Lottery Prediction by Recurrence Analysis

Reply #170

Raven62 — Wed, 05 Oct 2005 15:07:53 GMT

Anyone getting good results with VRA

Reply #169

liuhuiguo — Sun, 02 Oct 2005 17:23:56 GMT

I used the number recurrence gap to get a several time serias and analyzed/forecast by using non-linear(Exponional smooth and ARIMA) method, from Aug. 2004, but it didn't work well

Reply #168

NemeSys — Thu, 29 Sep 2005 08:56:34 GMT

I'll get back to you on this (unless someone else does first), as I'm going away for the week.

Reply #167

superluz — Tue, 27 Sep 2005 06:19:42 GMT

Hi, Maybe somebody here can help me. I ve downloaded the Tisean zip file which contains a lot of .exe programs. I have a history file with n values and I would like to train the algorithm up to n-5 and predict the next 5 values to compare them with the original ones.I have tried the predict.exe and the zeroth.exe with some quite strange output results.Does anybody know how those programs really works or has an example or a batch file? Thanks in advance for any kind

Reply #166

NemeSys — Mon, 26 Sep 2005 18:43:11 GMT

Excellent! Well done, that man! I'm always on the look-out for new nonlinear wares, and these look very interesting (though I already have TISEAN). We'll see how they compare to VRA regarding predictions.

Reply #165

superluz — Mon, 26 Sep 2005 11:06:01 GMT

Hi,I ve discovered this on the web: http://kmop.mff.cuni.cz/Chaos/Programs.html Maybe it can help .

Reply #164

NemeSys — Thu, 22 Sep 2005 20:15:17 GMT

Well, after struggling with the EuroMillions results it appears that VRA predictions start to break down after six numbers. EuroMillions uses seven, and I can barely get two numbers, no matter how I juggle dimension and delay. The alternative is to split them into the five main numbers, and the two 'lucky stars'.

Reply #163

paurths — Wed, 14 Sep 2005 19:34:25 GMT

True, notepad is very slow at replacing. I just mentioned it b/c any windows-system has it.Or from excel, just save it as .txt, should also work.

Reply #162

vick — Wed, 14 Sep 2005 19:15:54 GMT

notepad takes forever to replace all those if you have alot, use Microsoft Word it does it MUCH faster, then copy it back to the notepad or whatever

Reply #161

paurths — Wed, 14 Sep 2005 19:06:25 GMT

if you save your excel file as .csv, then the numbers will be seperated by a ;(that is, if the numbers are each in a different column)Now open it with notepad and use the replace function. Replace all the ; by a space. The file is now ready to be used by VRA.

Reply #160

valelli — Wed, 14 Sep 2005 10:36:01 GMT

thanks maryland ive downloaded it, so now ive got to try it, seems a bit like the long way round, i hope vra will soon be able to read excel files that were i keep all my data. anyway thanks for the info, i will give it a go when i get back from work. is there anything i should know about goldenpen before i start dabbling.

Reply #159

Maryland — Tue, 13 Sep 2005 20:46:32 GMT

Your not doing anything wrong valelli. I can't use excel sheets either as (.csv) so I save it as a .csv then use goldenpen (a free program at LDIR website) open it with that and save it as a text file.

Reply #158

valelli — Tue, 13 Sep 2005 12:59:26 GMT

hi eugene, i am still having problems with vra shutting down, it looks like i cant use it with my excel sheets, ive tried changing them to csv am i still doing something wrong, and how long will it be before the new version is out? many thanks

Reply #157

NemeSys — Tue, 13 Sep 2005 10:10:53 GMT

For use with VRA your data should have the earliest result at the top and the latest at the bottom.

Reply #156

BigJer — Tue, 13 Sep 2005 09:50:31 GMT

Hi again the question was ................is it best to have the newest numbers at the top of column or at the bottom . BigJer

Reply #155

FrankieH — Mon, 12 Sep 2005 19:08:03 GMT

Thanks Nemesys, Vick and BigJer!Now i'm back on course! Thanks a million guys! FrankieH

Reply #154

vick — Mon, 12 Sep 2005 16:30:39 GMT

find the 1st empty column near your other column[s] and enter from top to bottom 1 to whatever the last row number is, best way to do this is just enter the 1 in the first row and 2 in the second row and then click on it and highlight both rows and roll it down to the bottom which excel should autofill with descending numbers. Then select all the colums including the new column of descending numbers and then click on Data - Sort - Sort by the new column - then select the descending and no header... [ More ]

Reply #153

NemeSys — Mon, 12 Sep 2005 10:11:22 GMT

All UPDOWN and CoolRevGui do is to reverse the order of a file. They aren't prediction programs.

