How I play the Lotto

Winning in a Lotto Game is not an easy task.
Since some times I'm correct with my selections
this is the time that I shall win.
The plan I follow is the following:

If I select 9 numbers to play I use the LottoDesign (wheel)
LD(9,6,5,6.1)=7
The design below will give me at least 5 correct if I have the
6 correct numbers within my 9 selected once with 7 tickets.
1 2 3 4 5 6
1 2 3 7 8 9
1 4 5 7 8 9
1 4 6 7 8 9
2 3 5 6 7 8
2 3 5 6 8 9
2 3 5 6 7 9


If 10 numbers selected,
LD(10,6,5,6,1)=14
1 2 3 4 5 6
1 2 3 4 7 8
1 2 3 5 7 9
1 2 4 6 9 10
1 3 4 5 8 10
1 3 5 6 7 10
1 4 7 8 9 10
1 5 6 7 8 9
2 3 6 8 9 10
2 4 5 6 7 10
2 4 5 8 9 10
2 5 6 7 8 10
3 4 5 7 9 10
3 4 6 7 8 9

If 11 numbers selected.
LD(11,6,5,6,1)=22
1 2 3 4 5 6
1 2 3 4 7 8
1 2 3 7 9 10
1 2 4 8 9 11
1 2 5 7 10 11
1 2 6 8 9 10
1 3 4 6 10 11
1 3 5 8 9 11
1 3 6 7 8 11
1 4 5 6 7 9
1 4 5 7 8 10
1 5 6 9 10 11
2 3 4 5 9 10
2 3 5 8 10 11
2 3 6 7 9 11
2 4 5 6 8 11
2 4 6 7 10 11
2 5 6 7 8 9
3 4 5 7 9 11
3 4 6 8 9 10
3 5 6 7 8 10
4 7 8 9 10 11

If 12 numbers selected.
LD(12,6,5,6,1)=38
1 2 3 4 5 6
1 2 3 7 8 9
1 2 3 4 7 8
1 2 3 5 10 11
1 2 4 6 9 10
1 2 5 7 9 12
1 2 6 7 11 12
1 2 8 10 11 12
1 3 4 7 10 12
1 3 5 9 10 12
1 3 6 7 11 12
1 3 6 8 9 11
1 4 5 7 9 11
1 4 5 8 11 12
1 4 6 8 9 12
1 4 7 9 10 11
1 5 6 7 8 10
1 5 6 7 10 12
1 5 7 8 10 11
2 3 4 9 11 12
2 3 4 6 9 12
2 3 4 8 9 11
2 3 5 6 8 12
2 3 5 7 10 11
2 3 6 8 10 12
2 3 6 7 9 10
2 4 5 7 10 12
2 4 5 8 9 10
2 4 6 7 8 11
2 5 6 8 9 11
3 4 5 6 7 9
3 4 5 8 9 10
3 4 6 8 10 11
3 5 7 8 11 12
4 5 6 10 11 12
4 6 7 8 9 12
5 6 9 10 11 12
7 8 9 10 11 12


Do not waste you chance to win big when time is ripe!

I'll like to see what you follow and play to win.

Nick
 
I have been impressed by the excellent quality of your wheels for a long time. However I think the wheeling process is not the major hurdle to be overcome on the way to the jackpot.

From personal experience it is more critical to decide on what lotto numbers to include in the pool to be wheeled. After losing so often, I realize that in order to have a fair chance of including all of the 6 main winning numbers or at least 5 of the main numbers plus the bonus number, in the pool, one must play more than 20 numbers.

At this level, if only 5 of the winners are found in the pool, one may not win anything at all, if one plays only about 40 lines or blocks, even if one uses filters wisely, and especially if the lottery pays only for 4 winners and higher combinations.

Last week I segmented two pools of 16 numbers each. Both contain 4 winning numbers each. Yet I won nothing.

** Nevertheless the wheels you posted will produce fantastic returns on investment, if the particular draw, results in the winning numbers being densely clustered, in a cumulative frequency chart for example.

I have experimented with this approach at various times but the long term results are unprofitable. Somtimes one cluster may spawn another reduced cluster in a subsequent draw but I have yet to pinpoint what it is that triggers the clustering.

Perhaps one may use the Constraints setters in your Lotto Designer program to give higher weights to numbers in the region of the previous clusters, in the frequency chart. This is something I may work on but it will have to wait because your latest Build has refused to run on my PC.

Cheers!
 
