Welcome Guest
You last visited December 7, 2016, 10:55 pm
All times shown are
Eastern Time (GMT-5:00)

# Nonlinear regression to predict lotteries

Topic closed. 119 replies. Last post 2 years ago by lottoswe.

 Page 8 of 8

United States
Member #59354
March 13, 2008
3972 Posts
Offline
 Posted: February 16, 2015, 2:09 pm - IP Logged

Ok BOB

Straight up, using regression to predict a random lottery drawing is a waste of time.  Selecting your

numbers randomly might even do better.   Study the data, make your best  guess, and forget regression

as it's not going to help.   Population densities in a lottery game are a product of random and while they

may continue there is no way to predict when.  I have ran thousands of simulations using regression and

tested the results against real random data, ie past drawings,  and could not find anything that made me

a believer.   At best it can be used to narrow the field but even then it is is not dependable.   Someone may

use some sort of regression and actually win something but regression was not responsible.

Straight up, no BS

RL

Working on my Ph.D.  "University of hard Knocks"

I will consider the opinion that my winnings are a product of chance if you are willing to consider

they are not.  Many great discoveries come while searching for something else

Trump / 2016 & 2020

United States
Member #59354
March 13, 2008
3972 Posts
Offline
 Posted: February 16, 2015, 2:23 pm - IP Logged

Take a look at the results below.  This data represents one rogue positions values 0 to 3

The asterisk's indicate the actual value that hit, the top line is the current data.  Notice that

the values rated highest and lowest both showed the same number of times, 5ea.   I don't

care how  you slice it one still has to guess.   If that translates to BS then so be it.

RL

These values are the results of millions of calculations.

Red = most probable

Blue = least probable.

(*) actual value that hit.

The overall average on the bottom line just includes the values that showed.

Working on my Ph.D.  "University of hard Knocks"

I will consider the opinion that my winnings are a product of chance if you are willing to consider

they are not.  Many great discoveries come while searching for something else

Trump / 2016 & 2020

bgonÃ§alves
Brasil
Member #92564
June 9, 2010
2122 Posts
Offline
 Posted: February 16, 2015, 2:57 pm - IP Logged

Rl, wanted his opinion, because in a lottery, 49/6 example Here we have 6 positions, or 6 last digit, to reach 100% is needed 100,000 (one hundred thousand in 100% = 000000-999999) But I got in ascending order 5005 ) clear is obvious in 100% a reduction of 95,000 ninety-five thousand, 100%

United States
Member #59354
March 13, 2008
3972 Posts
Offline
 Posted: February 17, 2015, 11:01 am - IP Logged

dr san

The problems we face trying to predict a random event are not limited to just regression.

We have to be very careful when we analyze the data so that we don't fall into a connect

the dot sort of analysis.  I think just about everyone has done this at some time or another.

How many times have you looked at the data after the drawing and seen all kinds of pointers

that somehow went unnoticed before the drawing.  The truth is that we will always be able to

do this regardless of which set of numbers come in.   Back-test often fall into this dilemma as

we are looking at events that have already happened.    Once the draw has come and gone

so has the randomness so to say.  It's always possible to work out a solution but don't expect

the solution to be correct.  The model may fit the history but the next draw is random.

I think gut-instinct is the best predictor we will ever have.    I analyze the data and then make

my best guess.   This works far better for me than any predictor program I have ever written.

If your guesses are not doing well then change the data you analyze.

RL

Working on my Ph.D.  "University of hard Knocks"

I will consider the opinion that my winnings are a product of chance if you are willing to consider

they are not.  Many great discoveries come while searching for something else

Trump / 2016 & 2020

bgonÃ§alves
Brasil
Member #92564
June 9, 2010
2122 Posts
Offline
 Posted: February 17, 2015, 2:23 pm - IP Logged

Hello RL perfect and yes because it is a only event (the next draw) but if it is to play 20,30 ... always the same game, we have more confidence, the problem is to find patterns
Rotary leaving 75% to 80% of the time, position by position, each position may have connection eat nearby position Example 49/6 The 1st position can be seen delays in 2nd place, 2nd place already seen delays eat 1st and 3rd position

United States
Member #59354
March 13, 2008
3972 Posts
Offline
 Posted: February 17, 2015, 11:18 pm - IP Logged

dr san

It's the patterns, trends etc..  not repeating when we need them to that makes prediction so hard.

Playing second digits is a fine reduction tool but many times requires additional filtering.   We can find

the expected for any value by looking at it's population within the entire matrix but again this only takes

us so far.  I have come to hate filtering which led to the creation of the rogue program.  The fewer the

choices we have to make, the fewer the mistakes.   The more I work on prediction the more I become

convinced that it's impossible.   I would like to think there is something that has been overlooked but

I think I will leave it to others to find.