Reply #152

BigJer — Mon, 12 Sep 2005 09:46:43 GMT

hi guys I like the ion updown for DOS but I would like to know what order the program works best with , newest at the top or newest at bottom. I have down loaded the software but have not tried it yet. BigJer

Reply #151

NemeSys — Mon, 12 Sep 2005 09:08:58 GMT

Hello FrankieH,I found this neat little DOS utility called UPDOWN.exe at:http://www.saliu.com/free-lotto-tools.htmlScroll all the way down the page, and it's just above the yellow rectangle. It takes a plain ASCII file and reverses it, even if the data is arranged in columns. Include the file extension for both the input and output file, e.g. in.txt out.txtIf you prefer a Windows program try CoolRevGui.exe (about half-way down the page just above MARKOV.exe) at:http://www.saliu.com/other-softwa... [ More ]

Reply #150

FrankieH — Mon, 12 Sep 2005 04:56:59 GMT

Can someone suggest software that will flip the order of my numbers vertically? My Excel tools aren't working. Help!!Thanks, any suggestion would be appreciated.FrankieH

Reply #149

LANTERN — Mon, 12 Sep 2005 00:04:27 GMT

It is very nice to have an alternative to old LottoSync, now we have a new plaything .It is better than Lottosync, because VRA can be used for any lottery games.

Reply #148

vick — Sun, 11 Sep 2005 04:51:07 GMT

i didn't play keno tonight because was playing the 649's, but using yesterday's numbers from my best parameters set, ie the same one that would have hit for $25,000 a few draws ago it would have hit 11 / 20 which is MUCH BETTER than random average of 6 / 20. A little wheel and some luck could have hit for some more money [note that in quebec if you hit 15 / 20 that's $1 MILLION win but we don't have that here in ontario but if you were to hit 15/20 your wheels would surely hit for large]. This w... [ More ]

Reply #147

valelli — Sat, 10 Sep 2005 12:37:50 GMT

nemesys how about putting a predictions page on as well for forthcoming draws with the del and dim used we can then compare them to our own results.

Reply #146

NemeSys — Sat, 10 Sep 2005 12:11:55 GMT

I'll add a sub-page to my site detailing the mathematics I'm using, then people can play with the equations and perhaps make their own discoveries. I'll add an analysis of EuroMillions too (the nights are drawing in and trawling through combinations of dimension and delay is a good way to kill time).

Reply #145

valelli — Sat, 10 Sep 2005 10:53:59 GMT

thanks eugene, ive checked and my spreadsheets are all in one column, they were working until recently and then stopped, but i was still testing it, ive got to do some amendments to my computer, and then i will try it again saving them as csv files and see if that works, and would like to say i love the programme, although still a newbie with it, how long have you been producing programmes like these.

Reply #144

NemeSys — Sat, 10 Sep 2005 08:54:49 GMT

The shareware version of VRA allows you to save graphs but not files, the full version does. I don't mind using the graph, even though it's a bit harder on the eyes, although there is a zoom function: hold down the left mouse button and drag the cursor from top left to bottom right. To zoom out again drag the other way (bottom right to top left). Perhaps Eugene will automate some of the procedures, now that there is a greater demand for VRA.

Reply #143

LANTERN — Sat, 10 Sep 2005 07:34:02 GMT

Instead of the chart is there a way of getting numerical predicted values?And yes, tasks such as those and others should be preprogrammed and automatic.To be run at night while sleeping for example. And numbers instead of just graphs would be good also.Perhaps somebody will code a suitable programable keyboard macro that can be used with VRA to automate certain needed tasks.

Reply #142

NemeSys — Sat, 10 Sep 2005 07:16:31 GMT

For the benefit of retxx (and anyone else interested) here's a screen shot of an actual prediction (for the EuroMillions draw):To explain: the numbers up to 574 are those drawn so far. The 573rd and 574th numbers are 3 and 6 (the 'lucky star' balls). Beyond that you have the prediction (575 to 581). This prediction was for the 9th of September. It gave 8, 12, 23, 40 and 50, with lucky stars 8 and 9. The actual result was 8, 12, 19, 31, 50, and 6, 7. So the prediction gave three out of the five m... [ More ]

Reply #141

LANTERN — Sat, 10 Sep 2005 06:32:22 GMT

Eugene.I don't know how to say this, but maybe what paurths and some others need is a sort of macro that will automate certain needed VRA tasks maybe a sort of pre-processed script to be run by VRA a sort of way to automate certain needed VRA tasks, like maybe a linear series of predictions (or repetitive predictions) with progressive and automatic varied steps of dimention and or delay and or other variables or whatever is needed.In other words, the software should run or process pre-programmed... [ More ]

