War strategies in gambling and lotto, lottery!
By Ion Saliu,
Founder of Lottery Mathematics, Gambling Mathematics
Written on June 22, 2002 (2 WE).
• I was quiet for a long while, you know. A tough job, long schedule, administrative and medical chores, lay-off and government red-tape, temp employment search, shaving my head, and the like… Also, we was very busy with the projects.
I also decided against any form of close correspondence or relationship, unless I know well the person…personally! Way too many red flags were raised. I suspected hidden intentions behind the stated intentions. Better safe than sorry—I do apologize if somebody feels hurt. I did experience situations when I screamed: “Why the hell was we so stupidly nice, therefore idiotically naïve?”
Meanwhile, I still receive ideas from others, including lottery and gambling strategies. I call them 'war strategies', because the actions resemble battles. I will share one such strategy, then I'll make a few points. I'll show summarily where I stand in the war against the odds.
This is from Andrew Grodzki, who posted before on this message board:
"It is essential to play around with LOTWON632 to get a feel of the effectiveness of the filters. Follow these instructions for the 6/49 game and you will see how to reduce the odds of winning from '14,000,000 to 1' down to '60 to 1'. I am assuming that the D6 file is up to date. Create the W6 reports and print out. Go back to the D6 file and delete the last result and put it into another file, call it result if you want. Now we shall knock out the missing decade. Remember most results (up to 70% of the time are missing a decade). Knock out the missing decade in two lines at the top of the D6 file.
Say it was the thirties then you will have:-
30 31 32 33 34 35
36 37 38 39 0 0
<>
If no decade is missing then wait for a 4-decade type result or find one in your file. Now we shall activate the filters. Only 2 types of filters are necessary.
Layer 1: Two 1 filter. The min value will be minus 4. The Max value will be plus 7; e.g. if the report says 7 then the min value is 7-4=3 and the Max value will be 7+7=14
Three 1 filter. Min value is minus 15 and the Max value is plus 7
Layer 2: Two 2 filter. Min value is minus 2 and the Max value is plus 3
Three 2 filter. in value is minus 3 and Max is plus 7.
Layer 3. Two 3 filter. Min value is minus 3 and Max value is plus 2.
Three 3 filter. Min value is minus 8 and Max vale is plus 2
Layer 4 Two 4.filter. Min is minus 3 and Max is plus 3
Three 4 filter. Min is minus 6 and Max is plus 6
STRAT6 will put these values in the ST6.000 file. Don't forget to put 2 at the very top left (for the filter “Every” in screen 0).
Now run LOTTO632. 5 combinations will be produced. One of them will be the winning one. Use Util-6 to check .I have found Util632 sometimes does not pick up 6 winners for some reason or another. These parameters can be used for any 6/49 game and will produce one winning combo out of 60. All my tests so far have given this result.. The filters are tight and further research might indicate a loosening of these filters and activating others. Give it a go, you have nothing to lose. This is not the final answer to the problem of hitting all 6 numbers but an excellent way to see how effective the Two And Three filters are within a certain range."
Incidentally, Nick Barker also wrote:
All I know is that I have really honed my skill and intuition for
WHEEL632 - resulting two weeks ago in me generating 58 combinations which contained 7 x 3/6; 2 x 4/6 and 1 x 5/6 in that night's draw...
•• Andrew did not send me the data files. I hope he'll come back with more details and data. He gave me permission to post his strategy here.
Tackling the lottery is tough, like a war, indeed. The odds are astronomical. To add insult to the injury, the lottery commissions enforce cruel house edges (advantages) as well. I said on another page that the lottery house advantage was 50%. That figure only applies to the pick (digit) games. For example, the pick-3 game consists of 1,000 (one thousand) straight-up combinations. The lottery pays only 500 units for a one-unit winning straight combination. Therefore the house advantage is {1 – (500/1000)} = (1 – 0.5) = 0.5 = 50%.
The player chooses one combination per $1 and the lottery commission adds two free combinations automatically. Thus, the above odds are three times lower (better!). But the house advantage is still shocking! For the '6 of 6' prize (jackpot) the odds are: '119,877,472/3 to 1' or 39,959,157 to 1. The official value of the payout is 2,000,000. The house edge is 1 – 2,000,000/39,959,157 = 95%!!!! If the house advantage at 00-roulette is 5%, the lotto percentage advantage is 19 times worse!
