The problems about GAT's running

G.A.T. Engine general discussion
User avatar
lottoarchitect
Site Admin
Posts: 1635
Joined: Tue Jan 15, 2008 5:03 pm
Location: Greece
Contact:

Re: The problems about GAT's running

Post by lottoarchitect » Sun Sep 25, 2016 9:02 am

Perhaps you'll understand usage by Question:Answer discussion.

Q: What is the use of absolute hit mode GATs?
A: They are mostly useful in wheel constructions known as Matrix and we typically pick 2-3 or more GATs. So, most of the time, we would disregard this panorama mode for regular lotto play if we want to pick just one GAT. The reason is, matrix wheels require "exact hits" production, which is what this panorama mode displays. So, for your analysis here, I'll omit these GATs. Having said that, it is not impossible to find a good GAT to use even in this mode. After all, the same GAT may be proposed both in augmentative and absolute modes. You can disable this mode at the options so to speedup the scan if you are not interested in matrix constructions. Matrix construction is supported by Wheel Generator. You may make a search for matrix at the forums to learn more on this topic.

Q: What is build-up mode?
A: Build-up is basically an one-off attempt for the next draw only. I'd pick a GAT at THE MIDDLE columns mostly. I mentioned that "middle" part at an earlier post. This will be explained more below. Also more info on build-up viewtopic.php?f=23&t=751&p=3321.

Q: What is the use of augmentative panorama mode?
A: Augmentative evaluates GATs based on their X+ hit performance, where X is the column the GAT appears in. So, your GAT ID 2877227 which is positioned in column 1, produced 100% output for 1+ hit. It did deliver this as you have seen: it produced the desired 1+ hit (exactly 1 hit in that case). GAT ID 10591326 which is positioned in column 2 indicates 76% for a 2+ hit. It did deliver 1 hit however, so if we consider this "loss" as part of the other 24% (less than 2-hits), this GAT is expected to produce the 2+ hit at the next draw (Run Factor utilization - will be explained later). GAT ID 407844 which is positioned at column 3, indicates 37% for a 3+ hit, so given it produced 1-hit only, obviously it wasn't the right draw to make that 3+ hit; we potentially have to wait about 100%/37%=3 draws to get that 3+ hit from that GAT, based on its hit performance. More accurate information on its hit performance is available at its red graph line, where there we can see if this GAT still delivers its hits or if it has entered a cold cycle - what we look for is regularity of hits. Inspecting its graph, I can see it does deliver the 3+ hit all over its range, so it is good to go. If I am to expect the 3+ hit from this GAT, I should keep playing it for at least the next 2 draws (100/37) based on its hit performance - using the Run Factor. Please note, using the same GAT ID for the next two draws means, we use the run factor mechanism to get the NEW prediction this GAT ID will propose. The predictions of a GAT change from draw to draw. Similarly for the GAT ID 4146045 which is positioned in column 4, indicates 15% chance for a 4+ hit, which also means 85% chance for less than 4 hits. However 99% of the time this GAT (on its 85% part) will deliver fewer 3-hits, 2-hits etc) compared to GATs positioned in column 3 or 2, otherwise this GAT would also appear in those columns too [check note 1 at the end]. So you should pick a GAT in that column only if you target a 4+ hit judged by the hit performance shown and the expected waiting time to get the desired 4+ hit. Since this GAT at column 4 indicates 15% for a 4+ hit, this means 100/15=6.67 or 7 draws of waiting time on average to get that 4+ hit. This is an ideal estimation because it assumes the GAT produces that 4+ hit at exactly 7 draws every time which is not the case (thus the need to also inspect the red graph). If you don't want to inspect the graph, then use this GAT for a couple more draws, i.e. instead of 7 draws (using Run Factor always) use it up to 9 draws till it produces the desired hit. It may produce the 4+ hit at the very next draw, or a bit later, or beyond the 7-9 draws but you should wait for that hit about that amount of future history draws by using this GAT, however the advice is to always check the red graph for regularity of hits. For a comparison, GAT indicates 6.1 times better performance (the Imp. ratio) to produce a 4+ hit. The typical classic probability for a 4+ hit in your game when picking 12 numbers is once in 40 draws or so. GAT indicates it produces that hit on average at about 100/15=7 or so draws.