Reply #140

vick — Sat, 10 Sep 2005 03:32:43 GMT

didn't too good tonight again in the keno. mabye that first time was a fluke

Reply #139

retxx — Sat, 10 Sep 2005 02:32:34 GMT

nemesys can you actually post a screen showing probable numbers and where they would appear on the chart. thanks

Reply #138

vick — Sat, 10 Sep 2005 01:29:55 GMT

Vick: didn't do good tonight but yesterday's was a firecracker.:result nearest neighbour 5 / 20 17-4 - % of mean size max = 500 3 7 10 13 17 20 24 27 30 34 37 41 44 48 50 54 57 61 64 67results here:http://lottery.sympatico.msn.ca/cgi-bin/english?job=show_results lottery=on_daily_kenoI know very little about the lottery, so can you translate the term firecracker into the money terms? That is, how much would you win if you used VRA to play that draw

Reply #137

paurths — Fri, 09 Sep 2005 23:46:29 GMT

nonlinear: I feel your pain, and for the large part, it's there because VRA is really not meant for lottery prediction, although I am thankful to Nemesys and others for popularizing it here. I'll consider how to make it easier for you, the lotto guys.... just throw in some pathintegralsn perhaps a little monte-carlo, let it calculate for hours hours, predicting most likely futures on knowledge of parameters it has used to create possible pasts that actually got real close to reality and we... [ More ]

Reply #136

nonlinear — Fri, 09 Sep 2005 23:09:36 GMT

Carbob: I have a couple of questions. If I input my draws, 1-500, new to old and I predict from 501-510. Are the numbers I get are they for draws 501 thru 510, or are they for the next draw? My input is descending.The data in the input file should be arranged from the oldest to the newest results. That is, record 1 for the oldest draw, record 500 for the newest draw. It's also important that the last record really reflects the last known draw. That's because to make a prediction, VRA uses the cu... [ More ]

Reply #135

CARBOB — Fri, 09 Sep 2005 22:03:20 GMT

I have a couple of questions. If I input my draws, 1-500, new to old and I predict from 501-510. Are the numbers I get are they for draws 501 thru 510, or are they for the next draw? My input is descending.Carbob

Reply #134

FrankieH — Fri, 09 Sep 2005 19:31:22 GMT

Thanks again Nemesys. You've been very helpful.

Reply #133

valelli — Fri, 09 Sep 2005 18:29:52 GMT

nemesys can you help, ive got several excel worksheets that ive been using with vra, and for some unknown reason it has stopped working, the vra programme starts up ok, but when i go to open a signal (ie one of my sheets) it closes down the vra programme, its only just started doing this, is there something wrong somewhere, and before anyone shouts virus, ive excluded that already, any thoughts

Reply #132

NemeSys — Fri, 09 Sep 2005 18:07:09 GMT

If the predictor you're using (under Model Options) is Nearest Neighbour then the number of neighbours defaults to one (the nearest possible neighbour), and no other value is permitted. Switching from neighbours to fixed size gives poor results, so I don't use the fixed size option.

Reply #131

FrankieH — Fri, 09 Sep 2005 15:48:04 GMT

Nemesys,Do you have a suggestion for the settings of Neighbors and Fixed Size ? My settings are set for Neighbors (1) and Fixed Size (25).Thank you.FrankieH

Reply #130

NemeSys — Fri, 09 Sep 2005 11:34:47 GMT

Paurths: it looks like you've come across the periodicity in predictions that the phase space plot indicated (the chess board pattern, alternating periods of hits and misses). I'm looking forward to seeing the program you're writing.FrankieH: for 6/49 I'd suggest starting with delay of 6, and a range of dimension values, say 10 to 20, then working out from there. Something I've noticed is that good dimension values tend to be small multiples of delay, such as 2, 3, 4 times the delay.... [ More ]

Reply #129

LANTERN — Fri, 09 Sep 2005 05:34:54 GMT

It seems to me that the software needs to be optimized for lottery predictions and then maybe it would be the new Holy Grail of prediction software.You (we) should all send suggestions to this effect to Eugene, it seems to me that he is beginning to get interested in using his software on the lottery.Good luck.

Reply #128

vick — Fri, 09 Sep 2005 03:51:35 GMT

didn't do good tonight but yesterday's was a firecracker.: Sept 8 keno vra result nearest neighbour 5 / 20 17-4 - % of mean size max = 50037101317202427303437414448505457616467 6 / 20 18-4 - % of mean size max = 50037101317202427313437414447505357616467 6 / 19 18-4 - #of neighbours17212722252831323326292849474463526866 6 / 20 12-5 - % of mean size max = 500371014172224273034... [ More ]

Reply #127

retxx — Fri, 09 Sep 2005 01:34:43 GMT

vick could you post the keno numbers that vra came up with. I would like to see how they fare here in Mass. where keno runs every 4 minutes.

