Summary of Options

Word Weight Modification in N3SCM

In N3SCL, the variation of the maximum Score was applied prior to word pairs being chosen, and thus contributed to the choice of the word pairs. The effect was that more significance was given to the word types (MAJMOD, SKIPMOD) or categories (CATSW) of word pairs, rather than to their likeness.
For example, with N3SCL,
(using
REFMIN, MAJMOD*120, NAME-FORMAT=R, NOSTD, NORAW
),
SEARCH: JOHN ALLEN ANDERSON
FILE: JOHN ANDERSON ALLEN
scored 035, because with the MAJMOD*120 setting, the ANDERSON vs ALLEN pair scored higher ((30/100)*12) than the ANDERSON vs ANDERSON pair ((100/100)*1), and would be chosen.
Using the default N3SCM processing, the same match would score lower (021), because the word pairs are first chosen on their natural likeness prior to applying the
MAJMOD
option. In this example, the word pairs ANDERSON vs ANDERSON and ALLEN vs ALLEN score higher ((100/100)*1) than ANDERSON vs ALLEN pair ((30/100)*1) leaving no major word match to apply the weight modifier to.
In N3SCL, the weight of a word pair was allowed to be less than 100. Using N3SCM default settings, the weight cannot be less than 100.
The effect of this can be shown by the following CATSW example.
*C NN R Nick-names
NN JONATHON                   >JOHN             <
OPTION CATSW
VALUE NN,9
SEARCH: JOHN SMITH
FILE: JONATHON SMITH
In N3SCL,
SCORE: 100 (9/9 + 10/10)
In N3SCM,
SCORE: 95 (90/100 + 100/100)
To override this default behavior, and have N3SCM operate the same as N3SCL, use options
CATSS
,
LIMWCAT
or
USECWAIT
as described below.

Local Options Addressing Truncation & Initials

*C PT D Personal Title
*P SYSTEMS INTEGRATION MANAGER
*R MANAGER
PT MANAGER><

SEARCH: SYSTEMS INTEGRATION MANAGER
FILE: SYSTEMS INTEGRATION MANAGER
SCORE: 100

SEARCH: SYSTEMS INTEGRATION MANAGER
FILE: SYSTEMS INTEGRATION MANAGER
SCORE: 000

U.T.S.
University of Technology Sydney

SCORE: 010

SCORE: 100

PETER
PETERS P

SEARCH: HELEN M RICHARDSON
FILE: MICHAEL RICHARDSON
SCORE: 095

SEARCH: HELEN M RICHARDSON
FILE: MICHAEL RICHARDSON
SCORE: 065

     LOPT=(INITLOW)
     OPTION SCORES
     VALUE INIT,10

SEARCH: JOHN SMITH
FILE: J SNITH
SCORE: 055

     OPTION SCORES
     VALUE ILOWTRIG,8

SEARCH: JOHN SMITH
FILE: J SNITH
SCORE: 090

SEARCH: ROBERT
FILE:   ROBERTA
SCORE:  080

SEARCH: ROBERT
FILE:   ROBERTA
SCORE:  100

SEARCH: RO
FILE:   ROBERTA
SCORE:  030

IDENTITY SYSTEMS LTD
IS

IDENTITY SYSTEMS PTY LTD
ISS

Local Options Addressing Concatenation

ROBERT HACKFORTH JONES

ROBERT HACKFORTHJONES

LOPT=(CONC+CINITA)

Local Options Addressing Word Order

JOHN SMITH
JOHN ALAN SMITH

SEARCH:  ANDREW JOHN SMITH
FILE:    JOHN SMITH
SCORE:   100

SEARCH:  ANDREW JOHN SMITH
FILE:    JOHN SMITH
SCORE:   099

VALUE PERFLAG,0

VALUE PERFLAG,1

VALUE PERFLAG,2

Local Options Addressing Word Type

Codes
Major Words
Skip Words
Exact Words
Edit-list Words

HONG, HOANG
HOANG, HANG FAI

LOPT=(MAJMOD*20)

*C SS S Skip Word
SS LABS            >         <

*C SN I Secondary Lookup Word
__
SN MARKETING        >ADVERTISING       <

*C CS S Skip Word
*C RR R Replacement Word
RR DEPT            >DEPARTMENT  <
CS DEPARTMENT      >            <

OPTION CATNREP
VALUE SND,0 <= change Secondary Names (SN) to Delete
VALUE PTS,0 <= change Personal Titles (PT) to Skip