Hi adrew3698,

You touch a subject of
spatial autocorrelation

A clustering pattern in the spatial distribution of some variable which seems to be due to the very fact that the occurrences are physically close together, that is, that they are in geographical proximity. They are not independent of each other, but somehow linked. In other words, the data are spatially dependent. Spatial autocorrelation is widespread: rich people move to areas where other rich people live; people only go to parties because other people go, and so on. If the values in the cluster are more alike than would be due to random processes, there exists a positive autocorrelation; if they are less alike than would occur through random processes, there exists a negative autocorrelation.

In illustration of this key point in the difficulty of using standard statistical techniques in geography, consider a random sample of a population, selected by using random number co-ordinates as applied to a map. The sample will not be random unless the individuals are also randomly distributed over the map area, yet a chief concern of, say, urban geographers is the way in which individuals cluster together.

Most standard statistical tests rest on the assumption that the observations made are independent of each other (i.e. are not autocorrelated), so techniques in spatial statistics have been, and are still being, developed in order to investigate spatial autocorrelations.


My recent LottoSelector (under construction) uses these areas to some extend but it is not an easy subject to investigate.

Can you please provide me with more details about the refusal of LD to run in your computer?

Nick
 
Thanks Nick for your ultra-quick response.

I believe there's a date-tag on the latest version of LD. My intention was to dabble with the consecutive-pair filter and the Range Sum filter. Upon downloading it, I seemed to get the message that the trial period has expired. The date stamp on my PC is correct, so it could be due the recent 'spring-cleaning' of the cluttered hard disk.

Anyway, further to my comments about the clustering of winning numbers, I notice numbers tend to link and delink in cycles. Of course the moot point is, how to identify and measure cycles consistently in a lottery. Sometimes I just resort to a simplistic approach. Take the total number of lotto numbers. Divide by two. Treat the result as a half cycle. Divide by two again to get a quarter cycle. This may even be wrong according to higher mathematics, like transforms.

But the pair that repeats within a 'quarter cycle' may repeat again before the end of the 'half cycle'. Similar patterns may emerge with 3-tuples. How does one exploit these types of cyclic behaviour?

** Perhaps you could incorporate a pairs and/or triples filter in LD to eliminate pairs that have surfaced thrice and triples that have appeared twice, within a 'half cycle', from the blocks.

Elimination of numbers certainly appears to be the optimal way to improve the chances of winning. But there seems to be a threshold to elimination where one has to be careful it's not overdone. Last week in a certain lottery, I decided to throw out three numbers that were downtrending, and from past patterns, were likely to remain so. Guess what, those numbers appeared with a vengence.
 

Geenie

Member
THis is one of my filter sets I used for the Super 7
Not too legiable because VBScript is not kind to me on this board but it
works for me in finding the proper
sets to play using min/max



RARITY 1 3 '"03183820421113304109121415012728"
FREQ 1 4 '"43161034373546060804230721323945"
CURRENT 1 4 '"02060814162024323707152125262731"
TREND 0 4 '"14201527091213184101033042112238"
PATTERN 1 3 '"02263233390141470515374344030609"
OPTIMUM 2 5 '"03060912131618192023243005153743"
INDEX-S 2 5 '"38451122423510342836251740460141"
INDEX-ALL 2 7 '"0109101112131722252830343536383940414243444546"
CURR-TREND 2 6 '"01020306070809111213141516182021222425262730313237384142"
PATT-OPT 2 5 '"010305060912131516181920232426303233373941434447"
PREV 0 3 '"04081421272931"
2ND 0 2 '"02142631323339"
3RD 0 3 '"06070813192024"
4TH 0 2 '"03080912161823"
EVEN 2 5 '"0204060810121416182022242628303234363840424446"
LAST 4 DRAWS 1 5 '"020304060708091213141618192021232426272931323339"
ONE TO 19 1 4 '"01020304050607080910111213141516171819"
20'S 0 2 '"20212223242526272829"
30'S 0 3 '"30313233343536373839"
40'S 1 2 '"4041424344454647"
1 - 9 1 3 '"010203040506070809"
20'S AND 30'S 2 5 '"2021222324252627282930313233343536373839"
LD-01 0 2 '"0111213141"
LD-02 0 2 '"0212223242"
LD-03 0 2 '"0313233343"
LD-04 0 2 '"0414243444"
LD-05 0 2 '"0515253545"
LD-06 0 2 '"0616263646"
LD-07 0 2 '"0717273747"
LD-08 0 1 '"08182838"
LD-09 0 2 '"09192939"
LD-0 0 2 '"10203040"
FAVOURITES 6 7 '"01020304060910111419202123242526272933343537394143444647"
2 5
Aggressive Favourites
'"0406111902370946"
 
Last edited:
Hi Nick, in an earlier post, I may have inadvertently given the impression that it's futile to use the small wheels you posted, to try and win the jackpot. Allow me to correct the picture somewhat.