Reply #126

FrankieH — Thu, 08 Sep 2005 23:13:07 GMT

Hi Everyone, Has anyone come up with a workable Dim and Delay for 6/49? Any help would be appreciated. ThanksFrankieH

Reply #125

vick — Thu, 08 Sep 2005 19:42:16 GMT

I've tried this program for the 1st time with my Ontario Keno 20 / 70 and it coulda hit for $25,000 CND or so yesterday the best setting was Nearest Neighbour / Euclidean / Dim = 17 / Delay = 4 / Neighbourhood = % of mean size 500 [every thing else at default settings] at the current time i have 3419 draws in the 20 excell files 1 for each ball.so i set train from =1 / train to 3419Predict from 3420 / predict to 3420and this gives me 1 exact number on the left hand side for each execl ball file... [ More ]

Reply #124

paurths — Thu, 08 Sep 2005 19:39:58 GMT

Hi Retxx,i've tested it for just about any game you can find.Pick 3 did not score pretty good here.The best results i got was when i sorted the combo's from low to high and exported them from my database to 3 textfiles, so each textfile would have one column.For column 1 and 3 i had some pretty good results, but for column 2, which fluctuated the most, the results were not that magnificent. But then again, not every pick3 game is the same. (And i can not address this enough, sometimes 2 or more... [ More ]

Reply #123

paurths — Thu, 08 Sep 2005 19:34:26 GMT

A list of numbers at the bottom could be a step forward, b/c sometimes its pretty vage when you try to determine the exact x-y.I ve been testing it quite to the bottom , until my fingers hurt, due to the fact that i ran prediction, typed in the numbers in Access and then on the drawday compared the results (which i automated ofcourse). I've tested it for 5 consecutive draws, each time producing 14 numbers. (in most case VRA produced the same number several times in one prediction --> which was... [ More ]

Reply #122

retxx — Thu, 08 Sep 2005 16:02:06 GMT

I agree with lantern can it actually show a list of playable numbers at the bottom.Also can you enter as many past pick 4 draws as wanted? Does it work with a pick 3 daily as well? thanks

Reply #121

NemeSys — Thu, 08 Sep 2005 14:55:57 GMT

That screen capture Eugene posted was merely to demonstrate a point about the correlation between the training data (the actual results) and a given predictive model. To make actual predictions the red line should extend beyond the blue one.Further site additions: comparison of VRA graphs of UK Lotto, Thunderball, and EuroMillions results:http://zarnia.250free.com/visual.htm

Reply #120

LANTERN — Thu, 08 Sep 2005 14:34:41 GMT

I have never used it yet, but it sounds to me as if it might be better than Guru's LottoSync program and it is very much better for sure than the Attrasoft Predictor.If Eugenes's chart could be translated into the actual numbers then it would be super great.

Reply #119

NemeSys — Thu, 08 Sep 2005 07:47:10 GMT

Of course! I should have realised that! I use post-dicting to build tables of dimension and delay. From these I select those values which give two or more of the actual numbers. It's still laborious, though.Your screen capture shows a Kernel regression predictor, a Tricube kernel, and Euclidean distance. I use the Nearest Neighbour predictor with Euclidean measure (the blue and red curves match exactly then, but that doesn't guarantee highly accurate predictions). As you know, the Nearest Neig... [ More ]

Reply #118

nonlinear — Thu, 08 Sep 2005 02:59:05 GMT

What I actually meant was that the analysis of the 200 predictions can be made using the past data. Since VRA pretends not to see the data in the test interval (if it is set to be after the training interval), the fair test of the prediction quality can be evaluated using the actual draws that have occured in the past. This is sometimes is refered as an out-of-sample prediction and is considered a standard method of evaluating the prediction models. That saves you 2 years of waiting before ma... [ More ]

Reply #117

nonlinear — Thu, 08 Sep 2005 02:52:23 GMT

Another thing that didn't immediately occur to me is that for the 6-ball loto, each prediction should be a set of 6 numbers. The only way to do this in VRA is to use the multi-step model type, which is using the prediction for time t to make a prediction for time t+1. This type of forecast is by nature inferior to a one-step type, where the actual result at time t is used to make a prediction for time t+1. I hope I am explaining it well. The analogy is forecasting the wheather: it's obvious... [ More ]

Reply #116

NemeSys — Thu, 08 Sep 2005 02:26:18 GMT

You had partly anticipated my reply. Also, the results are presented as a continuous string of numbers, when in reality each group of numbers is separated by two or three days. As some people have noted perhaps a fairer test would be to analyse the results in draw order rather than by ascending order. However, the predictions I'm getting don't really justify re-ordering the data (though I could always be wrong). The whole point of ascending numerical order is that it 'helps' VRA in making predic... [ More ]

