CAN 6/49 no Bonus
Wanna know what I think? Well I'll tell you anyway

. This may apply to "law of averages".
Look at snapshots of most/least drawn numbers, at 500 draw intervals. I'm only gonna use 2 top bananas and 2 bottom bananas to make it easy on myself.
Draw/#/expected hits/actual hits/Ratio
------------------------------
500/31/61.2/79/129.0%
500/43/61.2/77/125.8%
500/15/61.2/48/078.4%
500/12/61.2/46/075.1%
1000/31/122.4/152/124.1%
1000/43/122.4/145/118.4%
1000/12/122.4/104/084.9%
1000/15/122.4/103/084.1%
1500/31/183.7/233/126.9%
1500/34/183.7/210/114.3%
1500/15/183.7/161/087.7%
1500/11/183.7/160/087.1%
2000/31/244.9/290/118.4%
2000/34/244.9/283/115.6%
2000/13/244.9/221/090.2%
2000/11/244.9/220/089.9%
It's interesting to look at the actual/expected hits Ratio difference between most and least: at draw 500 it is 53.9%, at draw 2000 it is 28.5%. My guess is that by draw 14,000,000 this difference will be negligible (we only have to wait about 140,000 years to confirm this).
What I'm saying is that at present we are looking at a tiny fraction (0.015%) of a span it may take to possibly cover all combinations and have things even out, as I've heard we should expect. So any difference now is just a blip and shouldn't last more than a few thousand years give or take a hundred.
Looks like the top banana 31 has slowed down a bit of late and maybe caught soon (in another few years) by the 47, or 34 or...?
But who knows? as Sheba put it more succinctly.
P.S. someone can check my figures but it's not really necessary as they're meant for illustration mainly.