Comparison of the standard syntax and the proposed syntax.

The standard syntax suggests the following use:

: foo-aux <what-must-be-done> ;
: foo
	begin-loop		\ different kinds of loops may be used here
		['] foo-aux CATCH
		<leave-the-loop-if-successful>
		<reactions-on-the-error-conditions>
	end-of-loop
	<raise-an-exception-if-there-has-been-no-success>
;

The proposed syntax suggests the following use:

: foo
	TRY
	CATCH{
		<what-must-be-done>
	}CATCH
	?DUP IF
		<reactions-on-the-error-conditions>
		N ( times to retry) RETRY
		( errc) THROW
	THEN
;


A more expanded example.
The lines responsible for counting and retrying attempts, and for
propagation of exceptions are marked by ( *** ). You see 7 such lines
in the standard syntax version, and only two in the proposed syntax version.

The standard syntax version:

\ unique error codes
CREATE error1
CREATE error2
CREATE error3

: do-something1
	... IF error1 THROW THEN ...
;
: do-something2
	... IF error1 THROW THEN ...
	... IF error2 THROW THEN ...
	... IF error3 THROW THEN ...
;
: do-something3
	... IF error3 THROW THEN ...
;

: foo-aux
	do-something1
	do-something2
	do-something3
;
: foo
	0 >R					( *** )
	BEGIN					( *** )
		['] foo-aux CATCH
		?DUP				( *** )
	WHILE					( *** )
		DUP
		CASE
		error1 OF reaction1 ENDOF
		error2 OF reaction2 ENDOF
		error3 OF reaction3 ENDOF
		         default-reaction
		ENDCASE
		( errc) R@ 3 > AND THROW	( *** )
		R> 1+ >R			( *** ) \ count attempts
	REPEAT
	R> DROP					( *** )
;

The version that uses the proposed syntax:

\ unique error codes
CREATE error1
CREATE error2
CREATE error3

: do-something1
	... IF error1 THROW THEN ...
;
: do-something2
	... IF error1 THROW THEN ...
	... IF error2 THROW THEN ...
	... IF error3 THROW THEN ...
;
: do-something3
	... IF error3 THROW THEN ...
;

\ -- till now the code has been the same --

: foo
    TRY			( *** )	\ the place from which attempt retrial begins
    CATCH{			\ the code where exceptions can happen follows
				\ (the application logic)
	do-something1
	do-something2
	do-something3
    }CATCH			\ the code that examines whether the last 
				\ attempt was successful follows
				\ (the error handling logic)
    ?DUP
    IF
	DUP
	CASE
	error1 OF reaction1 ENDOF
	error2 OF reaction2 ENDOF
	error3 OF reaction3 ENDOF
		 default-reaction
	ENDCASE
	( errc)
	3 RETRY THROW	( *** )	\ besides TRY, this is the only line
				\ responsible for attempts counting&retrying
				\ and for error propagation
    THEN
;

Conclusion

The same amount of the application code and the error analysis code, besides 
that the auxiliary procedure is not required;
less code required to organize retrying attempts.