There are great odds against the *individual* player using *one* 12-number wheel to win the jackpot.

For a syndicate the prospects may be brighter. How can it be done using LD for example?

Some people sneer at frequency charts but let's work with the 300-draw frequency charts. Sort the chart in descending order and trisect it equally or near equally. We may call the sections High (H), Medium (M) and Low (L) respectively. These may be combined into three smaller lotteries, namely HM, ML and HL.

Depending on the bahaviour of the parent lottery, one may find the 'core sixer' subsets surfacing in HM, ML or HL in above average frequencies.

Still getting the winning line out of a total pool of 28 - 34 numbers is a long shot.

Set LD to crank out a 'sufficient' bank of 12-number pools, using the 6 out of 6 option. Optimistically speaking, the 'core sixer' is resident in one of the pools.

Now wheel all the 12-number pools generated using the formula you posted. On a good day, the jackpot is hooked in together with heaps of other pari-mutuel and fixed prize payouts.

The catch is, the operator must have deep pockets to indulge in this kind of 'double wheeling'. Sigh, I remember the saying now. You must have money, before you can make more money.

Regards.
 
Last edited:

Geenie

Member
andrew3698 said:
Hi Geenie

I see your approach is strongly influenced by Lottery Director.

What's the success rate like?

I really just started using this method because
of Nick's LDxl but I do know by looking at these past LDIR predictions there are always a certain
amount from each. For example if one of your
sets you wanted to play had 5 or 6 numbers
in "pattern" that would be highly unlikely and
would save me a loonie or two!
 
Hello andrew,

-----------
Thanks Nick for your ultra-quick response.

I believe there's a date-tag on the latest version of LD. My intention was to dabble with the consecutive-pair filter and the Range Sum filter. Upon downloading it, I seemed to get the message that the trial period has expired. The date stamp on my PC is correct, so it could be due the recent 'spring-cleaning' of the cluttered hard disk.
=========
This week I'll post a new build which
I think that will be compatible with all PCs...



---------------
Anyway, further to my comments about the clustering of winning numbers, I notice numbers tend to link and delink in cycles. Of course the moot point is, how to identify and measure cycles consistently in a lottery. Sometimes I just resort to a simplistic approach. Take the total number of lotto numbers. Divide by two. Treat the result as a half cycle. Divide by two again to get a quarter cycle. This may even be wrong according to higher mathematics, like transforms.


But the pair that repeats within a 'quarter cycle' may repeat again before the end of the 'half cycle'. Similar patterns may emerge with 3-tuples. How does one exploit these types of cyclic behaviour?
============
I use samples of equal size.
I interimpose the binomial distribution
This way I always have a general idea
of the distribution's shape.


----------
** Perhaps you could incorporate a pairs and/or triples filter in LD to eliminate pairs that have surfaced thrice and triples that have appeared twice, within a 'half cycle', from the blocks.
=====
The upcoming version will include that.
Many users requested the same addition.



---------------
Elimination of numbers certainly appears to be the optimal way to improve the chances of winning. But there seems to be a threshold to elimination where one has to be careful it's not overdone. Last week in a certain lottery, I decided to throw out three numbers that were downtrending, and from past patterns, were likely to remain so. Guess what, those numbers appeared with a vengence.
==========
You are correct and not the only one that eliminated the upcoming numbers.
I'll agree that eliminating numbers is the best procedure.
I work on small steps and always keep some statistics as of the frequency of that Type I error happening.


Nick
 
Hi again,

-------
Hi Nick, in an earlier post, I may have inadvertently given the impression that it's futile to use the small wheels you posted, to try and win the jackpot. Allow me to correct the picture somewhat.

There are great odds against the *individual* player using *one* 12-number wheel to win the jackpot.

For a syndicate the prospects may be brighter. How can it be done using LD for example?
=====================================
I *always* use my 9 numbers 5if6 with
7 blocks wheel when I play.




------------
Some people sneer at frequency charts but let's work with the 300-draw frequency charts. Sort the chart in descending order and trisect it equally or near equally. We may call the sections High (H), Medium (M) and Low (L) respectively. These may be combined into three smaller lotteries, namely HM, ML and HL.

Depending on the bahaviour of the parent lottery, one may find the 'core sixer' subsets surfacing in HM, ML or HL in above average frequencies.
======================================
My approach in predicting is different.
One of my procedures is the Frequency Distribution of the last 5 draws.
This distribution has the property that includes ~1/2 of the 6/49 Game numbers.
Farther more, from the ones that have not appear (the complement) 3 will appear on the average the next time. From the ones that they have appeared once 2 will come and 1 from the ones that have appear more than one time.