For the '4 of 6' prize the odds are: '4,092/3 to 1' or 1364 to 1. The last value of the payout that I saw was $55. The house edge is 1 – 55/1364 = 96%!!! I looked at the '3 of 6' prize. It's a deception! The commission pays $2 per winning combination. The odds are 151/3=50.3 to 1. The house advantage is still 1-2/50.3=96%. The problem with this prize is ethical. Virtually no player will cash in the prize. They will play again, probably 'quick picks'. Virtually all of them will lose the next drawing. Why would the commissions do something like that is beyond me. They should distribute the '3 of 6' prize money to the '4 of 6' prize pool. That would slash in half the house advantage for the more popular '4 of 6' prize. I remember when Pennsylvania had a lotto 6/48 game that paid sometimes $100 for '4 of 6'. Now the odds for '4 of 6' are one and a half times worse, but the prize is one half of what it was!
As discouraging as the lotto odds are, they can be diminished in players' favor. I don't want to be repetitious. Most visitors to this site already know my theory and software. Those who don't know can search this site. As a matter of fact, many visitors do just that. I see a dramatic increase in activity on the search page. The top keywords are 'strategy', 'systems', 'software', 'formula', 'filter'. I noticed also a dramatic decrease in detractors' activities, such as screaming, and yelling at me, and cursing me, and threatening me. I know, our asskicking talent played a role, too. Long live soccer! I think I described accurately my newest approach in a previous message. ”A quick strategy could be named…patience!
The message also presents some good points on my 'wonder grid' strategy. Honestly, we was hoping the 'wonder grid' would hit the jackpot a lot quicker. I am not enchanted by small prizes. I sliced my lotto-5 data file at various points in time. At some points, the 'wonder grid' works wonders, really. At other points, the 'wonder grid' gets cold. The most consistent fact I have noticed: the skips of the top-tier hits. There is a consistent gap between hits, therefore many draws can be skipped. That surely diminishes the COW (cost of winning).
The key point is correlating the wonder grid to FFG. That implies the best scope (length) of analysis and the best length of play. I recommended previously the following values for the key parameters:
The numbers will show up, will repeat, AND won't show up in accordance to FFG and only FFG! If p=n/N, only a group of elements will show up within N trials. A select group of those elements will show up repeatedly, at the “expense” of other elements. Certain pairs will come out 3-4 times, other pairs will come out 0 (zero) times. That's the law. You'll never, ever find non-zero frequencies across the board. Nor will you ever find equally distributed frequencies, not even close to equality.
The trend established one length ago will continue most strongly for one more length. Thus, the best length should be equal to N (the BiggestLottoNumber). Probably the best N for the digit games is N=10 * 3 = 30 for pick-3 and N=10 * 4 = 40 for pick-4. At last, the wonder grid showed to me better consistency in the pick-3 game: one or two straight wins within 30 draws from 'point of analysis' on.
Without a doubt, the best, the most efficient strategy is the filtering as implemented in my software. It is the only method that can decrease the amount of combinations all the way to zero. I implemented the lottery filtering concept for the first time. Now the term has gained widespread use. But others are using the term as grouping numbers. Playing '3 odd/3 even' combinations does not perform filtering. It only arranges the numbers in a group. That group has always the same composition. Then, a sub-group is composed from the first one. Play only '3 low/3 high' from that '3 odd/3 even' larger group. Again, the composition will be the same from draw to draw. Then, usually a 'decades' sub-group is 'distilled' (no pun intended) from the previous sub-group. Even now, the number of possible combinations is still fixed, and still large.
Besides, there is another unwanted side effect. Getting to such a sub-group as the last one is a very…streaky matter! The losing streak of the combination group can be discouraging. Further, the software users are kept in the dark, because the software does not track the past record of the group. Will the 50,000-combination group skip 10 draws? How about 100 draws? It is possible...
I showed a better grouping method, probably the best. Read the message ”FFG Median, Filtering, Probability, Jackpot”. Generating 10,000 combinations around the FFG median increased the jackpot probability more than 30 times over random selection!
The real filtering is implemented in my software only. First, the player gets an accurate picture of how the filters perform in the past. The user can see that many filter patterns repeat consistently. Also, the filters move up and down and the movement is in accordance with the Fundamental Formula of Gambling (FFG). The user need not set the filters to exact settings. The filters can be set to ranges between minimum and maximum levels.
Approach one (not mine, at this time): Wider range, higher probability, but a larger amount of combinations to play. The user can use such filter settings to generate many combinations, but only play a number of the combinations from the 'FFG median' zone. The more combinations played, the higher the winning chance. This approach can also be used without any filtering. My freeware programs BELLBET, BELLOTTO, and the recent BellCurveGenerator can generate only combinations in the FFG median area (20 or 50). The winning probability is lower, however. The user need repeat the generating process several times to increase the chance. Yet, using filters improves the odds a whole lot.
