1180]

[attachment=1181]
Reply
#29
Understanding Forecast Levels and Methods...

https://docs.oracle.com/cd/E16582_01/doc...EOAFM00177
Reply
#30
So much reading!

Anthony, do you read all this stuff you are posting links to?

Also where are we at with 5 forecast 5 predictions, abandoned?

Is this (these 2 links) home work for that?

PS you changed title RE: Forecasting & Analysis I thought I was in different thread.

I was hoping you would teach, so I wouldn't have to read a book on subject.
Have I over dominated discussion or your lecture? Sorry, if I did.

I was thinking for next step we need a third person who would make up race results on some sort of pattern.
Not tell us the pattern, but we try different methods of forecast to predict the pattern.
Others could join in with their predictions.

We need a dog.
B += x
Reply
#31
Yes! bplus I read everything I Post and more! Wink

I changed the Title because there are going to be many other Types of Forecasting & Results Examples on here,and you as one already showed an interest in other ways to do things which is fine.
Everything is still the same in General I will make the Program I spoke about where it will Predict not only (5) but a Range of Forecasts and Results,this make the whole thing more Flexible! then we can include things like Lotto's which I know Walter has an interest in,so he can give up Normal Work ? Smile and devote more time to this Forum,and all the other 1000's of things he as an interest in,when he just Wins enough to do that.
As I said the Program I have in mind will give the first R1 Results and then it will make a Prediction for R2 and save that to a File (Where all the Forecast Info will be Stored,for all Players) That way it Can't Cheat as you can then see using a Test run... the Forecast before a Player makes their Forecast,then a Player can Input their Forecasts then Press Enter or Reveal ? to see the R2 Results plus the Programs and players Forecasts,all saved in a File that can be Paste into a Spreadsheet,those are my first Ideas! Idea

Anthony.
Reply
#32
Cool!
B += x
Reply
#33
News! Update...

I have Got the Programming BUG! again and made BIG! Progress today with My...

"THE A.R.B FORECASTER TEST PROGRAM" 

Watch this Space...


Anthony.
Reply
#34
Hi Anthony,

I am watching this space with eager anticipation.

Why are you messing around with sort?

Can you outline the project you have in mind?

I would like to help build it, if you are stuck, but we need a plan.

Why are you still messing around with sort?  Smile
B += x
Reply
#35
You have to bear with Me bplus Wink

I am working on a number of things and most are related to this Project...
In the meantime Swat up on your Forecasting / Results methods...because you will need them when you play against "The System!"  Smile


Anthony.
Reply
#36
grrrrrrrr

Oh man! not The System!

Yikes! Wink
B += x
Reply
#37
"The System" Long... for the  "The SYS" Smile  Big Grin
Reply
#38
Do you know the mystery of the Sphinx? Smile
B += x
Reply
#39
No! but I have heard of the Mystery of the "Sphincter"  Big Grin
Reply
#40
Oh man, let's not go there! Wink
B += x
Reply