Reply #115

nonlinear — Thu, 08 Sep 2005 01:29:15 GMT

I decided to give it try, fully expecting to see an unambiguous confirmation that there is nothing to predict when it comes to random numbers. The data for this experiment was the UK six main balls dataset that I downloaded from NemeSys' website. Below are the prediction results for the 100 balls, along with the model parameters. As you can see, the correlation between the predicted and the actual results is enormous. That visual result is also confirmed by a very low value of NMSE.I am very... [ More ]

Reply #114

NemeSys — Wed, 07 Sep 2005 23:16:14 GMT

Glad to be of help, FrankieH and Bolade (and thank you to everyone else for making this a well-viewed thread).Provided you have a complete set of results you can analyse any game with VRA.I have begun looking into the EuroMillions game, and although the set of results is small it already looks promising. The best values for dimension seem to be 21 or 22, and 7 for delay. In fact, I've noticed the optimal value of delay appears to correspond to the number of balls per draw. For EuroMillions there... [ More ]

Reply #113

mybolade — Wed, 07 Sep 2005 20:49:35 GMT

FrankieH, I am interested in the software if it words for pic 4. Have you had any positive results? I do not play the lottery, only pic 3 and pic 4. I have followed the forum from the beginning and appreciate NemeSys and others for your help. However, you have been talking about the Lotto.I am glad to see someone is working with pic 4. Bolade

Reply #112

FrankieH — Wed, 07 Sep 2005 12:57:20 GMT

Hi NemeSys, Thank you very much for your advice, i did it and it worked. I think i might have a handle on comprehending the software too (albeit, not in a analytical fashion!) Thanks again for your help.FrankieH

Reply #111

NemeSys — Wed, 07 Sep 2005 07:44:06 GMT

Delete all non-result values from your file (such as the date), otherwise the program will treat them as part of the data. Also, remove any non-numerical entries (i.e. alphabetic characters and symbols).

Reply #110

retxx — Wed, 07 Sep 2005 02:42:11 GMT

another question; do I enter just the 4 digit numbers or do I need the date for any reason?thanks

Reply #109

NemeSys — Tue, 06 Sep 2005 21:27:05 GMT

VRA will take any ASCII file of numerical data. Click on 'File', then 'Open signal'. Select 'all files' from 'Files of type' then click on 'Open', then you're ready to start experimenting.

Reply #108

retxx — Tue, 06 Sep 2005 15:22:55 GMT

can I load all the pick daily 4 Mass. numbers since 1979 into vra and if so how is it done.Thanks also will it analyze for play.?Thanks

Reply #107

NemeSys — Tue, 06 Sep 2005 06:46:11 GMT

Hello and welcome FrankieH.Have you loaded a data file in first? Without that VRA won't do anything. Go to my web site and you'll see some screen shots of VRA taking you through the initial steps. Let me know if you're still stuck:http://zarnia.250free.com/For those already familiar with my site, it seems I made a mistake interpreting the chessboard pattern. The curve that generates it isn't a sinusoid but a square wave (which can be neatly expressed as a Fourier series: I'll put it up on the si... [ More ]

Reply #106

FrankieH — Tue, 06 Sep 2005 00:03:47 GMT

Hello Everyone,I'm new to L.P. and i'm happy to be here. I've just downloaded the Visual Recurrence Analysis and after following instructions, i found that it's taking a long time to generate numbers. I'm still waiting. Maybe it's not working. Is that normal? Any advice will be welcomed. Help a girl out. Thanks in advance.FrankieH

Reply #105

NemeSys — Sun, 04 Sep 2005 19:32:05 GMT

The author of VRA himself! We're honoured. I'm still looking for related software to augment the analysis I have done so far. This list looks promising:http://usti.net/home/search/od/dir.cgi?d=/Science/Math/Statistics/Software/Time_Series_Analysis

Reply #104

Hyperdimension — Sun, 04 Sep 2005 04:21:02 GMT

Hi Eugene,Your program VRA is excellent to detect patterns, I was trying to predict a fractal Sierpinski Triangle pattern using 0 and 1 in VRA and it predicted correctly, the data:11 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 01 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 01 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 01 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 01 1 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 01 0 1 0 1 0 1 0 0 0 0 0 0 0 0 0 0

Reply #103

nonlinear — Sun, 04 Sep 2005 03:47:29 GMT

Wow, you guys take lottery seriously! I thought the jury was out -- the lottery results are random! If not, wouldn't there be a few smart people who figured that thing out to become billionaires and thus proving it otherwise? Well, anyway, here are a few points that I'd like to make:1. Visual Recurrence Analysis (VRA) assumes the data points are arranged in time. The most recent points should be at the end of the data file.2. If the lottery numbers are generated by a mechanical device, it better... [ More ]