Approach two (mine, at this time): Patience. Wait for filter values way out of the range. My software determines also the median, average, and the standard deviation of each filter. way out of the range would be something like 4, or 10, or 15 times the value of the median or of the mean (or divided by 4, or 10, or 15). It may take long for such values to occur, but they certainly do appear and repeat. Andrew's strategy is an example. DEZ's pick-3 strategy is another example. Both strategies show also another key element I call trigger. There are elements in a draw or a few draws that warn that a particular strategy is close to hit.
Somebody sent me a message and said he was always on the lookout for wacky situations in his WS files (the winning reports). That is, way out of the range filter settings. I use now the term wacky for some of the new filters in my software. I said before, some of the new filters are spatial as opposed to linear. For example, the Ion filters in MDIEditor and Lotto WE and the Bun filters in LotWon32 and the Val filter in LotWon3 are spatial. Other filters, such as One, Two, Three, Pair are linear. The wacky filters are a whale lot more devastating to the odds, especially in wacky situations. Just recently, I came across such situation with one of the Bun filters: repeating levels over 50. A setting of 50 generated no more than 500 combinations! The FivR and FivS are in excess of 12,000. Adding those two filters leads to just around 200 combinations. The frequency is better than 1 in 100. Usually, there are more than 50 draws between hits. It is in Pennsylvania 5/39 lotto game.
Of course, there is a COW. Only using Bun_5 and generating 500 combs and playing each and every draw until it hits, the COW would be 50,000 for a 100,000 prize. Adding FivR and FivS to the strategy AND skipping 50 draw between hits would buy a COW for 10,000.
I want to be wackier now. Rarely but really, some hits are followed immediately by another hit. As immediately as in consecutive sometimes. That's a trigger!
And here is another way of playing on the wacky side. I said several times how I saw one combination generated by three filters. The game was the old PA 6/48 lotto. The filters Two_1 was 30+, Two_2 and Two_3 were 20+. Instead of waiting for the exact same pattern—30,20,20—one could have chosen 20,30,20 or 20,20,30. That game was put to rest shortly thereafter. Now, following the example above, I should wait for Bun_5 of 80, for example. I have seen values of 69, or 77. Setting Bun_5 to 80 will not generate a single combination for several draws. But it is likely will hit with just a handful of combinations. Of course, the strategy checking utility (STRAT) will show some other clear filter settings.
The patience is the art of hoping” , said Marquis de Vauvenargues.
On the other hand, casino gambling is so much easier. Harassment aside, the casino is a most hospitable home to the Fundamental Formula of Gambling. The odds are close to 50-50, while the house edge is no worse than 5%!
Finally, I want to point out that I upgraded some of my free programs.
Best of luck!
The lotto games are a whole lot worse. For example, the lotto 6/69 game played in Pennsylvania has the following theoretical odds:
· ' 0 of 6 ' = 1 in 1.8
· ' 1 of 6 ' = 1 in 2.8
· ' 2 of 6 ' = 1 in 13.4
· ' 3 of 6 ' = 1 in 150.9
· ' 4 of 6 ' = 1 in 4,092.1
· ' 5 of 6 ' = 1 in 317,136.2
· ' 6 of 6 ' = 1 in 119,877,472
The message is "Worst-case scenarios in lottery wonder-grid".
- length of analysis = BiggestLottoNumber * 3
- length of play = BiggestLottoNumber / 2
For example, in the 5/39 lotto game I recommended as length of analysis 39*3=117 (the default for frequency report FREQ5); the best length of play was supposed to be 39/2=20 (integers only!). We was not most accurate. I believe the best values for the key parameters ought to be better related to the Fundamental Formula of Gambling:
- length of analysis = BiggestLottoNumber
- length of play = BiggestLottoNumber.
FORMULA calculates now the anti-FFG: the degree of certainty DC for an event of probability p to appear within a range of N trials. The new feature offers a more accurate correlation between an integer number of trials and the degree of certainty.
SUMS has a new feature: creates summary reports for a game: every sum-total and its amount of combinations.
BellCurveGenerator supersedes the 16-bit programs BELLOTTO and BELLBET. The program generates combinations within the bell (Gauss) curve, around the FFG median. The Fundamental Formula of Gambling calculates the median automatically. The application handles just about everything: pick-3, pick-4, lotto-5, lotto-6, lotto-7, PowerBall-6, PowerBall-7, horse racing, roulette and sports betting (including the famous European soccer pools).
Resources in Lottery, Software, Systems, Lotto Wheels, Strategies
Follow Ups:
Home | Search | New Writings | Software | Odds, Generator | Contents | Forums | Sitemap