OPTION CATNIGN
VALUE NN,0 <= disable the Nickname (NN) category

OPTION CATTREP
VALUE SD,0 <= change all Skip (S) to Delete (D)

OPTION CATTIGN
VALUE R,0 <= ignore all Replace (R) types

*C NM 7 Do Not Match
NM PETER    >PETA      <
NM PETER    >PETRA     <

Local Options Addressing Spelling

OPTION FLAGS
VALUE TRANSLEN,5

OPTION SCORES
VALUE STD,8

SEARCH: NATHAN
FILE:   NATHON
SCORE:  055

ORIGWTHR,90
ORIGWSCR,100

SEARCH: NATHAN
FILE:   NATHON
SCORE:  083

OPTION FLAGS
VALUE MATCHEND,1

Tiene vs Tienne scores 8/10

SEARCH: CECILIA M SMITH
FILE:   JAQUELINE M SMITH
SCORE:  076

SEARCH: CECILIA M SMITH
FILE:   JAQUELINE M SMITH
SCORE:  066

SEARCH: MICKALSEN
FILE:   MICKALSON
SCORE:  080

SEARCH: MICKALSEN
FILE:   MICKALSON
SCORE:  087

new score = raw score + ((100 - raw score) * stabilized score / RAWSTBVL)

Local Options Addressing Long Names or Addresses

(Original Score + CLIMIT Score * N/10) * 1/2

GOPT: REFMIN

SEARCH: 3, 94 MILLER ST SYDNEY 2060
FILE: 94 MILLER ST NORTH SYDNEY NSW 2060

OPTION CLIMLIST
VALUE SMI,0

Local Options Addressing Multi-valued Fields

Account names & Compound names
Secondary names
Repeating fields

Account Names & Compound Names

The method will automatically compare Account and Compound Names if these features are turned on in the Algorithm being used by the method. For a description of Account & Compound names see the

Multi-Valued Fields

section. As an example, if the Compound Name feature is turned on and

T/AS

is defined as a Compound Name Marker and the

REFMAX

Global Option is defined in the Matching Scheme, then:

SEARCH: SNAPPY INVESTMENTS
FILE:   ABC HOLDINGS T/AS SNAPPY INVESTMENTS
SCORE:  100

Secondary Names

To use the Secondary name feature, Secondary name entries must first be defined in the Edit-list being used by the method’s Algorithm. For a description of Secondary name Edit-list entries see the

Multi-Valued Fields

section.

When comparing a word that is defined as a Secondary name, the method can then compare all its replacement values at the same time.

To specify what types of Secondary names should be expanded, the following

Local

Option must be set:

Option	Description
OPTION FLAGS VALUE SECOND,{0/1/2/3/4/5}	0 – do not expand secondary names (default) 1 – expand only if leftmost minor word (only Y types) 2 – expand only if rightmost minor word (only Y types) 3 – expand all words (all types) 4 – expand only minor words (not M or N types) 5 – expand only major words (M or N types) As an example, when matching addresses, this feature can be used to give equality to neighboring localities, provided the Edit-list is set-up with the appropriate values. For example, assuming NEWTOWN is next to MIDTOWN, and MIDTOWN is next to OLDTOWN, but NEWTOWN is not next to OLDTOWN, then the Edit-list should contain at least the following: *C SN I Secondary Names SN NEWTOWN >MIDTOWN < SN OLDTOWN >MIDTOWN < Then, if using: OPTION FLAGS VALUE SECOND,3 SEARCH: 25 MAIN STREET NEWTOWN FILE: 25 MAIN STREET MIDTOWN SCORE: 100 SEARCH: 25 MAIN STREET NEWTOWN FILE: 25 MAIN STREET OLDTOWN SCORE: 066
OPTION FLAGS VALUE SECPHRSE,{0/1/2}	This option allows you to create secondary phrase names. This matching option is equivalent to NAMESET function keyword SECPHRASE or SECPHRASEALL . Value 0 turns this feature off, 1 turns the feature on (NAMESET SECPHRASE ) and 2 creates all secondary names (NAMESET SECPHRASEALL ). The default is 0.
OPTION FLAGS VALUE SECPORIG,{0/1}	This option allows you to include original names before secondary phrase rules are applied. This matching option is equivalent to NAMESET function keyword SECPHRASEORIG . Value 0 turns this feature off and 1 turns the feature on (NAMESET SECPHRASEORIG ). The default is 0.
OPTION SCORES VALUE SECPHRSE,n	This option allows you to specify the maximum score to apply to secondary phrase matches. Value 0 turns this feature off and will assign a maximum score of 100. The default is 0.
OPTION SCORES OPTION SECCAT VALUE [Edit-list Category],1	Requires that both Secondary words being matched are in the [Edit-list Category] specified. Multiple Edit-list Categories can be specified using multiple VALUE statements. Requires OPTION FLAGS , VALUE SECOND to be set to non-zero value. For example, OPTION FLAGS VALUE SECOND,3 OPTION SECCAT VALUE SN,1 will only perform Secondary name matching on words which are in the SN Edit-list category.
OPTION SECTYPE VALUE [Word-type],1	Requires that both Secondary words being matched have the [Word-type] specified. Multiple Wordtypes can be specified using multiple VALUE statements. Requires OPTION FLAGS , VALUE SECOND to be set to non-zero value. For example, OPTION FLAGS VALUE SECOND,3 OPTION SECTYPE VALUE S,1 will only perform Secondary name matching on words which have a Word-type of S (Skip).
OPTION SCORES VALUE SEC,[Number]	This option allows you to specify the word Score, from 0 to 10, for Secondary Word matches. The default is 10.