Reply #102

NemeSys — Sat, 27 Aug 2005 21:42:22 GMT

New sub-page added to my site:http://zarnia.250free.com/rng.htm

Reply #101

NemeSys — Thu, 25 Aug 2005 00:32:20 GMT

The formulae are readily available, but most of the really useful ones have already been written into software, namely VRA, TISEAN, AVTS, Cint, Fractan, among others. Enter 'recurrence' 'analysis' 'software' into AltaVista will find most of the best software that's been written so far (I don't like Google, btw - how it got so popular is a mystery to me...)I'll take a look at sourceforge, too.

Reply #100

paurths — Wed, 24 Aug 2005 23:29:09 GMT

Well, i considered the same thing, installing linux!Today i've spend quite some time on www.sourceforge.net where everything is opensource (and much only for linux), and so far i have not found anything that pins it down, but i guess there might be some things there. If i find something thats interesting i'll surely post it here.(Hopefully it ain't linuxplatform only...)Provide me with the formula's, Nemesys, and i'll write the damn program

Reply #99

NemeSys — Wed, 24 Aug 2005 23:24:33 GMT

The time-consuming aspect I can live with. It's consistent results that need pinning down. Automating the search would speed it up, but you would still have lots of results to go through.Somewhere there might be recurrence analysis software which does cycle through values automatically and implement other features; I'm always looking for new programs that implement RA, but so far there doesn't seem to be much (a few very interesting programs I've seen only run on other platforms - I've been temp... [ More ]

Reply #98

paurths — Wed, 24 Aug 2005 20:14:53 GMT

Well, i ve been testing VRA for several weeks now and it pretty much got me to the nuthouse.Dim Del work for about 2 draws, then it shifts, the dims dels become pretty much useless, and i'm in pursuit of the new best dimensions and delays . 2 draws later it shifts again.Also it is very time consuming. One way or the other, it would be just magnificent if the calculations (for ALL dimension-delay combinations) could be automated. Then the results could simply be written to a DB-structure whe... [ More ]

Reply #97

NemeSys — Wed, 24 Aug 2005 01:30:23 GMT

Changing the dimension and delay alters how the recurrence model looks at the data. The plot is a two dimensional view of what's going on in a higher dimensional phase space, which of course we can't see if it's higher than three anyway. The 2-d plot essentially tells you how closely related the vectors are. As dim and del change so do the numerical values connected with the colours (the lengths of the vectors are changing).There's no way to change the axial lengths - whatever appears on the x-a... [ More ]

Reply #96

Hosaka — Tue, 23 Aug 2005 23:24:18 GMT

Hi NemeSys, Thanks for responding. Yes, I am using VRA.1) The spectrum boundaries do correspond to the limits of the actual dataset entered when using values embedded dimension=1 and time delay=1, but what then do they represent when those are changed? The upper bound seems to shift drastically with changing dimension and delay. I understand it is the Euclidean distance away from the actual value, but how can that be interpreted?3 4) I don't understand why the fact that lottery results require... [ More ]

Reply #95

NemeSys — Tue, 23 Aug 2005 22:20:53 GMT

The coloured plots are there only to give a broad indication of the time series structure (or indeed if there is structure at all). The graph is really what you need to look at for actual predictions. But in answer to your queries (I'll assume you're using VRA):1) The colour spectrum seen at top right corresponds to actual numbers. For a 6/49 lottery the white end corresponds to 0, the black end to 48 (it goes from 0 to 48 rather than 1 to 49 because that's just the way VRA was programmed, but i... [ More ]

Reply #94

Hosaka — Tue, 23 Aug 2005 19:50:52 GMT

NemeSys:This work is intriguing. I have gone over your webpage, but never came away with a clear sense of how to read a recurrence plot. Which is better - the smaller magnitude, red end of the spectrum or the larger blue? When a checkerboard structure appears, why is the number of squares always an integer number within the time series? Are there ever parameters which only show partial structure at the largest - or smallest - values in our time series, thus displaying a partial square? Is the ch... [ More ]

Reply #93

NemeSys — Tue, 23 Aug 2005 01:27:54 GMT

Site update: placed three links at the top of the main page, added four more PDFs to the research papers list, added a link to the University of Southampton lottery work, and a new sub-page about visualizing lottery numbers:http://zarnia.250free.com/visual.htmNow I can back to analysing actual results.

