lucene-java-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Doug Cutting <>
Subject Re: exact the same score from different documents
Date Fri, 14 May 2004 17:34:54 GMT
hui wrote:
> I am getting the exactly same score like 0. 04809519 for different size
> documents for some queries and this happens quite frequently. Based on the
> score formula, it seems this should rarely happen. Or I misunderstand the
> formula?

Normalization factors (& document boosts) are represented in the index 
using a one-byte float format with a 3-bit mantissa, which means that 
differences of plus-or-minus 1/8 are rounded to a single value.

For example, a field with 256 tokens by default has a lengthNorm() of 
16.0.  With a three-bit mantissa, values 16.0 to 18.0 are rounded to the 
same value, which means that fields with between 256 and 324 tokens will 
have the same effective length normalization.


To unsubscribe, e-mail:
For additional commands, e-mail:

View raw message