Repeating fields

Repeating fields are distinct fields which repeat n times in the search and file records being passed to Matching. Examples might be where you want to match a search name against either a current name or a former name and get the best Score. Another example might be to match a search address against either a residential address or a postal address.

To enable this feature, use the

REPEAT=

option on the

FIELD

keyword as described in the

Definition File Structure

section at the beginning of this chapter. For example,

METHOD NAME=MNAME,WEIGHT=1,              X
            GOPT=(LENGTH*50+REFMIN),     X
            LOPT=(CONC+CINITA+INITLOW)
FIELD OFFSET=0,REPEAT=2

This tells the N3SCM Method to expect two (2) fields of length 50 in both the search record and file record. For example,

Search: John Smith
File:   Mike Taylor                 John Smith

would return a Score of 100, due to the match of the search name John Smith against the second of the two fields in the file record. You could also put two names in the search record and the method would try up to four matches before returning with a Score, although if any one of those matches scores 100, it returns early.

To modify this behavior the following Local Options may be used.

Option	Description
OPTION FLAGS VALUE WORSTSCR,{0/1/2} 0 - return the highest score detected (default) 1 - return the worst "best" score detected. 2 - return the worst "worst" score detected.	In the descriptions below, the search and file records are assumed to contain the following: The search record contains 3 fields in a repeating group: John Smith Mary Smith Peter Smith And the file record contains 4 fields in a repeating group: John Smith Mary Smith Peter Smith Paul Jones WORSTSCR=0 Each field in the search record is compared to every field in the file record, always remembering the highest score. This process continues until either all possible comparisons have been performed or a score of 100 is detected. The highest score is then returned. Using the sample data, John Smith will score 100 and this score will be returned. No further comparisons will be necessary. WORSTSCR=1 Each field in the search record is compared to every field in the file record, remembering the highest score calculated for each search field. Then, the lowest remembered score is returned. Using the sample data, each of the fields in the search record will score 100. The lowest of these scores, 100, will then be returned. WORSTSCR=2 Each field in the search record is compared to every field in the file record, remembering the lowest score calculated for each search field. Then, the lowest remembered score is returned. Using the sample data, each of the fields in the search record will score 0. The lowest of these scores, 0, will then be returned.
OPTION FLAGS VALUE WSCRNOEX,[number]	This option only has an effect when WORSTSCR is set to 1 or 2. It is used to penalize the score in the case where the number of repeats differs between the search and file records. The score is reduced by this value for each extra field present. So, if WSCRNOEX is set to 3 and using the sample data above, the score will be reduced by 3 (3 * (4 - 3)). If specified, it should be in the range 1-10. The default is 0, which means that no reduction in score occurs.