Q: So, why do we typically pick from the middle columns?
A: Because these are the columns that offer the maximum performance of X+ hits production where X is a mostly meaningful good win division typically a 2 or 3 or 4 hit, which also demonstrates good overall hit performance compared to natural probability and additionally demonstrate much more easily their "regularity of hits". You cannot judge regularity having just one or two hits indicated. For all these reasons, the best GATs to use are typically the middle ones and use any picked GAT for at least 100/X% indicated by that GAT. This 100/X% is known here as the 100/X rule-guideline.

Q: I picked a GAT ID which failed at the draw I tested it on. How to use the Run Factor for the forthcoming draws?
A: Record down the option parameters Tested Draws, Stat.Data and Start point. Also write down the GAT ID you want. Add the new history draw as it comes. When you are ready to make the new scan to get the prediction of your GAT ID for the new draw, go at the options and verify the 3 parameters ( Tested Draws, Stat.Data and Start point) are the same. You'll notice Run Factor now becomes 1, which means one new draw has been added to the history since the initial run. You can think of RF as how many new history draws have been added since the initial scan you picked the GAT ID (where you had RF=0). Also, place the GAT ID at the "GATs to retain" panel at the options too. This is to forbid the engine from discarding that GAT if during the scan potentially finds better GATs at that time which will move the GAT out of the panorama in which case the system discards the GAT [check note 2]. You may also want to put the maximum GAT ID (if you monitor more than one GAT ID) at the "Maximum GATs to scan" also found at the options so to stop the engine when it reaches the maximum ID you want to monitor. Then start the scan. You'll have to wait again 72 hours (plus a bit more for Run Factor overhead) to reach the ID, so you really have to decide how much time you want the engine to run so to get predictions in time. [check note 3]

PS: You reset Run Factor to 0 by pressing the R button next to the Start point parameter at the option. Doing so, you break any ability to keep using GAT IDs picked, so essentially you start a new fresh run to pick new GATs to use.

[note 1] Some users monitor if a GAT shows up in more than 1 column as potentially an overall even better GAT. So eg if it shows at column 3 and column 4, this means it is a very good 3+ hitter and also a very good 4+ hitter, so it has good qualities in both situations. Might be indeed a very good GAT to pick.

[note 2] If my GAT ID is discarded, does that mean it is a bad GAT since other better emerge? Absolutely not! During a scan with Run Factor, potentially new GATs may emerge as new best performing. However, the reason you initially picked a GAT, is the reason you keep expecting this GAT to deliver. After all, a GAT is not judged by one draw only; it is judged for its delivery ability over the 100/X draws or so, so if this 100/X draws span is lengthy enough, alot of things may change in terms of other GATs, however you haven't given the ability to your original GAT to deliver what it promises. To protect this GAT from getting discarded - if this situation arise - we also indicate it to the engine at the "GAT IDs to retain" at the options. If your retained GAT does not show at the panorama, you can access it by clicking at the ID number shown next to the slider (when viewing the details of a GAT ID). Right-click on this same ID to bring up a menu with some handy options too.

[note 3]
When we scan with RF>0, we essentially care only to reach the retained GAT IDs. In that sense, it is pointless to consume time for the panorama display. There are ways to speedup this scan process.
1) Disable any modes you are not interested in. Basically we do not care at all on what the panorama display; we do not look for new GATs to pick anyway.
2) At the options set the Maximum GATs per column to 3 and also the req. numbers range to 1-1. These settings will minimize all the comparisons needed by the panorama to place GATs in there, thus it will free some computing power to reach quicker the IDs you seek.

lottoburg
Knowledgeable
Knowledgeable
Posts: 117
Joined: Sun Sep 18, 2016 11:06 pm

Re: The problems about GAT's running

Post by lottoburg » Mon Sep 26, 2016 4:29 am

Hi, AT:

Thanks for your detail and patient reply!
I still want to talk my understanding with you even though I need more time to digest your instructions.

The first of all I'm very dispirited to the First Show of GAT Engine for me!
The best result of prediction for the next draw (09/01/2016: 07-17-23-27-32) just is 3 x 2/5 from 15 Top-most GTAs based on 14.667 Million GATs which are generated by the scan of more than 72 hours! Also, I have checked out the all 285 GATs on the screen and found the best prediction for the next draw only is 5 x 3/5 ($25.66 x 5)
which is not placed on the Top-most or the Middle either? IMHO, from the First Show, your GAT Engine is similar to a random Quick Picker! I do hope it would have a good performance in the next backtests.

