I would like the system to some scenario testing capability when it comes to selling shares.
I would like to scenarios test particularly around the effect of selling a holding or number of holdings. What would be the impact on my:
By being able to test a scenario, I won’t have to pretend to carry out the actions and subsequently have to re-instate be real position.
Your question is not clear enough to know what you are looking to achieve.
Are you trying to measure what the return would have been if you had not sold the stocks, or are you trying to measure the return of the stocks you sold?
I think he would like to set up a tracking portfolio to see what would happen if he changed certain parameters in his real portfolio. He may simply need to set up a tracking portfolio that is like his real portfolio except for the desired changes. I think he is allowed to do this if he is a paying customer. One can try various scenarios at “Portfolio Visualizer” using historical data.
It is possible to set up “What-if” portfolios in Sharesight. For example,
“Reality”: My real life portfolio with all purchases and sales entered correctly.
“Never-Sold”. Same portfolio as the "Reality portfolio, but without entering any of the sales.
“All-Sold” All the stocks that I sold, but entered as if they were purchases. I want this portfolio to under-perform my “Reality” portfolio.
.
Thinking it through, the way this would work best is to allow a list of shares to to be individually switched “on” or “off” for the performance calculation by the way of a checkbox.
That way, any combination of the stocks in a portfolio could be used to calculate performance without losing the history and without having to have multiple portfolios.
Developers of Sharesight: Please take note and think about adding this feature. 
Thank you for your thoughts on this.
Giving this a little more thought, it would be really great if you could just create a copy of your current portfolio with the all transactions from that portfolio. This would allow all of the current functionality to work against the ‘Test Scenario’ portfolio so that you test out various scenarios for example - end of year tax planning.
The functionality should allow you to delete the ‘Test Scenario’ portfolio and create a new one.
I would expect that ShareSight would want to put a time limit on the ‘Test Scenario’ for example 24 hours before it is automatically deleted.
There is a way you can do this utilising the All Trades Report. We currently have a guide for recreating your portfolio in a different tax residency below:
The same rules would apply, you just set up the test scenario under the same conditions as your actual portfolio. You would then just remove any dividend reinvestments and sales you don’t want to include in the test scenario from the Excel file download of the All Trades Report 
I just tested this and can confirm it works. Thank you for being so responsive.
It’s a solution for testing multiple scenarios without too much work.
However your linked instructions are insufficiently explained for those who are not good at Excel.
In particular you need to expand on instruction 5 (Convert to CSV).
You also need to specify that the date column (which formatted by default as dd.mm.yyyy on my Mac) needs to be reformatted to a custom date format of dd/mm/yyyy, otherwise every date reverts to tomorrow’s date, which throws up an error (date must be in the past).
Lastly, when uploading, share splits show as an error. They need to be manually adjusted to “Buy” at a price of Zero, in order to proceed.
So with your Mac, are you using Excel or Numbers to open the file? I know with the Apple-branded Numbers it can change the date formatting as you have outlined.
Sorry, I should have specified. The file import should only show buy & sell trades. DRPs need to be actioned manually and splits/consolidations will generate automatically as well. So as long as you have the correct buy & sell data, the other stuff should generate internally
I am using Excel (Microsoft 365) for Mac. I just tried it again, this morning and did not need to correct the date format any more. Maybe something changed?
That’s pretty interesting, I’m not sure why that would happen 
The formatting that you saw the second time is how it should be and what it is designed to show. So it sounds like it was just a one-off event.