Local Options Controlling Word Score

OPTION CATSW
VALUE NN,9

RICK THOMSON
RICHARD THOMSON

OPTION CATSW
VALUE NN,95

RICK THOMSON
RICHARD THOMSON

Local Options Controlling Reference Record Matching

   GOPT=(REFMIN)

SEARCH: JOHN PEEL
FILE: KEN JOHN PEEL
SCORE: 100

   GOPT=(REFMAX)

SEARCH: JOHN PEEL
FILE: KEN JOHN PEEL
SCORE: 066

   GOPT=(REFMIN)
   OPTION NOEXCESS
   VALUE TRIGGER,95
   VALUE PENALTY,5

SEARCH: JOHN PEEL
FILE: KEN JOHN PEEL
SCORE: 095

10 VALLEY RD
10

    GOPT=(REFMIN)
    OPTION NOEXCESS
    VALUE TRIGGER,95
    VALUE PENALTY,5

SEARCH: JOHN
FILE: KEN JOHN PEEL
SCORE: 090

    GOPT=(REFMIN)
    OPTION NOEXCESS
    VALUE TRIGGER,95
    VALUE PENALTY,5
    VALUE REFCNT,1
    VALUE REFMULT,2
    VALUE REFS,1

SEARCH: JOHN
FILE: KEN JOHN PEEL
SCORE: 080

    GOPT=(REFMIN)
    OPTION NOEXCESS
    VALUE TRIGGER,95
    VALUE PENALTY,5

SEARCH: MOHAMMED SAID
FILE: SA-ID
SCORE: 100

      OPTION NOEXCLST
      VALUE YM,0

SEARCH: MOHAMMED SAID
FILE: SA-ID
SCORE: 095

OPTION REFN
VALUE WORDS,2
VALUE GOOD,8
VALUE SCORE,95

JOHN MICHAEL THOMPSON
JOHN CHRISTOPHER THOMPSON

Local Options Addressing Matching of Codes

OPTION CODESCOR
VALUE CODEWGHT,<codeweight>
VALUE CODEUDIF,25
VALUE CODEUNON,90
VALUE CODEUONE,50
VALUE CODEMAXD,6
VALUE CODEPOSS,1
OPTION SORTSCOR
VALUE CLN,<clnweight>
VALUE FMT,<fmtweight>
VALUE SORTWGHT,<sortweight>
VALUE NGRAMC,<weight>
VALUE NGRAMCLV,<level>
VALUE NGRAMF,<weight>
VALUE NGRAMFLV,<level>

CODEWGHT
must be set to 0 - 100 (0 means
CODESCOR
is disabled and 100 means that the other scoring calculations will be ignored. Default value is 0.
Both
CODESCOR
and
EXCTCODE
cannot be used at the same time. The method will exit with error if these conflicting options have been specified.
CODESCOR
does not make sense if
FORMATTING-OPTIONS #1
is set to ’D’ (delete codes). The method will exit with error if these conflicting options have been specified.
The sum of the specified
CODEWGHT
and
SORTWGHT
must not exceed 100.
If
SORTWGHT > 0
then calculate
SORTSCOR
score.
If
CODEWGHT > 0
and if either search or file name contains one or more codes, then calculate
CODESCOR
score.

Normal Score Weight = 100 - (Sort Score Weight + Code Score Weight)
Final Score =  ((Sort Score * Sort Score Weight)
               + (Code Score * Code Score Weight)
               + (Normal Score * Normal Score Weight))
               / 100)

Normal Score Weight = 100 - Sort Score Weight
Final Score =     ((sortScore * Sort Score Weight)
                  + (normalScore * Normal Score Weight))
                  / (100 - Code Score Weight));

Code score calculation

Code Score = 100
            - (100 * Number of Codes in Stack
              / Number of Entries in the Stack That Has Codes)
if (Number of Entries in the Stack That Does Not Have Codes <
    Number of Entries in the Stack That Has Codes) then
     Weight = 100
              * Number of Entries in the Stack That Does Not Have Codes
              / Number of Entries in the Stack That Has Codes
     Code Score = Code Score * Weight / 100
endif

Comparing codes that consist of or begin with digits

"G "     grams
"L "     litres
"M "     metres
"KG "    kilograms
"MG "    milligrams
"KL "    kilolitres
"ML "    millilitres
"CL "    centilitres
"DL "    desilitres
"MM "    millimetres
"CM "    centimetres

Intermediate Score = 100 * Smaller Number / Larger Number

if both numbers are followed by a unit of measure, then:
if the units match exactly, then no intermediate score weighting is done
otherwise weight the intermediate score with 25%; eg. "400 oranges" scored against "600 apples" produces score of 25 * 66 / 100 = 16.
if only one number is followed by a unit of measure, then:
weight the intermediate score with 50%; eg. "400 oranges" scored against "600" produces score of
50 * 66 / 100 = 33.
if neither number is followed by a unit of measure, then:
weight the intermediate score with 90%; eg. "400" scored against "600" produces score of 90 * 66 /
100 = 59.

