Welcome Guest
Log In | Register )
You last visited December 3, 2016, 6:31 am
All times shown are
Eastern Time (GMT-5:00)

Back testing systems

Topic closed. 11 replies. Last post 5 months ago by adobea78.

Page 1 of 1
52
PrintE-mailLink
amber123's avatar - OpIFNim

United States
Member #164727
March 12, 2015
2506 Posts
Offline
Posted: July 11, 2016, 7:05 pm - IP Logged

There are a few people on this site, won't mention names, who believe back testing systems are useless. What I'd like to know is how those same people who are opposed to it would rate a system. How would they know the efficacy of the system when they don't know how well it works in the past?

Although it's true that a good past performance is not a guarantee of how well the system will work in the future, I can't think of any other way to find out. Even the best systems have their dry periods, so if a system seems to be working recently, or...not working that well, only means it's in a trend/pattern deviation. 

So anyone who can tell me how they validate a system, NOT using back testing, please comment. 

    Avatar
    South Carolina
    United States
    Member #18322
    July 9, 2005
    1704 Posts
    Offline
    Posted: July 11, 2016, 10:01 pm - IP Logged

    There are a few people on this site, won't mention names, who believe back testing systems are useless. What I'd like to know is how those same people who are opposed to it would rate a system. How would they know the efficacy of the system when they don't know how well it works in the past?

    Although it's true that a good past performance is not a guarantee of how well the system will work in the future, I can't think of any other way to find out. Even the best systems have their dry periods, so if a system seems to be working recently, or...not working that well, only means it's in a trend/pattern deviation. 

    So anyone who can tell me how they validate a system, NOT using back testing, please comment. 

    Back testing is very necessary, b/c it's the only way to find out how well a system works in the state that you are playing in. In my opinion, It would be STUPID to waste money playing a FREE system in forward motion, if there's no evidence that it works well enough in your state through back testing.  I have eliminated many FREE systems posted here on LP through back testing, systems that allegedly worked in other states, but NOT in South Carolina. 

    If its a PAID system, and you don't have the opportunity to back test it before paying money $$$ for it, then that's a different situation.

    It's just COMMON SENSE to back test a FREE system, if you have the opportunity to do so, so you can save your lottery budget that you are playing the numbers with, if the system doesn't work well in your state.

      Avatar
      Madison, WI
      United States
      Member #172977
      February 11, 2016
      515 Posts
      Offline
      Posted: July 11, 2016, 10:40 pm - IP Logged

      Hi Amber and Destiny,

      My concern with simply backtesting, is that a system could be designed or created around satisfying those past results, in order to create an appearance that the system works. If I know the past 30 days results, I could make a system that happens to cover those 30 days very well. Or maybe a longer period if necessary.

       

      As far as a replacement for backtesting? You can simply forward test without playing real money. If you are in this for the longhaul, then it makes sense to take some time to use the system, make the predictions based on the system, and track how much would have been spent to play those predictions versus how much winnings would have came in, it won't be very hard to see how the system is working, and there is no danger that the system has been crafted just to cover the previous results.

      You also eliminate the risk of clouding the backtesting with ideas such as, "oh I wouldn't have played all of those predictions, just the ones I would have further worked out with my own ideas". You can actually use whatever other screening or systems you would use with your forward testing, and then know how the system and your additions to it are working in reality.

      In the end, if the system turns out to actually be working, you have plenty of time with the rest of your life to capitalize on your find and get rich.

        Tialuvslotto's avatar - Jailin
        Texas
        United States
        Member #150797
        December 31, 2013
        814 Posts
        Offline
        Posted: July 12, 2016, 6:26 am - IP Logged

        Why would you expect the game to be any different tomorrow than it is today?  The game is the game -- its the same everywhere.  Its not tougher in Ohio than it is in Calfornia.  It is exactly the same. Bang Head

        I agree with Wisconsin3054 that a 30 day period can display trends that may vary from the norm -- that is why 30 days is not a sufficient backtest.  However, comparing year-by-year, every state will have approximately the same statistics -- same proportion of doubles, same sum distribution, same proportion of even and odd, same for whatever parameter you choose to display.  This is especially true for RNG states as the system is designed to produce those distributions. 