RL

Working on my Ph.D.  "University of hard Knocks"

I will consider the opinion that my winnings are a product of chance if you are willing to consider

they are not.  Many great discoveries come while searching for something else

Trump / 2016 & 2020

bgonÃ§alves
Brasil
Member #92564
June 9, 2010
2122 Posts
Offline
 Posted: February 18, 2015, 6:32 am - IP Logged

Ok rl, putting the last digits in ascending order, 49/6
6 positions = the last digit (terminals 0-9) low of 100,000
(One hundred thousand) for 5005 at 100% clear !!

Tahiti- Polynesia
Tuvalu
Member #34524
March 4, 2006
54 Posts
Offline
 Posted: February 18, 2015, 5:03 pm - IP Logged

Nonlinear regression

17fev2015 - Euromillions

Data file processed with B-spline Order =3 Interval = 0.0125 (insert 80 intermediate datas between 2 numbers)

Predictor : Nearest neighbour
Distance : Euclidian

Dimension 3
Delay 2
Bandwidth 9

Data window used = 48000

17/02/2015 02 05 18 30 43 /1 10

Results at 80-160-240-320-400

Model Results
---------------------

Time Index Actual Value Predicted Value Error (NMSE)
48001 0.5440586     1
48002 -4.8893911    2
48003 6.3601678     3
--------------------------
48078 29.1009668 78
48079 29.5949999 79
48080 30.1223886 80----30
48081 30.6559556 81
48082 31.1638777 82
--------------------------
48155 43.0777338 155
48156 42.8645999 156
48157 42.6683116 157--43
48158 42.4350446 158
48159 42.1979336 159
48160 41.980545 160-------
48161 41.7508774 161
48162 41.5136002 162
--------------------------
48235 17.9162668 235
48236 17.5252889 236-18
48237 17.1408666 237
48238 16.705722 238
48239 16.2460999 239
48240 15.8173555 240--------
48241 15.4068667 241
48242 15.021411 242
------------------------
48317 1.8668622 317
48318 1.8733389 318
48319 1.8780378 319
48320 1.9465211 320-----2
48321 2.0723756 321
48322 2.1588112 322
--------------------------
48398 30.6559556 398
48399 31.1638777 399
48400 31.6756001 400---

Just an example.

That's a multi-step prediction, 400 datas ahead predicted with results at each 80 datas.

4 first numbers are good or near good. Fifth number is more difficult to find.

I generally use 2 numbers at each 80 datas. 3 good numbers on that drawing plus one star.

One-step prediction would be more accurate but the task is long and tedious.

I have a program beeing coded to automatize the process.

I have others examples like that.

I have won a lot with nonlinear regression. The only way to win.

Bob

New Member
Wheeling, West VA
United States
Member #115308
August 17, 2011
8 Posts
Offline
 Posted: February 18, 2015, 6:12 pm - IP Logged

dr san

It's the patterns, trends etc..  not repeating when we need them to that makes prediction so hard.

Playing second digits is a fine reduction tool but many times requires additional filtering.   We can find

the expected for any value by looking at it's population within the entire matrix but again this only takes

us so far.  I have come to hate filtering which led to the creation of the rogue program.  The fewer the

choices we have to make, the fewer the mistakes.   The more I work on prediction the more I become

convinced that it's impossible.   I would like to think there is something that has been overlooked but

I think I will leave it to others to find.

RL

When RL  says, "The more I work on predictions the more I become convinced that it's impossible," the more hope I feel for him.

Tahiti- Polynesia
Tuvalu
Member #34524
March 4, 2006
54 Posts
Offline
 Posted: February 18, 2015, 8:46 pm - IP Logged

Another example

23/01/2015 for 27jan2015

27/01/2015 0* 10 31 33 40 /8 10

Dimension: 4
Delay: 72
Bandwidth: 8

Predictor Type Kernel Regression
Kernel Type: Tricube
Distance Type: Euclidean

Interpolation : B-spline order3. Interval =0.0125 (80n inserted between 2 data)

