I have listed below several values for a function F(a,b), where a and b represent decimal digits. The samples are tabulated in 12 rows, numbered 0 to 11:

Row 0: F(1,0)=10
Row 1: No samples
Row 2: F(2,8)=28; F(3,9)=39
Row 3: F(2,7)=27; F(4,9)=49
Row 4: F(2,6)=26; F(3,8)=38; F(5,9)=59
Row 5: F(2,5)=25; F(6,9)=69
Row 6: F(2,4)=24; F(3,7)=37; F(4,8)=48; F(7,9)=79
Row 7: F(2,3)=23; F(8,9)=89
Row 8: F(2,2)=22 ; F(3,6)=36; ; F(5,8)=58; F(9,9)=99
Row 9: F(2,1)=21; F(4,7)=47;
Row 10: F(2,0)=20; F(3,5)=35; F(6,8)=68
Row 11: No samples.

One might deduce that F(a,b)=10*a+b, were it not for the rows 1 and 11.
For those rows there are no "red herring" samples to lead you into false conclusions.

Actually you are requested to:
1. Find a row-dependent function G(a,b,r), r representing the row number, for which all the above values fit, - and in rows 1 & 11 the following is true:

G(7,3,1)=32; G(6,8,1)=59; and G(6,4,11)=45; G(1,11,11)=32;

2. Evaluate G(3,7,8).

3. Explain, why there are no "misleading examples" for row 1, 11 and higher.