        So there is absolutely no reason that a backtest of 3 months should not produce very similar results to a forward test of 3 months.

        "There is no such thing as luck; only adequate or inadequate preparation to cope with a statistical universe."

        ~Robert A. Heinlein

          str8ca$hhomie's avatar - Cash

          United States
          Member #146028
          August 22, 2013
          842 Posts
          Offline
          Posted: July 12, 2016, 9:22 am - IP Logged

          Back testing is very necessary, b/c it's the only way to find out how well a system works in the state that you are playing in. In my opinion, It would be STUPID to waste money playing a FREE system in forward motion, if there's no evidence that it works well enough in your state through back testing.  I have eliminated many FREE systems posted here on LP through back testing, systems that allegedly worked in other states, but NOT in South Carolina. 

          If its a PAID system, and you don't have the opportunity to back test it before paying money $$$ for it, then that's a different situation.

          It's just COMMON SENSE to back test a FREE system, if you have the opportunity to do so, so you can save your lottery budget that you are playing the numbers with, if the system doesn't work well in your state.

          From a slightly different viewpoint if I may, destinycreation. Why should back testing be placed upon the prospective "BUYER" of the FREE SYSTEM. What else in the real world works exactly like that ??? If only we had rules and/or regulations in place that prevented someone (I won't mention any name) from posting twenty-seven (27) different free systems and then relying on other people to tell him whether his free system works or not ! Shouldn't it be the other way around. The individual wishing to post the system does all the work in back testing if only showing number statistics for one (1) week. That way he shows the advantages in using his system and members could make an educated decision as to whether to utilize the system or not.

          Sometimes it's extremely difficult if not practically impossible to get people to disregard the smoke and mirrors.  Instead, they seem to enjoy the ride down the proverbial Garden Path....... helpless to extricate themselves from being totally deceived by known forces in their midst who would argue that they have come here for the sole purpose of helping people.......str8ca$hhomie

            Sunglasses's avatar - nicebear
            Zaperlopopotam
            Belgium
            Member #173932
            March 26, 2016
            950 Posts
            Offline
            Posted: July 12, 2016, 9:52 am - IP Logged

            There are a few people on this site, won't mention names, who believe back testing systems are useless. What I'd like to know is how those same people who are opposed to it would rate a system. How would they know the efficacy of the system when they don't know how well it works in the past?

            Although it's true that a good past performance is not a guarantee of how well the system will work in the future, I can't think of any other way to find out. Even the best systems have their dry periods, so if a system seems to be working recently, or...not working that well, only means it's in a trend/pattern deviation. 

            So anyone who can tell me how they validate a system, NOT using back testing, please comment. 

            Eventually your new system never worked in the past and just will work today and tomorrow the lottery shuts down. As you stopped for a pint in your local pub, you were too late for validating your ticket. You missed your chances, so you go back to the pub to have plenty of pints.


              United States
              Member #175910
              July 10, 2016
              40 Posts
              Offline
              Posted: July 12, 2016, 10:08 am - IP Logged

              Go ahead and test backward and or forward. The true test is with cold hard cash on the line! Shocked

                bobby623's avatar - abstract
                San Angelo, Texas
                United States
                Member #1097
                January 31, 2003
                1394 Posts
                Offline
                Posted: July 12, 2016, 11:04 am - IP Logged
                I'm not sure I understand what 'back testing' means in the real world.
                I have a workout and I know it works - now and then.
                Sure, I could go back and second guess myself.
                Oh, why did I choose 2 when 3 was the correct answer. Woe is me!!
                Face it.
                Lotteries are not predictable.
                We can always make educated guesses about what might happen next, but we won't know what
                the winning permutation is until it's known.
                The 'randomizers' the lotteries use for the games are working as intended, and no amount of back testing is going to change the future.
                Why people are so intent on trying to make randomizer outputs match the design, pre-mix statistics is a mystery, and a path to bankruptcy.
                Perhaps a majority of lottery gamblers don't really know how to analyze lottery history.
                It's always easy to blame someone else or some system for our losses.
                It's about gambling where there are a lot of losers and few winners - and few lucky souls who rarely lose.
                In my humble opinion, of course.
                  RJOh's avatar - chipmunk
                  mid-Ohio
                  United States
                  Member #9
                  March 24, 2001
                  19816 Posts
                  Offline
                  Posted: July 12, 2016, 11:27 am - IP Logged