Data Window used: 48000

xx 80 and multiple
** 79 and multiple

Results

48001 5.5625935 1
48002 5.2665372 2
48003 5.1143618 3
48004 4.9848709 4
48005 4.8905682 5
48006 4.7926008 6
48007 4.7137892 7
48008 4.6846332 8
48009 4.6061714 9
48010 4.5314446 10
48011 4.4389207 11
48012 4.3709692 12
48013 4.2485473 13
48014 4.1449217 14
48015 3.9648433 15
48016 3.8127105 16
48017 3.5413542 17
48018 3.2303695 18
48019 2.7320614 19
48020 2.1079218 20
48021 1.3993239 21
48022 0.8067466 22
48023 0.4121742 23
48024 0.1018203 24
48025 -0.1147591 25
48026 -0.1870762 26
48027 -0.2923274 27
48028 -0.3610562 28
48029 -0.6730756 29
48030 -1.0375281 30
48031 -1.3602798 31
48032 -1.4885052 32
48033 -1.6614458 33
48034 -1.916087 34
48035 -2.0031789 35
48036 -2.0622956 36
48037 -2.1212777 37
48038 -2.1853451 38
48039 -2.2605403 39
48040 -2.4019322 40
48041 -2.439503 41
48042 -2.4207806 42
48043 -2.3905713 43
48044 -2.320971 44
48045 -2.2720667 45
48046 -2.2799619 46
48047 -2.2389951 47
48048 -2.196441 48
48049 -2.1007782 49
48050 -2.0782619 50
48051 -2.1590237 51
48052 -2.3086557 52
48053 -2.4368237 53
48054 -2.5507459 54
48055 -2.6871059 55
48056 -2.9827468 56
48057 -3.6206471 57
48058 -4.0594217 58
48059 -4.2979458 59
48060 -4.3995872 60
48061 -4.3819931 61
48062 -4.3314843 62
48063 -4.5155307 63
48064 -5.0503119 64
48065 -5.4680654 65
48066 -5.3607272 66
48067 -5.0367847 67
48068 -4.4905593 68
48069 -4.349921 69
48070 -4.1643907 70
48071 -3.7054767 71
48072 -2.8082123 72
48073 -1.6497172 73
48074 -0.715752 74
48075 0.2052701 75
48076 1.3126842 76
48077 2.4709379 77
48078 3.6723353 78
48079 4.7280846 79**5
48080 5.3052369 80xx--5
48081 5.6337938 81
48082 5.9071448 82
48083 6.1157256 83
48084 6.1988502 84
48085 6.5005937 85
48086 6.6166741 86
48087 6.6261581 87
48088 6.8590747 88
48089 6.9280099 89
48090 7.0639738 90
48091 7.1313773 91
48092 7.1702691 92
48093 7.1041231 93
48094 7.1980466 94
48095 7.0701694 95
48096 7.0845043 96
48097 7.082953 97
48098 7.1560568 98
48099 7.2420476 99
48100 7.2157996 100
48101 7.3293115 101
48102 7.3574902 102
48103 7.4602416 103
48104 7.4137234 104
48105 7.4919556 105
48106 7.4643039 106
48107 7.4989112 107
48108 7.3910033 108
48109 7.2316695 109
48110 7.1201394 110
48111 6.795804 111
48112 6.384007 112
48113 6.2432572 113
48114 5.9795012 114
48115 5.6914885 115
48116 5.5820573 116
48117 5.4989392 117
48118 5.1474938 118
48119 4.9081839 119
48120 4.6383597 120
48121 4.5759564 121
48122 4.6534066 122
48123 4.9333472 123
48124 4.9780892 124
48125 5.0532771 125
48126 5.3270812 126
48127 5.6861793 127
48128 6.2639633 128
48129 6.7629827 129
48130 7.4655775 130
48131 8.5152165 131
48132 9.847192 132
48133 11.6842558 133
48134 13.8090873 134
48135 15.7214338 135
48136 17.3135597 136
48137 18.5382812 137
48138 20.0126197 138
48139 21.6615186 139
48140 23.3778375 140
48141 25.14117 141
48142 27.3145428 142
48143 29.3253846 143
48144 30.9560386 144
48145 32.663222 145
48146 34.0454409 146
48147 35.1932775 147
48148 35.6489983 148
48149 36.0247145 149
48150 36.2134839 150
48151 36.1538325 151
48152 35.8221141 152
48153 35.1647418 153
48154 34.3517533 154
48155 33.7605067 155
48156 33.4396555 156
48157 33.2814428 157
48158 33.2475094 158****33
48159 33.196325 159
48160 33.1597382 160xx--33
48161 33.1929356 161
48162 33.2124828 162
48163 33.2423185 163
48164 33.234216 164
48165 33.2803393 165
48166 33.3682384 166
48167 33.5502607 167
48168 33.8037103 168
48169 34.1195629 169
48170 34.5622492 170
48171 35.0405718 171
48172 35.6565703 172
48173 36.5002951 173
48174 37.5492333 174
48175 38.6517034 175
48176 39.7289237 176
48177 40.7275158 177
48178 41.5111115 178
48179 42.2677053 179
48180 42.8356277 180
48181 43.2860762 181
48182 43.6170336 182
48183 44.2769357 183
48184 45.1558507 184
48185 45.8518066 185
48186 46.4789863 186
48187 46.9242772 187
48188 47.1971927 188
48189 47.5348325 189
48190 47.6752588 190
48191 47.7779864 191
48192 48.0331448 192
48193 48.0326601 193
48194 47.9228481 194
48195 47.9015539 195
48196 47.6402778 196
48197 47.4960245 197
48198 47.1930762 198
48199 46.9266679 199
48200 46.6779108 200
48201 46.4567172 201
48202 46.3311891 202