This is the most concise distribution on any 6/49 Game.

Distributions of 300 draws+ for me they carry too much error to be used.


------------------
Still getting the winning line out of a total pool of 28 - 34 numbers is a long shot.

Set LD to crank out a 'sufficient' bank of 12-number pools, using the 6 out of 6 option. Optimistically speaking, the 'core sixer' is resident in one of the pools.

Now wheel all the 12-number pools generated using the formula you posted. On a good day, the jackpot is hooked in together with heaps of other pari-mutuel and fixed prize payouts.

The catch is, the operator must have deep pockets to indulge in this kind of 'double wheeling'. Sigh, I remember the saying now. You must have money, before you can make more money.
=====================================
We have to reduce the field first and then apply this procedure.
So the elimination of *any* number is very important.
 

Beaker

Member
<snip>
But the pair that repeats within a 'quarter cycle' may repeat again before the end of the 'half cycle'. Similar patterns may emerge with 3-tuples. How does one exploit these types of cyclic behaviour?

** Perhaps you could incorporate a pairs and/or triples filter in LD to eliminate pairs that have surfaced thrice and triples that have appeared twice, within a 'half cycle', from the blocks.
Elimination of numbers certainly appears to be the optimal way to improve the chances of winning. But there seems to be a threshold to elimination where one has to be careful it's not overdone. Last week in a certain lottery, I decided to throw out three numbers that were downtrending, and from past patterns, were likely to remain so. Guess what, those numbers appeared with a vengence. [/B]
Great discussion here andrew3698 :agree2: I believe Peter is doing some work on pairs repeating based on the last digit of the draw number - essentially a 9-10 draw cycle.

I am intrigued by the cycle strategy :agree: and have looked into your 3-tuple question. Based on what I see, for Canada's 6+B/49 lotto, 30 draws seems to be interesting for 3/6 numbers repeating.

Also, as Nick mentioned, I look at 5, 10, 21, 30, 300, 391 frequencies for various history subsets.

andrew3698, have you done any work with decade elimination? Seems to me this is a ripe area for elimination. In the Can 6/49 lotto we see at least 1 decade miss 7/10 draws.

Again great discussion here :agree2:
 
Hi Nick

============================================
My approach in predicting is different.
One of my procedures is the Frequency Distribution of the last 5 draws.
This distribution has the property that includes ~1/2 of the 6/49 Game numbers.
Farther more, from the ones that have not appear (the complement) 3 will appear on the average the next time. From the ones that they have appeared once 2 will come and 1 from the ones that have appear more than one time.

This is the most concise distribution on any 6/49 Game.

Distributions of 300 draws+ for me they carry too much error to be used.

=============================================

To be honest, I do not use the 300-draw frequency chart alone. However it may be a useful tool. Look for the frequency band that encompasses the median lotto number. You may find the majority of the winners are distrubuted either above or below the median group. Sometimes you may have to include the median group in the count though. Based on this you could turn the long-term frequency chart into a filter.

As for the 5-draw or slightly higher frequency charts, there is always a frequency band that divides the winners into majority and minority groups, assuming the lottery draws 7 winners each time. For the 5-draw chart, the freq=2 band is most likely to be the interface.

So you can set two Min-Max filters as follows:

2 5 Lotto numbers with freq = 2, 3, 4 or higher

2 5 Lotto numbers with freq= 1, 0

These are generally fail-safe ratios but they introduce too much latitude into the wheeling process, unless one plays with a large number of blocks.

Using these ratios one is almost always right in setting the filter but is unable to intercept winning combinations in a limited number of blocks.

Min-Max ratios of 3:3 or 3:4 may be better.
 
Last edited:
Hi Beaker

==============================================

andrew3698, have you done any work with decade elimination? Seems to me this is a ripe area for elimination. In the Can 6/49 lotto we see at least 1 decade miss 7/10 draws.

==============================================

For some time I did fiddled and fumbled with decades. The five decades may be paired to form 10 pools of 19 to 20 numbers each. Pick the right pool and you can have 3 to 6 winners, to be wheeled out together by some appropriate method.

Then I switched tactics. The simple hit chart formed by plotting the draw date against the numerically ordered list of lotto numbers caught my attention. There are interesting geometric patterns in them.

Now there are 'gurus' who caution against making anything out of these charts. It seems the dots are strung out or dispersed around each other by pure coincidence. Any inference you make is just hogwash, they advised.

