The graph starts on the 10th death not the 10th day.
The problem is you don't know when day 1 was. Picking the 10th day, approximately the 1000th case starts with a statistically significant number of cases, normalizing the time 0 for staring the exponential growth.
CT, do you think NY is overstated or understated? Have you taken a look at the respective wiki pages I referenced? I don't think the OCD autistics who are painstakingly collecting the state by state data are off. If anything is incorrect, it would be the tallies being reported ie source docs that they are using.