                  There's nothing wrong with back testing systems you find hard to believe have worked in the past but most reasonable systems probably have worked in the past and if played long enough will probably work again in the future.  For example if some told me all six winning numbers in a 6/49 game are likely to be in a pool of 35 numbers or more 50% of the time I wouldn't back test it but if they said the same about a pool of 10 numbers I would.

                   * you don't need to buy more tickets, just buy a winning ticket * 
                     
                               Evil Looking       

                    Avatar
                    Madison, WI
                    United States
                    Member #172977
                    February 11, 2016
                    515 Posts
                    Offline
                    Posted: July 12, 2016, 12:28 pm - IP Logged

                    Why would you expect the game to be any different tomorrow than it is today?  The game is the game -- its the same everywhere.  Its not tougher in Ohio than it is in Calfornia.  It is exactly the same. Bang Head

                    I agree with Wisconsin3054 that a 30 day period can display trends that may vary from the norm -- that is why 30 days is not a sufficient backtest.  However, comparing year-by-year, every state will have approximately the same statistics -- same proportion of doubles, same sum distribution, same proportion of even and odd, same for whatever parameter you choose to display.  This is especially true for RNG states as the system is designed to produce those distributions. 

                    So there is absolutely no reason that a backtest of 3 months should not produce very similar results to a forward test of 3 months.

                    I don't expect the game to be any different looking back than looking forward. What I know is that the system could not have been designed just to fit the numbers coming in the future as those numbers haven't been picked yet. There is always the potential looking backwards that you could find a way that if you had just played it that way for the 30 days, 90 days, 1 year, whatever time period, I would have come out ahead. Its also too easy to say, oh it would have worked if i just had done the mirror, or the adjacent, or whatever else i want to make up to make it look better than it really was. Unless it actually accomplishes something going forward, its meaningless.

                    I guess I'd say to me backtesting is useful as a starting point, to give you some idea if there is anything to the idea. But then forward testing actually proves out whether it holds water. I don't claim to be a long-time expert, but I have yet to see a "system" that ends up holding water. Of course, as noted, I would expect that if someone had found it they wouldn't just share it in a public forum.

                      Avatar
                      Madison, WI
                      United States
                      Member #172977
                      February 11, 2016
                      515 Posts
                      Offline
                      Posted: July 12, 2016, 12:30 pm - IP Logged

                      Go ahead and test backward and or forward. The true test is with cold hard cash on the line! Shocked

                      Very true!

                        Avatar

                        United States
                        Member #116344
                        September 8, 2011
                        3919 Posts
                        Offline
                        Posted: July 13, 2016, 8:01 pm - IP Logged

                        Hi Amber and Destiny,

                        My concern with simply backtesting, is that a system could be designed or created around satisfying those past results, in order to create an appearance that the system works. If I know the past 30 days results, I could make a system that happens to cover those 30 days very well. Or maybe a longer period if necessary.

                         

                        As far as a replacement for backtesting? You can simply forward test without playing real money. If you are in this for the longhaul, then it makes sense to take some time to use the system, make the predictions based on the system, and track how much would have been spent to play those predictions versus how much winnings would have came in, it won't be very hard to see how the system is working, and there is no danger that the system has been crafted just to cover the previous results.

                        You also eliminate the risk of clouding the backtesting with ideas such as, "oh I wouldn't have played all of those predictions, just the ones I would have further worked out with my own ideas". You can actually use whatever other screening or systems you would use with your forward testing, and then know how the system and your additions to it are working in reality.

                        In the end, if the system turns out to actually be working, you have plenty of time with the rest of your life to capitalize on your find and get rich.

                        Thanks, very insightful !