Repeat analysis

[Frank]

It's been a long time since I posted in this thread, but I have persevered with this system and recorded over the weeks how well or how badly my lists did against each predicted result.
My pick ones and pick fours are derived from this analysis, and now and again it seems to get one or two right, such as in the last draw 1494, I got two right for the purposes of pick four.

I have a scoring system based on how probable a number is to repeat to the next draw but many numbers share the same score, so I have to put them into colour groups, same colour = same score. In the last draw for example 1494 ,The scores taken after draw 1493 were

Score >3............. 17,24 (purple)
Score =3...............3,10,29,36 (blue)
Score=2 ...............1,5,6,7,11,18,26,27,28,31,34,42,44,47,49 (yellow)

and there are more as you will see in the attached image (shows the last few weeks only) and white ones score zero and expected to be eliminated. The Actual result matches have red fonts for the main balls.

What I'm striving to do by tweaking a few things and adding some extras based on Lens Data sheet is get results which fall in the purple and blue zones, the leftmost groups.
   

I just got lucky in the last draw as 3,10,29,36 were drawn in 1494, and 31 and 44 were in the next block (yellow) I really need my scores to be wider ranging than 0 to 4 or 5 as they are currently coming out, so I can have more (and smaller) colour groups. Then the aim of getting more results in the (smaller) leftmost groups can be aimed for. As you can see, not all past attempts have been as good as the last draw. But I keep going on and tweaking. My prediction (for what its worth) is already in the chart and my pick four happen to be in the purple zone for this draw.