Hope y'all are being facetious...'cuz I was...

Anyway...I've got enough data now to figure how things are going to look, at least for me. I've got three S7s in Labrador that are running at a 50% overhead, including power and psu's, based on the pool's production since the halving. Those particular units were ROI already, so that's a moot question in this instance. An S9 would run at the same absolute overhead as an S7, but produces around 12THs, v. 4THs. The math isn't complicated. It will make money on a monthly basis, as do the S7s.
The problem, of course, is in the ROI. Until Bitmain gets the cost down on the S9, I'm looking at running A6s instead. If I ran an S9 at the datacentre, it would take (given spot staying stable) around ten months for ROI.
Am I missing something in my math? S9s run ~$2000. 13T/s gets you ~.003/block. That's ~1000 blocks for roi. Where do you get your 10 months from?From kano.is block stats
Description Time MeanTx% Diff% Mean% CDF[Erl] Luck%
Last 5 Blocks 2.1days 104.9% 187.78% 187.78% 0.9568 53.25%
Last 10 Blocks 4.8days 104.6% 206.85% 206.85% 0.9967 48.34%
Last 25 Blocks 7.5days 103.8% 129.08% 129.08% 0.9190 77.47%
Last 50 Blocks 13.5days 103.6% 115.96% 115.96% 0.8687 86.23%
Last 100 Blocks 25.3days 103.0% 104.70% 104.69% 0.6899 95.52%
Last 250 Blocks 8.4wks 102.2% 95.47% 95.11% 0.2222 105.14%
Last 500 Blocks 16.9wks 101.7% 96.27% 96.15% 0.1954 104.01%
2 x this is 33.8 weeks that is 8 monthsLast 1000 Blocks 63.4wks 101.4% 93.61% 93.08% 0.0128 107.44%
this is 14 monthsAll - Last 1232 Blocks 97.2wks 101.2% 94.01% 94.24% 0.0200 106.11%
she may have extrapolated 10 months based on those 2 stats. it is reasonable to think it could be 10 months