That being said, as I understand it, reality still is fuckin’ weird. Consider this experiment where I’ll avoid using the terms particles or photon or anything that might not exist. Just describing what happens.
Say we have a “doohicker” device with a light on it. And then we have two “globber” devices. The ‘globber’ device is a cylinder with an ‘input’ on one end and two outputs on the other end. The output end has two lights on it, one on the top half and one on the bottom half.
If we line up the doohicker device with two globber devices on opposite sides, input ends pointing toward doohicker device, then once/minute the doohicker device flashes and then a short time after, the globber devices flash, each either the top light or the bottom light. So we say the doohicker device is doing something which makes the globber devices flash.
Now say we set up these devices and the globber devices are 2km away from the doohicker device. We notice btw that ~6.7 microseconds after the doohicker device flashes, the globber devices flash.
Near each globber device we have a piece of paper and we write down the minute, the rotation of the globber device, and whether the globber device read ‘up’ or ‘down’. And we randomly rotate it as the experiment goes on.
We notice that each globber device individually always gives 50/50 up or down, regardless of the rotation.
Now we bring the two pieces of paper together to compare the results of the same emission across the two globber devices. And we notice they are correlated!!
Namely: whenever the globber devices are rotated 180 degrees away from each other, they both either flash top or bottom. Perfectly correlated. If they are rotated the same degree then when one flashes top, the other flashes bottom, and vice versa.
Whenever they were rotated 90 degrees away from each other, then there was no correlation. Whether the first flashed top or bottom, the other flashed top 50% of the time and bottom 50% of the time.
Whenever they were rotated 45 degrees apart, then there was some correlation in between. When one flashed top, the other flashed top 14.64% of the time and bottom 85.36% of the time. When one flashed bottom, the other flashed top 85.36% of the time and bottom 14.64% of the time.
This happens even though we were randomly rotating the globber devices!
This is what actually happens in reality ^ .
You might imagine that that’s because the doohicker devices emits some real “thing” that definitely has some property, and the globber devices are measuring this property, and they have some chance of yes/no based on the actual property of the thing. But Bell showed that any model you come up with that assumes this ^ will not be able to model what actually happens! He showed that, mathematically, you can only model this if you model the probabilities of both devices agreeing/disagreeing, based on the angles of both devices. Even though they are far apart from each other!
But locally each device always gives 50/50. You can’t tell how the other device is rotated when you look just at yours. But when you combine the results together, you notice weird things like “hey whenever I was at angle X and the other device was at angle Y I was only getting up 14% of the time if it got down…”