I try to make a summary for your last post as below:

1) Disregard Absolute mode for regular lotto play;
2) The Build-up is one-off attempt for the next draw only; To use it, I need to search for the best GAT(s) from THE MIDDLE columns.
3) Focus on Augmentative mode; To use it, I need to :
a. Search for the best GATs (at least two) from THE MIDDLE columns first;
b. Search for the best GATs according to their red graph lines in which I can find the
Regularity of Hits (Signatures);
c. Save these best GATs by using "Add GAT to list" & "GAT IDs to retain" before making a new
scan;
d. Utilize these best GATs and Run Factor correctly for the next draws.
4). To speed up the scan process, I need to:
a. Disable any modes but the retained GATs only;
b. Select the 3 parameters (Tested Draws, Stat.Data and Start point) correctly
e.g. Maximum GATs per column=3, req. numbers range=1-1.

Are they above right? If so, my questions are as below:

1) Where is the area of MIDDLE columns? In other words, how can I select the MIDDLE columns from the both modes: Absolute & Augmentative?

2) How can I read the red graph lines in which I can find the Regularity of Hits correctly? Could you explain it by using some examples with some pictures? The best way should be to use some Videos. (I can not get the clear answer from the Help File and Forum)

3) How many draws can I use these best GATs continuously for? Just calculate 100/X or read the red graph together? What are their special patterns in this condition generally?

4) Can I use the next winning #s as the Comparing Numbers which in the window of Predicted Number in the backtests? If so, what's the differences between GAT IDs with the comparing numbers and the GAT IDs without the comparing numbers?
Also, how can I select these Comparing Numbers for the next draw in the real play?

5) How can I disable any modes to speed up the scan? I can not find the options bar from the Engine.

Best regards,
lottoburg

draughtsman
Site Admin
Posts: 112
Joined: Thu Jan 17, 2008 4:22 am
Location: Melbourne, Australia

Re: The problems about GAT's running

Post by draughtsman » Mon Sep 26, 2016 7:50 am

lottoburg,

With all due respect I suggest you read the Help File.

It is all in there.

draughtsman

User avatar
lottoarchitect
Site Admin
Posts: 1635
Joined: Tue Jan 15, 2008 5:03 pm
Location: Greece
Contact:

Re: The problems about GAT's running

Post by lottoarchitect » Mon Sep 26, 2016 10:26 am

Hi, lottoburg
The first of all I'm very dispirited to the First Show of GAT Engine for me!
The best result of prediction for the next draw (09/01/2016: 07-17-23-27-32) just is 3 x 2/5 from 15 Top-most GTAs based on 14.667 Million GATs which are generated by the scan of more than 72 hours! Also, I have checked out the all 285 GATs on the screen and found the best prediction for the next draw only is 5 x 3/5 ($25.66 x 5)
which is not placed on the Top-most or the Middle either? IMHO, from the First Show, your GAT Engine is similar to a random Quick Picker! I do hope it would have a good performance in the next backtests.
I highly doubt you have counted correctly. Fun fact, over 285 tries of 12 picked numbers should produce on average around 38 GATs that deliver the 3 hit based on natural probability. So you say you have counted only 5!!! That is not even close to a quick pick mechanism, it is much much worse. This, if ever occur would be a rare occurrence of such bad performance and I don't think I ever recall such a bad result in that many years of using GAT; actually I don't recall any instance that GAT did worse than natural probability and most of the time is quite better anyway. You did this manually, by hand, how sure can you be that you did it properly and counted accurately? I'll surely run this myself to verify you did something wrong over there. However, I am not happy with your bold statement at all. How you'll feel for your statement if this eventually (and probably) turn out this was a miscalculation on your end? Very unfair comment given you still don't know how to use the system or properly understand what it shows to you.
1) Disregard Absolute mode for regular lotto play;
2) The Build-up is one-off attempt for the next draw only; To use it, I need to search for the best GAT(s) from THE MIDDLE columns.
3) Focus on Augmentative mode; To use it, I need to :
a. Search for the best GATs (at least two) from THE MIDDLE columns first;
b. Search for the best GATs according to their red graph lines in which I can find the
Regularity of Hits (Signatures);
c. Save these best GATs by using "Add GAT to list" & "GAT IDs to retain" before making a new
scan;
d. Utilize these best GATs and Run Factor correctly for the next draws.
4). To speed up the scan process, I need to:
a. Disable any modes but the retained GATs only;
b. Select the 3 parameters (Tested Draws, Stat.Data and Start point) correctly
e.g. Maximum GATs per column=3, req. numbers range=1-1.
3a. One GAT for normal play. 2 or more are for absolute mode mostly. You pick the GAT and keep using it for the next future draws (RF utilization). More advanced utilization usually involve usage of many GATs, even in augmentative mode.
3c. No need to add GAT to list. GAT IDs to retain so to be able and retrieve your picked GAT ID (RF > 0 usage).
4a. One mode must always be active. Enable the augmentative if you picked a GAT from that mode.

