Bug: Frequency of ὁμολογέω in 1 John per 1000 words

Page 1 of 1 (13 items)
This post has 12 Replies | 4 Followers

Posts 8
Bill & Janet Curd | Forum Activity | Posted: Mon, Apr 29 2019 8:26 PM

On 2 different family computers in Logos 8 I look at "confess" in (NASB 1995) 1 John 1:9, lemma ὁμολογέω, Bible Word Study. Specifically, I look at ὁμολογέω in NASB95 | Number of Results in Book / 1000 Words in Book.

As best I can tell I am doing exactly the samething on both computers.Both computers correctly show 5 occurrences of ὁμολογέω in 1 Jn, but one computer reports 0.98 and the other reports 1.96 occurrences per 1000 words in 1 Jn.

Any ideas?

Posts 26458
Forum MVP
MJ. Smith | Forum Activity | Replied: Tue, Apr 30 2019 12:00 AM

I would need screen shots (attach via the paper clip icon) to see what the difference is in your queries.

Orthodox Bishop Hilarion Alfeyev: "To be a theologian means to have experience of a personal encounter with God through prayer and worship."

Posts 1012
Harry Hahne | Forum Activity | Replied: Tue, Apr 30 2019 1:07 AM

William Curd:
Both computers correctly show 5 occurrences of ὁμολογέω in 1 Jn, but one computer reports 0.98 and the other reports 1.96 occurrences per 1000 words in 1 Jn.

Is there any chance that you are searching a different Bible on the other computer?

If I do this search on the NASB 1995, the number of results / 1000 words is 0.98. For the NIV, I get 1.98. On the NA28 Greek New Testament, I get 2.34.

Of course, the most meaningful statistic would be in the Greek New Testament, if you are interested in the relative frequency of a Greek lemma. For English Bibles, the number would vary because the translators may choose to translate a single Greek word with a phrase rather than a single word.

Posts 10029
Forum MVP
NB.Mick | Forum Activity | Replied: Tue, Apr 30 2019 2:13 AM

Harry Hahne:
If I do this search on the NASB 1995, the number of results / 1000 words is 0.98.

Hm. Sometimes 0.98, sometimes 1.96 for me - and I even get it to display both results next to each other:

This is definitely a bug - one of the displays is wrong. Maybe it has a counter for occurrences that is not reset to zero and thus counts twice on the second run?

Running Logos 8 latest beta version on Win 10

Posts 10029
Forum MVP
NB.Mick | Forum Activity | Replied: Tue, Apr 30 2019 2:14 AM

Harry Hahne:

Of course, the most meaningful statistic would be in the Greek New Testament, if you are interested in the relative frequency of a Greek lemma. For English Bibles, the number would vary because the translators may choose to translate a single Greek word with a phrase rather than a single word.

possibly the only meaningful statistic here.

EDIT:

Since the question is there, even if it is not meaningful, it would be interesting to see which of both figures is correct. I ran a concordance on 1 Jn in NASB - the surface text has 2,553 bible words, i.e. the five occurrences of the Greek lemma will give a relative frequency of 1.96 per 1,000 words in 1 Jn. This seems to explain the 1.96 calculated ratio, the 0.98 seems wrong.

Running Logos 8 latest beta version on Win 10

Posts 1012
Harry Hahne | Forum Activity | Replied: Tue, Apr 30 2019 12:43 PM

NB.Mick:
This is definitely a bug

I can confirm that there is a bug causing inconsistent statistics for Number of Results in Book / 1000 Words in Book.

This is a very significant bug, because a scholar publishing research conclusions based on incorrect statistics would at best be embarrassed but at worst would be discredited.

I did a search on the NA28 for ὁμολογέω and generated the graph for Number of Results in Book / 1000 Words in Book. When I did this twice I got 1.17 one time and 2.34 the second time. Sadly, when turned on logging Logos decided to behave and be consistent, so I cannot show a log with the 2 outcomes. However, Logos is reporting half of the correct number.

The concordance tool reports 2140 Greek lemmas in the NA28 Greek New Testament. There are 5 occurrences of ὁμολογέω in 1 John. Therefore, in the NA28 the correct Number of Results in Book / 1000 Words in Book should be 2.3364 (or rounded to 2.34). However, the graph shows 1.17, which is exactly half of the correct number.

I am using Logos 8.5 Beta 1. I have attached a log for troubleshooting.

4604.Logos Log Files - Hahne -04-30-19 (1219).zip

Posts 10029
Forum MVP
NB.Mick | Forum Activity | Replied: Tue, Apr 30 2019 4:10 PM

I assume the OP runs 8.4 stable, but since we both see the bug on beta (and the beta forum is more closely monitored regarding bugs), I crossposted at https://community.logos.com/forums/t/181064.aspx to bring this to FL's attention.  

Running Logos 8 latest beta version on Win 10

Posts 24919
Forum MVP
Dave Hooton | Forum Activity | Replied: Tue, Apr 30 2019 7:00 PM

I couldn't reproduce the above results in stable 8.4 with NA28, no matter how I generated the Graph e.g. from Results on 1 Jn and All Passages. My initial results in 8.5 Beta 1 were correct, but the second Graph from the same Search Results showed half the expected frequency.

Dave
===

Windows & Android

Posts 2
Lance Dare | Forum Activity | Replied: Thu, May 2 2019 6:09 PM

I am getting the correct result if I search the NASB, but in searching the Greek New Testament I am getting the "half answer".

Posts 24919
Forum MVP
Dave Hooton | Forum Activity | Replied: Fri, May 3 2019 4:45 PM

Bump for Faithlife.

Dave
===

Windows & Android

Posts 1826
LogosEmployee
Phil Gons (Faithlife) | Forum Activity | Replied: Fri, May 3 2019 7:34 PM

Got it. I added it to our list of important academic bugs to fix in Q2/Q3.

Posts 8006
LogosEmployee
Bradley Grainger (Faithlife) | Forum Activity | Replied: Tue, May 14 2019 11:01 AM

NB.Mick:
Sometimes 0.98, sometimes 1.96 for me - and I even get it to display both results next to each other:

We fixed a bug that could cause the results to be consistently good, or consistently bad (half the expected number) on a per-machine basis (what the OP reported).

However, we were not able to reproduce variation in the results on a single machine, especially not side-by-side. NB.Mick, please test again with 8.5 Beta 5 and let us know if you can reproduce the behaviour you reported above.

Posts 10029
Forum MVP
NB.Mick | Forum Activity | Replied: Tue, May 14 2019 12:32 PM

Bradley Grainger (Faithlife):
NB.Mick, please test again with 8.5 Beta 5 and let us know if you can reproduce the behaviour you reported above.

I just did - I couldn't get Logos 8.5 Beta 5 produce a wrong number in the results/1000 words in book chart. Apparently, it's good now, thanks!

Running Logos 8 latest beta version on Win 10

Page 1 of 1 (13 items) | RSS