My Top 10 TCIs

[CarsonWid]

1. Topaz - 1.3640 (228 Films in Common)
2. Akikaze - 1.4207 (290)
3. MALG - 1.4400 (150)
4. tendbundy - 1.4530 (118)
5. Moonrooster - 1.4554 (213)
6. Randoom - 1.4556 (439)
7. sheeper - 1.4835 (395)
8. RotciV - 1.4854 (171)
9. joske1230 - 1.4942 (431)
10. Night - 1.5051 (396)

[MmzHrrdb]

#1 1.3708 ozufan
#2 1.4715 hungrybear
#3 1.4754 alfredlee
#4 1.5370 DirktheJerk
#5 1.5591 Blevo
#6 1.5777 Lone Wolf
#7 1.5959 Llamadeus
#8 1.5962 egi_taltos
#9 1.5972 FakeAntonyo
#10 1.6038 Kerc

[ShogunRua]

I used to not have any TCL higher than 2.00, but then I watched a number of classic movies, and changed my number of films in common to a freakishly low 3%;

1. TCL 1.6739 1qw3er5 138 Films in Common
2. TCL 1.8000 Gnawdude 50 Films in Common
3. TCL 1.8610 naikou 187 Films in Common
4. TCL 1.9369 Stabwound 111 Films in Common
5. TCL 1.9792 Razzll 96 Films in Common
6. TCL 1.9821 lena 56 Films in Common.
7. TCL 1.9848 wanders 66 Films in Common
8. TCL 1.9894 nothing 94 Films in Common
9. TCL 2.0000 Shafour 53 Films in Common
10. TCL 2.0339 fa.thor 59 Films in Common

[djross]

x

[shalev]

#1 imdb_ws TCI: 1.3366 (101)
#2 Dingoes TCI: 1.3630 (136)
#3 popinio TCI: 1.3729 (129)
#4 UcanPenguen TCI: 1.4416 (154)
#5 Critic2010 TCI: 1.4420 (138)
#6 gs_undertow TCI: 1.4537 (118)
#7 nickhaag TCI: 1.4661 (118)
#8 Forilas TCI: 1.4688 (96)
#9 Antikroppar TCI: 1.4873 (158)
#10 sirius TCI: 1.5000 (92)

My minimum is 12% (86)
Only 2 are recently active. (sirius and Forilas)
I wonder about the imdb_ws user, seems like the rankings are from imdb, mostly more popular films - 266 films ranked on a period of 2 weeks. The location reads Bulgaria. Huh?
It's strange, cause my TCI with IMDB is always around 2.1.
What does it say, that 2 'users' with basically the same 'taste' have such a different TCI?

Edit: Oh, seems like popinio is also just a selection of rankings from IMDB, posted over a week. What is going on here?

[TonythePony]

What does it say, that 2 'users' with basically the same 'taste' have such a different TCI?

It's more about the high proportion of low quality obscure movies that have been added. This does not correspond to the way your average person selects movies.

When the imdb user had about 5000 rankings my TCI was 1.10. That placed it in my top ten.
With the 34000 rankings my TCI is over 2 (have it ignored so not sure exactly). My top 1000 ends at 1.77.

By adding in every movie possible the account has been rendered useless for recommendations.

[djross]

x

[TonythePony]

Not useless for everybody, though. My TCI for IMDb may only be 2.3232, but this still equates to 390th place, which means that for many movies, the IMDb rating still counts.

Fair enough. It's not completely useless for recommendations, just much less useful than it could be.

What exactly is supposed to really be measured by the IMDb score may indeed be rather obscure, but there are still many movies that few or very few Criticker users have ranked, so in those cases IMDb can at least add one (collective) voice.

In that sense it is worthwhile as a reference rather than a source of recommendations.

[djross]

x

[TonythePony]

In that sense it is worthwhile as a reference rather than a source of recommendations.

I don't understand the difference: if user IMDb is within one's 1000 top TCIs, and within the top ten TCIs for a particular film, then how is this user any different from any other user who falls within those two conditions?

What I meant by reference was that it shows the score on IMDb without having to go to the site. For recommendations the score is converted to a percentile and that is where the problem creeps in.

An extreme example of the skew in the imdb user rankings is the "theyshootpic" accounts here at criticker. They provide a list of the 1000 greatest movies with 250 other rankings added to attempt to provide balance. The greatest 1000 movies of all time should be tier 10 or at worst tier 9. Instead the lowest on the 1000 become tier 3.

You might assume that those lists would have an extremely high TCI, but when most of the "films in common" are at the top of the list the TCI's are quite close. Three or four of those lists were in my top 1000 TCI's before I ignored them. So a lowish TCI does not always mean an accurate recommendation.

Another example. Lets assume I locked you away and forced you to watch and rate 2000 movies which would normally be tier 1 for you. Adding those to a new account ("Tortured_djross" perhaps ) with your old rankings would create a user with a vastly different TCI from your original despite tastes being exactly the same. This is sort of similar to the imdb user.

Apologies for the lack of clarity there.

No, I think you were clear enough. I can't say the same for myself.

What I'm trying to say amidst my stream of waffle is that the imdb user, despite best intentions, is an inaccurate representation of the imdb site rankings.