I recently got the combo changed using that combo generator.
I underestimated this lock and it promptly taught me a lesson. But let's proceed one step at a time.
I made the first graph testing all wheels left and was very happy because despite a random combo I found so many gates.
So I took my "ruler" and checked which gates could have been the true ones. I found two options.
I was so sure about those gates being the true ones that I only made a couple high-low tests.
So instead of making a few high-low tests more, I concluded Left 10 was the true gate on wheel #2. I converted it to Right 9.5 and started graph #2: wheels #1 and #2 at Right 9.5, wheel #3 around Left. It seemed the quickest option at the time. I found a few gates but only amplified around the ones I thought was the true one, finding its center at Left 78.5. This time I had no doubt because I was testing only wheel #3 so my ruler identified the true gate with ease.
It was getting late and I was hungry, but since I thought I was close to an opening, I went on manipulating.
Just to keep track of what I was doing I made a quick and small graph of wheel #1 around left. I wish I paid more attention to it. Based on gates location I concluded wheel #1 may be 17. Here is the graph.
A bit disappointed I left the lock alone and grabbed something to eat. A couple days later I had some time to play so had a look at my graphs. In theory the lock should be open already. In practice I had just found false gates. Since it's the quickest test to run, I decided to verify wheel #3 before going on. My graph confirmed its position at Left 78.
So either wheel #1 or wheel #2 needed to be checked again. I started with wheel #2 with wheel #1 at 17 and wheel #3 at 78, as you can see on graph #5.
Having two wheels at their gates position, this test gave me huge indications on the gates on wheel #2. Contact points were more than one whole number closer than on the rest of the wheel.
It was easy to say that wheel #2 was 90.
The last thing I needed to do was brute force wheel #1. It took me less than 15 minutes to do so. The lock opened with 48 - 90 - 78.
In conclusion: If I had paid more attention at the tests after the first graph, I may haven't needed to make four more.
Cheers