VALUE CODEUDIF,25
VALUE CODEUNON,90
VALUE CODEUONE,50

Comparing codes that consist of (or begin with) large numbers

VALUE CODEMAXD,6

CODEMAXD

(default value 6) specifies the maximum number of digits in a code for which a numeric comparison is allowed to take place. When the number of digits exceeds this limit, a string comparison is performed instead. If a numeric comparison should always take place then set this value to a large value (at least 24 which is the maximum size of an entry in the formatted word stack and therefore the maximum length of a code).

VALUE CODEPOSS,1

CODEPOSS

(default value 1) specifies that this string comparison is position sensitive, eg. the first digit in the first code is matched only against the first position in the second code etc. Setting this value to 0 specifies that synchronization is allowed to take place. This synchronization follows the same rules as described in the section Comparing codes that do not begin with digits below. Other values are reserved for future use.

The position sensitive score is calculated as follows:

First count matching characters in codes until the end of one code is reached.

If the lengths of the codes are equal, then the score is calculated as follows:

score = Number of Matching Characters * 100 /
                Total Number of Characters in the Code

If the lengths of the codes differ, then the score is weighted down relative to the difference between the code lengths. The formula is as follows:

score = (Number of Matching Characters *
                            100 * Length of Shorter Code) /
                (Length of Longer Code * Length of Longer Code)

Comparing codes that do not begin with digits

if (reflen <= Number of Matching Characters) then
       score = 100
else
       score = 100 * Number of Matching Characters / reflen
endif

OPTION SORTSCOR

When

SORTSCOR

has been specified, the method tokenizes the search and file names, sorts these tokens and performs additional scoring on these sorted tokens. This happens after all the usual exact match checking etc. usually performed by N3SCM. The logic for sorted scoring is explained below.

Scoring Cleaned Strings ("CLN Score")

VALUE CLN,<clnweight>

If

CLN

has been set to a value greater than 0, each cleaned input name is sorted in byte order ignoring any spaces.

UNICODE support has not been implemented). For example, if the search name is

"AC/DC          "

and the file name is

"EDC...BA       "

then these names are first cleaned and the resulting names are sorted, resulting in

"ACCD"

(assuming that the used cleaning routine gets rid of the ’/’)

and

"ABCDE"

(assuming that the used cleaning routine gets rid of the ’...’)

The resulting pair is compared byte by byte in a synchronised manner. In the above example the following comparisons happen:

A vs A

--> match, skip to the next character in each input string

C vs B

-->does not match, skip to the next character in the second string (as

C > B

)

C vs C

-->match, skip to the next character in each input string

C vs D

-->does not match, skip to the next character in the first string (as

C < D

)

D vs D

-->match, skip to the next character in each input string

at this point the end of the first string is detected (spaces are ignored) and the matching stops. 3 matches were counted.

Calculate the sorted score for the cleaned string as follows:

(number of matching characters) * 100
     / (Length of reference string)

if

REFMAX

has been specified, then use the longer string as reference. In the above example length of

ABCDE

(5), resulting to,

3 * 100 / 5 = 60

if

REFMIN

has been specified, then use the shorter string as reference. In the above example length of

ACCD

(4), resulting to,

3 * 100 / 4 = 75

Scoring Formatted Word Stacks ("FMT Score")

VALUE FMT,<fmtweight>

If

FMT

has been specified (greater than 0), the formatted word stacks are sorted. For example, if the search name word stack is.

"01 DC          "
"02 AC          "

and the file name word stack is

"01 BC        "
"02 AC        "
"03 DC        "

then these stacks are first sorted, results being

"AC        "
"DC        "

and

"AC       "
"BC       "
"DC       "

Then the sorted stacks are compared in a synchronised manner. In the above example the following comparisons happen:

AC vs AC -->

match, skip to the next entry in each stack

DC vs BC -->

does not match, skip to the next entry in the second stack (as

DC > BC

)

DC vs DC -->

match, skip to the next entry in each stack

at this point the end of the first stack is detected and the matching stops. 2 matches were counted.

