qeFT returning rounded errors difficulting selection of the best one.

Hi, consider:
```
data( "oliveoils" )
qeFT( data = oliveoils, "Area", pars = list( k = c( 5, 10, 15, 25 ) ), qeftn = "qeKNN", nXval = 5, nTst = 100 )
```


When accuracy values are below 1, as in the example above, they all round to 1. I believe this is incorrect since I ran ```
replicMeans(nrep = 5, "qeKNN(oliveoils, 'Area', k = 5)$testAcc")```, for each different k tried before and it produced different values between them.

Start by running the oliveoils qeFT example with qeKNN and compare its displayed accuracy values with replicMeans results for each k. Trace the qeFT output and formatting path; done when distinct accuracy values below 1 remain distinguishable instead of all displaying as 1.

Written by the indexing model from the issue text.