Are they above right? If so, my questions are as below:

1) Where is the area of MIDDLE columns? In other words, how can I select the MIDDLE columns from the both modes: Absolute & Augmentative?
> Middle columns are those that are in the middle of the panorama! Not the 1-hit nor the maximum hit column.

2) How can I read the red graph lines in which I can find the Regularity of Hits correctly? Could you explain it by using some examples with some pictures? The best way should be to use some Videos. (I can not get the clear answer from the Help File and Forum)
> I did that a few times already in the previous posts. If it delivers the desired hit all over the range, is has a good regularity.

3) How many draws can I use these best GATs continuously for? Just calculate 100/X or read the red graph together? What are their special patterns in this condition generally?
> Typically 100/X plus a couple more draws. If it don't deliver over that range, probably it is in cold cycle where another GAT might be more appropriate to pick.

4) Can I use the next winning #s as the Comparing Numbers which in the window of Predicted Number in the backtests? If so, what's the differences between GAT IDs with the comparing numbers and the GAT IDs without the comparing numbers?
Also, how can I select these Comparing Numbers for the next draw in the real play?
> Compare numbers must have the next real draw numbers (and that draw must not be in your history). If you enter numbers at the compare numbers box, you enable the coloring of the panorama where the greener the cell, the more numbers that GAT ID predicted. Read the help file for more on this function.

5) How can I disable any modes to speed up the scan? I can not find the options bar from the Engine.
> There is not options bar anywhere. All the options are accessed from the main window screen (before starting a scan) where it says "GAT Analysis Options". If you are at the panorama display, you cannot change any options.

lottoburg
Knowledgeable
Knowledgeable
Posts: 117
Joined: Sun Sep 18, 2016 11:06 pm

Re: The problems about GAT's running

Post by lottoburg » Mon Sep 26, 2016 1:00 pm

Hi, draughtsman
With all due respect I suggest you read the Help File.
IMHO, your Help File is not a good production since its fuzzy and abstract.
It lacks operability really even though I have read it again and again.

Also, a lot of Posts including the pictures in the Forum can not be read and learned since their pictures had been removed or these posts were too old since 2012.

I do hope you can update and improve it further.

I don't understand why not use some videos which are very clear and simple instead of reading a lot of fuzzy and abstract written documents. Of cause, you can set an authority in these videos for the license users only.

Best regards,
lottoburg

lottoburg
Knowledgeable
Knowledgeable
Posts: 117
Joined: Sun Sep 18, 2016 11:06 pm

Re: The problems about GAT's running

Post by lottoburg » Mon Sep 26, 2016 1:24 pm

Hi, AT:
I highly doubt you have counted correctly. Fun fact, over 285 tries of 12 picked numbers should produce on average around 38 GATs that deliver the 3 hit based on natural probability. So you say you have counted only 5!!!....
I'm so sorry about that but it's true. It's my dream that GAT can work for my Take 5 really
and I'm working for it all days now. I do hope you can play it by yourself and give me some smart and operable strategies for Take 5/39. I'm looking forward to it now.

I will reply your other answers in detail later.

Best regards,
lottoburg

User avatar
lottoarchitect
Site Admin
Posts: 1635
Joined: Tue Jan 15, 2008 5:03 pm
Location: Greece
Contact:

Re: The problems about GAT's running

Post by lottoarchitect » Fri Sep 30, 2016 1:04 pm

I have made the run for NY-Take 5 draw 1 September 2016. As I said, I highly doubt you have counted correctly lottoburg and below I present the outcome per panorama scan mode that verify what I said. I also include your own initial image of the scan to verify it is the same run. As you can see, all three modes have predicted 1-2 4-hits and more than 5 3-hitters in each mode. I'll not dive in more detail and analysis here. This is just to debunk unfair comments made.
Augmentative mode http://anastasios-tampakis.net/images/f ... tative.png
Absolute mode http://anastasios-tampakis.net/images/f ... solute.png
Build-up mode http://anastasios-tampakis.net/images/f ... uildup.png

lottoburg
Knowledgeable
Knowledgeable
Posts: 117
Joined: Sun Sep 18, 2016 11:06 pm