Reply #92

NemeSys — Mon, 22 Aug 2005 21:18:46 GMT

Hello Zulu, I tried the values you mentioned with the Lyapunov colouring scheme. I think the Lyapunov colouring shows up details that the others tend to swamp. Phase space portraits are often filled with more structure than the eye can see at first glance. Fractals coloured with the sepia tones of Lyapunov rather than by a more typical 'rainbow' spectrum often look clearer and more striking.

Reply #91

Zulu — Mon, 22 Aug 2005 18:31:55 GMT

I meant to add Volcano and Lyapunov colour schemes give a pretty regular colour of the parameters I gave.

Reply #90

NemeSys — Mon, 22 Aug 2005 18:27:18 GMT

You'd be surprised just who does lottery research! Here's some work from the University of Southampton:http://www.hpcc.ecs.soton.ac.uk/lottery_f.html

Reply #89

Zulu — Mon, 22 Aug 2005 18:09:10 GMT

Try this to test RP. Generate a 6/10 lottery of say a thousand combinations. Apply RP to predict. It struggles even with this simplified lottery(+- 220 possible combinations). Test column by column and then on an all six combination basis. I am pretty sure the Maths wizards at Oxford, Cambridge or wherever have had a go with their Cray supercomputers. Or perhaps they don't research lotteries as it is too vulgar for those chaps and they could be kicked out of academe. As a matter of interest Dime... [ More ]

Reply #88

NemeSys — Mon, 22 Aug 2005 16:15:31 GMT

Infinitysymbol: you're right about dimension. Of the two parameters it's always the harder one to pin down (there's tons of theory about finding the best value for a given data set). Regarding 'post-draw'predicting, it's the only way to do it. By building up a table of values for dim and del you can then begin to see (hopefully) which ones to use for real predictions. As for patterns, yes, the human mind is especially good at finding them where they don't necessarily exist. But VRA is just a set... [ More ]

Reply #87

Zulu — Sun, 21 Aug 2005 18:04:58 GMT

The usefulness of RP is it tells you what numbers will not come up column by column. If you divide the 6/49 lottery into patterns then it tells you which patterns will not be forthcoming in the next draw or the next few draws. In other words it is very good at anti-predicting. For example randomly delete 10 numbers. Then use RP to anti-predict 10 numbers. I tend to delete 2 of the winning numbers while RP anti-predicts more successfully than me. Nemesys has to be congratulated in presenting this... [ More ]

Reply #86

infinitysymbol — Sun, 21 Aug 2005 15:41:12 GMT

Hi Nemesys, I have several queries. Firstly, with regards to the results that you have received with Thunderball, although the best delay to use seems to be mostly 4 or 5, with regards to dimension, the numbers you use range from between 10 to 20 with no one number appearing more dominant than any other. Therefore, this can be interpreted either as there being a pattern, more specific with delay but less so with dimension , likewise one can see your results as specifically selected for ma... [ More ]

Reply #85

NemeSys — Sat, 20 Aug 2005 15:37:30 GMT

I'm having problems with my ntl connection at the moment (I'm sending this from the local library). I don't know when they'll sort it out, only a few days I hope.

Reply #84

NemeSys — Wed, 17 Aug 2005 15:17:49 GMT

The arrangements look different on a word processor, but VRA treats them the same, namely as a continuous string of numbers. However, if you want to analyse each column separately then they do have to be loaded as individual files. You can extract single columns of numbers from the Excel file on my site (which I still haven't had time to update recently, I've a job interview Thursday, so early to bed tonight...).As to which values of dimension and delay are the best...I honestly don't know for s... [ More ]

Reply #83

valelli — Wed, 17 Aug 2005 07:35:07 GMT

hi nemeys, this is a very basic question, ive been playing around with vra, what would you say was the best approach to input the data, one single column with all 5 numbers in one cell, or all the numbers in 1st postion in one column, and to check them a column at a time, and which del dim is recommended, im not having much luck, but i think thats down to my ignorance.....so ie whats the best format for inputing 1 15 24 26 342 16 25 27 33 as one column or12 as one column....many thanks

Reply #82

NemeSys — Tue, 16 Aug 2005 22:15:40 GMT

You don't need any extra figures for AVTS, just the data file. The UK main lottery results give a shape like a tetrahedron. This is only part of the attractor since it's only one trajectory, but I'm guessing whole thing has the same basic shape, like the Sierpinski Tetrahedron, some beautiful models of which you can see here:http://www.public.asu.edu/~starlite/sierpinskitetrahedron.html

Reply #81

Hyperdimension — Tue, 16 Aug 2005 21:55:18 GMT

Hi nemesys, I downloaded fractan and avts from your site, I have next question, in the avts program in order to see the fractal you need all data of loto and avts calculate fractal dimension or only knowing the fractal dimension 2-H +/- D is enough to see it? Regards