Calculate the sorted score for the formatted stacks as follows:

(Number of Matching Stack Entries) * 100
   / (Total Number of Entries in the Reference Stack)

if

REFMAX

, then use the larger stack as reference. In the above example the larger stack has 3 entries, resulting to,

2 * 100 / 3 = 66

if

REFMIN

, then use the smaller stack as reference. In the above example the smaller stack has 2 entries, resulting to,

2 * 100 / 2 = 100

Combining CLN And FMT Scores

Score = ((CLN Score * <clnweight>) + (FMT Score * <fmtweight>)) / 100)

Score = ((75 * 75) + (100 * 25)) / 100

Score = ((60 * 75) + (66 * 25)) / 100

Additional Scoring (if any)

VALUE SORTWGHT,<sortweight>

Finally the

SORTWGHT

is processed. The value of

<sortweight>

must be a number between 1 to 100 or an error is issued. If 100 is specified, then the above calculated combined

CLN

and

FMT

Score is the final result. But if

<sortweight>

is less than 100, then all usual score calculation done by N3SCM is performed ("normal score"). After this, the scores are weighted using the following formula:

Final Score = ((Combined Sort Score * Sort Weight)
                  + (Normal Score * (100 - Sort Weight))) / 100

(Remainder – if any – is dropped.)

For example, if the combined sort score was 81 (the above

REFMIN

example),

<sortweight>

was 35 and the normal score was 90, then the final score would be calculated as follows:

Final Score = ((81 * 35) + (90 * (100 - 35))) / 100

which results to 86.

NGRAM Scoring

OPTION SORTSCOR also recognises the following parameters:

VALUE NGRAMC,<weight>
VALUE NGRAMCLV,<level>
VALUE NGRAMF,<weight>
VALUE NGRAMFLV,<level>

NGRAM scoring for cleaned input strings ("NGRAMC Score")

NGRAM

scoring for cleaned input strings is controlled with parameters

NGRAMC,<weight>

and

NGRAMCLV,<level>

(both or neither must be specified).

The cleaned input strings are split into tokens, the length of which is specified with the parameter NGRAMCLV,<level> These tokens are scored as explained in the

OPTION SORTSCOR,VALUE CLN

section above, with the exception that the length of each token is not 1, but is specified with the

<level>

parameter.

If the

<level>

parameter is 2, then 2-grams are scored, if

<level>

is 3, then 3-grams are scored etc. The achieved

NGRAMC

Score is weighted with the value specified with the

NGRAMC,<weight>

parameter.

NGRAM scoring for formatted word stacks ("NGRAMF Score")

VALUE CLN,<clnweight>
VALUE FMT,<fmtweight>
VALUE NGRAMC,<ngramcweight>
VALUE NGRAMF,<ngramfweight>

Score = ((CLN Score * <clnweight>)
        + FMT Score * <fmtweight>)
        + (NGRAMC Score * <ngramCweight>)
        + (NGRAMF Score * <ngramFweight>)) / 100)

N3SCM Default Options

OPTION CLIMIT
VALUE CHARDS,100
VALUE NSACTF,0
VALUE TRIGS,100
VALUE NOINCR,0
VALUE AVERAGE,0
VALUE NOEXPNTY,0

OPTION CONCAT
VALUE PLURALS,0

OPTION FLAGS
VALUE ENABLDNM,0
VALUE EXACTCAT,0
VALUE EXACTINI,1
VALUE EXACTWRD,1
VALUE INITCODE,1
VALUE LIMWCAT,1
VALUE MATCHEND,1
VALUE OPTIMILW,1
VALUE ORIGCSW,0
VALUE SECOND,0
VALUE SYNCS,2
VALUE TRANSLEN,1
VALUE USECWAIT,0
VALUE SKIPMAJM,0
VALUE SKIPSMOD,0
VALUE SCALEFTR,10
VALUE CATSWEXT,0
VALUE CATSWD,0
VALUE CATSWF,0

OPTION NOEXCESS
VALUE PENALTY,1
VALUE TRIGGER,101
VALUE REFCNT,1
VALUE REFMULT,0
VALUE REFS,1
VALUE REFF,0
VALUE SREFCNT,1
VALUE SREFMULT,0

OPTION ORDER (ORDER options are disabled by default)
VALUE POS,9999
VALUE SEQ,9999
VALUE TRIGGER,9999

OPTION REFN
VALUE GOOD,9
VALUE SCORE,99
VALUE WORDS,0

OPTION SCORES
VALUE ILOWTRIG,10
VALUE SEC,10

OPTION CODESCOR
VALUE CODEWGHT,0

OPTION SORTSCOR
VALUE SORTWGHT,0