48203 46.1511178 203
48204 46.0720889 204
48205 46.0197575 205
48206 45.9892517 206
48207 46.3659734 207
48208 46.6913243 208
48209 46.807274 209
48210 46.8824446 210
48211 46.9402539 211
48212 46.9928195 212
48213 46.8045305 213
48214 46.2920506 214
48215 45.5364924 215
48216 44.8521051 216
48217 44.2138277 217
48218 43.4850547 218
48219 42.5092327 219
48220 41.2639931 220
48221 40.0946336 221
48222 39.3562223 222
48223 38.9772031 223
48224 38.6876813 224
48225 38.632789 225
48226 38.6962575 226
48227 38.9106005 227
48228 39.2332185 228
48229 39.541386 229
48230 39.7686815 230
48231 39.9191472 231
48232 39.9979201 232
48233 39.9841766 233
48234 40.0130866 234
48235 40.0407666 235
48236 40.0167817 236
48237 39.9174674 237***40
48238 39.7331442 238
48239 39.5179936 239
48240 39.2683569 240xx
48241 38.9255645 241
48242 38.6634097 242
48243 38.3161221 243
48244 37.934152 244
48245 37.5207099 245
48246 37.066043 246
48247 36.5789672 247
48248 36.1497302 248
48249 35.8814137 249
48250 35.6613121 250
48251 35.487689 251
48252 35.2274446 252
48253 35.1589125 253
48254 35.0368194 254
48255 34.9019296 255
48256 34.7494736 256
48257 34.6694897 257
48258 34.6685671 258
48259 34.7078632 259
48260 34.7454281 260
48261 34.7862455 261
48262 34.8469551 262
48263 34.8550725 263
48264 35.0467666 264
48265 35.1824334 265
48266 35.2198852 266
48267 35.2232445 267
48268 35.2283591 268
48269 35.1883453 269
48270 35.0754514 270
48271 34.8204699 271
48272 34.7584658 272
48273 34.4485952 273
48274 34.1667006 274
48275 33.8742659 275
48276 33.5482028 276
48277 33.1220315 277
48278 32.4561424 278
48279 31.7606397 279
48280 31.0088231 280
48281 30.1835368 281
48282 29.404329 282
48283 28.7527179 283
48284 28.1492528 284
48285 27.6199925 285
48286 27.1653558 286
48287 26.6919416 287
48288 26.2608462 288
48289 25.7784077 289
48290 25.1422205 290
48291 24.6996915 291
48292 24.3288558 292
48293 23.923026 293
48294 23.5373527 294
48295 23.1575692 295
48296 22.3424844 296
48297 21.5095278 297
48298 20.5115928 298
48299 20.0842538 299
48300 19.7123771 300
48301 19.3830166 301
48302 18.8377742 302
48303 17.8708813 303
48304 16.5990112 304
48305 15.6093827 305
48306 14.819322 306
48307 14.2337023 307
48308 13.7501772 308
48309 13.360248 309
48310 12.9545381 310
48311 12.4362732 311
48312 11.7633806 312
48313 11.2179972 313
48314 10.6098778 314
48315 10.0241901 315--10
48316 9.2502023 316***
48317 8.2631235 317
48318 7.2775229 318
48319 6.7274139 319
48320 6.3963068 320xx
48321 6.3131454 321
48322 6.1216863 322
48323 5.9215632 323
48324 5.6826038 324
48325 5.6486918 325
48326 5.583607 326
48327 5.6431479 327
48328 5.8008728 328
48329 6.4447156 329
48330 7.4605208 330
48331 9.2171713 331
48332 11.4567384 332
48333 13.1476088 333
48334 14.3790842 334
48335 15.1972725 335
48336 15.7311749 336
48337 16.2007586 337
48338 16.5978946 338
48339 16.9507152 339
48340 17.2848114 340
48341 17.607984 341
48342 17.9885604 342
48343 18.2741121 343
48344 18.4837705 344
48345 18.7368582 345
48346 18.8005178 346
48347 18.8395906 347
48348 18.9648045 348
48349 19.1397058 349
48350 19.1506065 350
48351 18.9841244 351
48352 18.7436133 352
48353 18.4355379 353
48354 17.8835116 354
48355 17.4132185 355
48356 17.0376083 356
48357 16.8509362 357
48358 16.9428499 358
48359 17.2793154 359
48360 17.9220495 360
48361 18.7042057 361
48362 19.9927175 362
48363 21.0273803 363
48364 21.1366915 364
48365 21.5439039 365
48366 21.6738526 366
48367 21.5243228 367
48368 21.1628666 368
48369 20.9318913 369
48370 21.1130315 370
48371 21.5908094 371
48372 22.2996255 372
48373 22.916707 373
48374 23.4012991 374
48375 23.8869005 375
48376 24.490168 376
48377 24.7797857 377
48378 24.5230619 378
48379 24.0565748 379
48380 23.7903559 380
48381 23.6662444 381
48382 23.5975343 382
48383 23.6186158 383
48384 23.902261 384
48385 24.238633 385
48386 24.4764442 386
48387 24.8920387 387
48388 25.3461032 388
48389 26.0101427 389
48390 26.8490496 390
48391 27.6337616 391
48392 28.6546227 392
48393 29.6479815 393
48394 30.5331663 394-31
48395 31.6578309 395**
48396 32.5463436 396
48397 33.3735578 397
48398 34.0209808 398
48399 34.5229446 399
48400 35.0174452 400xx