Reply #80

NemeSys — Tue, 16 Aug 2005 21:47:48 GMT

To calculate the fractal dimension I use Fractan, available at:http://www.tsm-soft.com/category/education/edupro03.htmlAfter unzipping and running, load in an ASCII file of results, then click on the 'Process' button to the right. Next, click on the Process menu, then 'Hurst exponent'. Save the file, and the values of Hurst exponent and fractal dimension will appear at the bottom of the window.To see the attractor I use AVTS, available at:http://www.cg.tuwien.ac.at/research/vis/dynsys/avts/This... [ More ]

Reply #79

Hyperdimension — Mon, 15 Aug 2005 23:21:05 GMT

Welcome back Nemesys,Do you know how to see a fractal dimension, I mean how to create the fractal (the image)? Which parameters?H=2-1.3090=0.691 (Random H=0.5, if H>0.5 less chaos)HDelta=0.1204HAlpha=? Ex.1.3090+/-0.1204Regards

Reply #78

NemeSys — Mon, 15 Aug 2005 22:39:07 GMT

Hello all, I'm back. At home I dug around through my library of software and maths/physics books and brought back a couple of things I forgot I even had, like the random number generator based on linear congruence, also a little program I wrote to display lottery numbers in interesting patterns (I wrote it years ago on a BBC Master in BASIC, but it can easily be translated to other languages). I'll put it all up on my site, I just have to catch up with a few things in the non-math world first.... [ More ]

Reply #77

NemeSys — Sun, 07 Aug 2005 21:43:00 GMT

Unfortunately there's no way to automate the search process from within the program. A batch file might work, but my programming skills aren't up to it. For the time being you just have to go through it manually.

Reply #76

rowdy117 — Sun, 07 Aug 2005 20:46:07 GMT

Nemesys,Thank you for all of your input on VRA. Everyone appreciates your effort.I have worked with this program a bit. Granted it is free, but I find it very difficult and time consuming to work with.It would be nice if you could loop the dimension and delay instead of using a slider bar. It takes a lot of time to manuallygrind out the results.

Reply #75

NemeSys — Sun, 07 Aug 2005 13:48:31 GMT

I'm going away for a week on Monday, so I thought I'd post this before I go.I was looking at the VRA analysis of the Thunderball results. The current values of dim and del accurately predicted two out of the five main numbers quite regularly. It occurred to me that if the predictions are good for two numbers, perhaps that's how I should look at the results. To explain: for the main 6/49 draw, analyse columns 1 and 2, then 3 and 4, then 5 and 6. It might provide that oh-so-elusive breakthrough. T... [ More ]

Reply #74

NemeSys — Wed, 03 Aug 2005 23:20:50 GMT

Spatio-temporal entropy is just a measure of noise (randomness) in a system. A more useful measure is correlation dimension (click on 'analysis' then 'General Nonlinear Analysis'). CD calculates the most likely optimal value of dimension for a given time delay.As for draw order files, you can find one in the zipped Excel file at: http://www.jcwchan.clara.co.uk/lottery1.htmGo down the page about halfway then click on 'Click here to download'. The password is 'extra'. After you've unzipped it and... [ More ]

Reply #73

paurths — Wed, 03 Aug 2005 22:28:52 GMT

I haven't read enough about recurrence analysis, but i guess the squares that show up when adjusting the dimension and/orf delay, somehow hints towards the prediction parameters. The spatio-temporal entropy looks promising, or so i thought....I figured, the more drawings, the more accurate the prediction will be (low entropy), but instead the entropy gets sometimes amazingly low when only using the last 3 months or so. But, like i said before, there's alot of reading learning left.Also with... [ More ]

Reply #72

NemeSys — Wed, 03 Aug 2005 19:54:11 GMT

The periodicity of predictions was something I had expected - a run of hits, followed by a longer dry period (you can see it in the phase space portraits).I didn't think RA would work on bingo (that's what we call keno in the UK). One behaviour of VRA predictions I have noticed is that as you gradually increment the values of dimension and delay, predicted values bounce around a certain number (often the one that actually gets drawn).Draw order is something I mean to get around to (unless some... [ More ]

Reply #71

paurths — Wed, 03 Aug 2005 00:32:54 GMT

Hi, i play Belgian lotto, 642 with powerball (or whatever it is named) Have spent the last few nights and days testing VRA, damn it takes incredible time to test fully and i'm dead beat, but finally came up with something, or nothing.Something that costs a lot of money... When asked the VRA-soft to return 19 numbers (instead of 6), i hit the 7 numbers serveral times (so far 7 out of 16), for 3 periods over the entire period of 22 years (only since 1983, b/c eversince then there are 42 balls ins... [ More ]