Re: The problems about GAT's running

Post by lottoburg » Sat Oct 01, 2016 1:59 am

AT,

I'm not sure why I can not input the next new drawing number into my database for backtests.
My operations are as below:

1) Click the item of "lottery" in the top left corner;
2) Open the window of "Lottery Draws History Management";
3) Click the bar of "Import" in the bottom middle;
4) Open the browser of "Drawing history File" and select the new drawing text file;

But the information (please see the picture as below) always appeared on the screen, then it's still nothing after clicking "OK". How can I do that next?

http://prntscr.com/coehdm

Best regards,
lottoburg

lottoburg
Knowledgeable
Knowledgeable
Posts: 117
Joined: Sun Sep 18, 2016 11:06 pm

Re: The problems about GAT's running

Post by lottoburg » Sat Oct 01, 2016 2:46 am

AT,
I have made the run for NY-Take 5 draw 1 September 2016. As I said, I highly doubt you have counted correctly and below I present the outcome per panorama scan mode that verify what I said. I also include your own initial image of the scan to verify it is the same run. As you can see, all three modes have predicted 1-2 4-hits and more than 5 3-hitters in each mode. I'll not dive in more detail and analysis here. This is just to debunk unfair comments made.
Augmentative mode http://anastasios-tampakis.net/images/f ... tative.png
Absolute mode http://anastasios-tampakis.net/images/f ... solute.png
Build-up mode http://anastasios-tampakis.net/images/f ... uildup.png
1) First, I highly doubt you have made the run for the draw of 09/01/16 of NY Take 5 by yourself!!! The proof is two:
a. Why your Currently testing GAT is the same thing with mine: 14666579? It's impossible absolutely if you made the new run for Take 5 by yourself instead of copying mine.
b. Why your all 285 GATs in the three models are the same thing completely with mine? It's impossible absolutely if you made the new run for Take 5 by yourself instead of copying mine.

2) Second, you said: "all three models have predicted 1-2 4-hits..."
Could you copy the bars of "Predicted Numbers" which include the 1-2 4/5 hits here? The picture like the as below: (The draw #s of 09/01/16 NY Take 5 = 07-17-23-27-32)

http://prntscr.com/cm0c0q

Best regards,
lottoburg

User avatar
lottoarchitect
Site Admin
Posts: 1635
Joined: Tue Jan 15, 2008 5:03 pm
Location: Greece
Contact:

Re: The problems about GAT's running

Post by lottoarchitect » Sat Oct 01, 2016 11:31 am

I really don't like what you imply with your post, honestly do you say I have "cooked" these results?? Everyone can verify this by running GAT with your same parameters and history. One great aspect of GAT is, the results are repeatable, so whatever you have done over there, I can re-run it and re-create your outcome, provided I run it with the exact same parameters and history. I have done exactly that and my only "worry" here was, if the run you performed there was wrong or you missed something like an incorrect entry to the history which would result in a different outcome. Since I have ended up with the exact same results, indeed I have performed the exact same run as yours.
1) First, I highly doubt you have made the run for the draw of 09/01/16 of NY Take 5 by yourself!!! The proof is two:
a. Why your Currently testing GAT is the same thing with mine: 14666579? It's impossible absolutely if you made the new run for Take 5 by yourself instead of copying mine.
b. Why your all 285 GATs in the three models are the same thing completely with mine? It's impossible absolutely if you made the new run for Take 5 by yourself instead of copying mine.
This gets funnier and funnier, apart the "cooked" part you imply I have done which is atrocious on your behalf. You are so sure what you say is "proof" (like your counting accuracy) however: a) I can end the scan at the exact same GAT ID you have ended yours; this function is known as "Maximum GATs to scan" at the options although naturally it is not used to face accusations but it was handy in this case here b) Explained above.

Finally, for everyone interested, I have uploaded the gsf file (scan export of this run) so to check everything you want, including the parameters, history and draw to be predicted.
https://www.mediafire.com/?faawui2r687tafj

Post Reply

Who is online

Users browsing this forum: No registered users and 21 guests