Look at it and make your own idea.

Lottery is chaotic but is deterministic, not stochastic (completely random).

If you know how to handle it, you'll get good results.

Those who don't know will not succeed.

Bob

New York, NY
United States
Member #39471
May 16, 2006
2696 Posts
Offline
 Posted: February 20, 2015, 6:34 am - IP Logged

Surely, believers and fans of statistical formulae and methods should be hitting the P3 daily.  :)

\$\$\$

New Mexico
United States
Member #86099
January 29, 2010
11119 Posts
Offline
 Posted: February 20, 2015, 11:22 pm - IP Logged

Put 59 numbers on  sheet of paper spread 1/2 inch apart.  Put chicken feed over the paper.  Turn  a chicken loose and wait 8 minutes.  Take the paper and use the 5 biggest holes that the chicken pecked.  Random but a good chance to win!

Atlanta, GA
United States
Member #3480
January 24, 2004
51 Posts
Offline
 Posted: February 24, 2015, 1:11 am - IP Logged

Why would you not sort the draw if the results are non-positional?

Tom Dempsey

Tahiti- Polynesia
Tuvalu
Member #34524
March 4, 2006
54 Posts
Offline
 Posted: February 24, 2015, 5:36 am - IP Logged

Why would you not sort the draw if the results are non-positional?

Tom Dempsey

Hi,

I'm not sure your question is for me but this is my answer about using unsorted drawings.

Regression is looking at real phenomenons to predict future in many domains like weather, etc...

If you twist the phenomenon by sorting drawings, you loss his true behaviour. You can get some

results with sorted  drawings from time to time  but forecasting is regularly far better when taking

numbers in the order they come out of the machine.

Forecasting column by column is an interesting method but in that case, there will be not enough

data to use for good regression. That's about Euromillions I play.

Bob

Sweden
Member #163023
January 17, 2015
34 Posts
Offline
 Posted: February 24, 2015, 8:35 am - IP Logged

lakerben

For now I have to say no, the program is still under development but so far it's proving to be the best

I have ever written.  I have deleted all my older stuff from my computer.  I have a daily game version

that operates the same way and it will reduce a P-4 to 3 lines with 6 rogues.  Each rogue value has a

range of 0 to 3 and it outputs straight hits.  I am in the process of testing all my old predictions tools

and analysis stuff to see what I can find that helps pin down the rogue values.  P-3 can be reduced to

5 lines with 4 rogue values or 2 lines if 5 rogues are set.   The hard part right now is trying to decide

which game to work on, number games or daily games,  both are really tempting.   These programs

don't need extra filtering which was the motive behind them.  Anyway, thanks for your interest and

maybe at some point in the future I will release it.   I kind of overbuilt these programs as they have

a number of hidden options, wheels etc.. that give the user many ways to generate the lines and set

the rogue values..

Pic of P-4 setup using 6 values.

Hi RL,

i trying  to understand " rogue values".

You said each rogue value have range 0-3. Are they represent numbers 0-9 like range 0 is 1 and 2 , 1 is 3 and 4 etc?

I hope you can explain  this.

regards

 Page 8 of 8