The long-term charts, together with hindsight, seem to show another perspective. You see, there are pronounced gaps in them, formed by clustering of hits or skewing of hits towards low or high values.

One type of gap is formed by a continuous string of numbers 'hibernating' for 7 to 8 consecutive draws. Join the appropriate dots in the chart with a pencil, and you get a rectangle of 'non-action'.

Another type widens gradually from a single dot or a pair of adjacent dots for several draws.

The point is, those gaps eventually get filled.

Suddenly, a string of winners may appear to fill a large gap.

Or, the process is more measured. At first, one or two numbers, followed by a quiet phase, then a frenzy of activity for several consecutive draws.

You just have to be there when the action unfolds.

Players who dislike dealing with fuzzy situations, may have to opt for more clinical analyses like skip patterns, hit ratios and the like.
 
Size of Group of Numbers...

Hi Andrew & Beaker...

From my point of view
I use sizes of groups that have certain
properties, like sizes of 4,12,20 and 28 numbers.
These groups of numbers theoretically shall have a success rate of ~.5, 1.5, 2.5 and 3.5, (in a 6/49 Game), which is not possible.
Then my inference (any conclusion drawn from a set of propositions) will show more clear the trend of these groups.
 
Re: Size of Group of Numbers...

Hi Nick

{Quote}

From my point of view I use sizes of groups that have
certain properties, like sizes of 4,12,20 and 28 numbers.
These groups of numbers theoretically shall have a success rate of ~.5, 1.5, 2.5 and 3.5, (in a 6/49 Game), which is not possible.
Then my inference (any conclusion drawn from a set of propositions) will show more clear the trend of these groups.

___________________________________________________

I saw you post the same idea in r.g.l. sometime last February. I filed it with the intention of understanding it sometime later.

Perhaps now is the right time to ask.

I suppose "success rate" means the theoretical number of hits.

By what characteristics are the groups of 4,12, 20 and 28 set up?

Why is the sequence of success rate "not possible"?
 

hot4

Member
Re: Re: Size of Group of Numbers...

andrew3698 said:
___________________________________________________

I saw you post the same idea in r.g.l. sometime last February. I filed it with the intention of understanding it sometime later.

Perhaps now is the right time to ask.

I suppose "success rate" means the theoretical number of hits.

By what characteristics are the groups of 4,12, 20 and 28 set up?

Why is the sequence of success rate "not possible"?


I'm waiting too, since February. :bawl:
 
Let us use a 6/49 Game.
When we select groups with 4 numbers our expected probability is: 4*6/49 or 0.48980
But this score is not possible.
Such a group will have success of 0 numbers (lower than the expected 0.48980 )
or 1+ numbers (more that the expected)
Since "you are predicting" the later shall be more probable for your case.
The question is what have we gain by that single assumption?

Some calculations will show the following:
If one (1) number appears, from that group, we have reduced the total game
combinations to: 4c1 * 45c5= 4,887,036 a gain of 9,096,780 combinations!

If zero numbers appear then the total combinations are 45c6= 8,145,060
a gain of 5,838,756 combinations.


When we select groups with 12 numbers our expected probability is: 12*6/49 or 1.46939
But this score is not possible.
Such a group will have success of 0 or 1 numbers (lower than the expected 1.46939 )
or 2+ numbers (more that the expected)
Since "you are predicting" the later shall be more probable.
The question is what have we gain by that assumption?

Some calculations will show the following:
If two (2) numbers appears, from that group, we have reduced the total game
combinations to: 12c2 * 37c4= 4,358,970 a gain of 9,624,846 combinations!

If more that two numbers appear then the reduction is even bigger.


What we are doing exactly is "forcing our predictions".

Of course, other sizes of groups can be used if we are "certain" with our
assumptions, with even greater gain in combinations.

Such "hypotheses" are very critical.
Since we are dealing with (True/False) cases (binomial)
the remaining number's probabilities fluctuate based on the above assumptions.


I hope that this is clear as of what I try to convey.
 

hot4

Member
Re: to Nick ...

Nick Koutras said:
You choice of words shall be more careful.

If you do not understand just ask for clarification!

Nick

When I made the same questions to you in rgl, you didn't tell me, you have not understood the questions.

How could I know you didn't understand my questions?

Now I know the answer. Thanks.

Although you don't consider groups of 16 numbers so perfect, I like to study them because they have 16*6/49= 1,96 numbers expected (+-2).

So, if they hit 3 numbers, that means 50% better than the odds. The amount of possible lines (max=3 min=3 of 16) is 3 055 360. Don't you agree that this is a very good reduction?

Frank
 

Sidebar

Top