Welcome to the ICM Forum. If you have an account but have trouble logging in, or have other questions, see THIS THREAD.
Polls: Favorite Movies (Results), 1945 (Results), 1929 awards (Apr 4th), South Asia (Apr 25th), Doubling the Canon (Ratings Apr 30th)
Challenges: Doubling the Canon, Nordic, 1950s
Film of the Week: Valkoinen peura, May nominations (May 1st)
World Cup S4: Round 1 schedule, 1F: Brazil vs Greece vs Japan vs Poland (Apr 5th), 1G: Germany vs Pakistan vs Ukraine vs USA (Apr 22nd)

iCM Favorite Underrated Movies - Discussion topic

500<400, Favourite 1001 movies, Doubling the Canon, Film World Cup and many other votes
User avatar
albajos
Posts: 6821
Joined: May 24, 2016
Location: Norway
Contact:

Re: iCM Favorite Underrated Movies - Discussion topic

#41

Post by albajos » November 22nd, 2019, 7:57 pm

sol wrote:
November 22nd, 2019, 2:31 pm
I like the idea of having a minimum 400 checks criteria. Not only would that reduce overlap with the 500<400 list, it kind of makes sense since it is hard to judge whether something is really under-rated or not if few out there have seen it.
95% of all norwegian movies are below 400 checks

If anything go with 1000 votes which is a regular size of a statistic group

User avatar
sol
Donator
Posts: 7844
Joined: Feb 03, 2017
Location: Perth, WA, Australia
Contact:

#42

Post by sol » November 23rd, 2019, 2:51 am

albajos wrote:
November 22nd, 2019, 7:57 pm
sol wrote:
November 22nd, 2019, 2:31 pm
I like the idea of having a minimum 400 checks criteria. Not only would that reduce overlap with the 500<400 list, it kind of makes sense since it is hard to judge whether something is really under-rated or not if few out there have seen it.
95% of all norwegian movies are below 400 checks

If anything go with 1000 votes which is a regular size of a statistic group
Yeah, I'm cool with a 1000 check minimum too. As long as there is a certain minimum, I think the results have a chance of being meaningful.
Former IMDb message boards user // iCM | IMDb | Letterboxd | My top 600 films // Long live the new flesh!
Image Image Image

User avatar
Fergenaprido
Donator
Posts: 3669
Joined: Jun 03, 2014
Location: Malaysia
Contact:

#43

Post by Fergenaprido » November 23rd, 2019, 3:31 am

Ivan0716 wrote:
November 22nd, 2019, 7:03 pm
Fergenaprido wrote:
November 22nd, 2019, 2:26 pm
Sure, ToL may be rated highly by a lot of people, but it's also rated lowly by a lot of people. I wouldn't have a problem if a film like ToL showed up in the final list, because it's sounds exactly like the type of divisive film to be mentioned.
Tree of Life is one of the most acclaimed films of this decade, and has been voted as one of the best American films ever, any attempts to call it underrated is laughable. I don't think anyone will actually vote for it, but you can't really justify its inclusion by saying "Yeah but a bunch of people hate it too": every film has its own share of detractors.

-

An underrated film shouldn't just be something you rate higher than everyone, it should be films that hasn't gotten the attention you think it deserves; one where its qualities and values have been unfairly overlooked by the "whole" rather than a conveniently selected section of the community. It's not always easy to justify, but it should go beyond "because I/my ratings say so". If something is widely regarded as one of the best - like Tree of Life or Salo - then it simply can't be called underrated.

Just to be clear, I'm not against the poll or any of the methods proposed so far, I just think it's worthwhile to discuss why this is such a flawed concept.
On the flip side, I would say what you're saying is slightly contradictory; it sounds like your case is that ToL shouldn't be considered overlooked because a conveniently selected section of the community appreciates it, when it's had a rather lukewarm reception by the "whole". If we exclude critically acclaimed films that have some pockets of love somewhere, then we effectively exclude all films except for the most widely panned, and this turns into a "Favourite bad films" or "Favourite films from the imdb Bottom 100". I don't think anyone is really interested in lists of those sort.

In all of our other polls "because I/my ratings say so" is good enough criteria... why should this poll be any different? Ratings are essentially just a proxy for how much we like a film, and are used as a guideline to help determine what we each individually think is underrated. No one is asked to justify why they include any particular film in their list (unless it's a questionable genre-inclusion or they're PdA tehe ), so I'm curious as to why you think people need to justify their selections for this list. I did a brief scan through my rated feature films yesterday and found about 35 films that I've rated at least 2.0 points higher than the imdb average (on the 1-10 scale). I wouldn't consider all of them underrated, and there are other films that fall into the 1.5-1.9 difference range that I would be more likely to consider underrated that some of the 2.0+ ones.

Tree of Life and Salo may be widely regarded by a highly visible section of the film community (namely critics), but they don't seem to have connected with film audiences in general. Therefore, if you like those films, it makes sense if you consider them underrated relative to the general film community perceives them.
sol wrote:
November 23rd, 2019, 2:51 am
albajos wrote:
November 22nd, 2019, 7:57 pm
sol wrote:
November 22nd, 2019, 2:31 pm
I like the idea of having a minimum 400 checks criteria. Not only would that reduce overlap with the 500<400 list, it kind of makes sense since it is hard to judge whether something is really under-rated or not if few out there have seen it.
95% of all norwegian movies are below 400 checks

If anything go with 1000 votes which is a regular size of a statistic group
Yeah, I'm cool with a 1000 check minimum too. As long as there is a certain minimum, I think the results have a chance of being meaningful.
I'm not really in favour of a minimum "seen" criterion, but if we are going to have one I'd prefer it to be 500 or 1000 votes on imdb, instead of the number of checks on icm. There are plenty of films with a couple of thousand votes on imdb but still less than 1000 checks on icm. I also think it makes sense that we keep the criteria limited to one website; if we want to use checks on icm then I would suggest we use icm favourite % instead of imdb ratings (but I think that's even more problematic, hence why I prefer imdb).

User avatar
Ivan0716
Posts: 1200
Joined: Feb 05, 2012
Contact:

#44

Post by Ivan0716 » November 23rd, 2019, 6:33 am

Fergenaprido wrote:
November 23rd, 2019, 3:31 am
On the flip side, I would say what you're saying is slightly contradictory; it sounds like your case is that ToL shouldn't be considered overlooked because a conveniently selected section of the community appreciates it, when it's had a rather lukewarm reception by the "whole".
I think you're really underselling The Tree of Life to make this point. I might even argue that it's a well appreciated film on the whole, but had gotten a rather lukewarm reception by a conveniently selected section of the community. Hey, see what I did there? :rolleyes:
If we exclude critically acclaimed films that have some pockets of love somewhere, then we effectively exclude all films except for the most widely panned, and this turns into a "Favourite bad films" or "Favourite films from the imdb Bottom 100". I don't think anyone is really interested in lists of those sort.
I'm sure you can find a few films that falls in between "critically acclaimed" and "widely panned".
In all of our other polls "because I/my ratings say so" is good enough criteria... why should this poll be any different? Ratings are essentially just a proxy for how much we like a film, and are used as a guideline to help determine what we each individually think is underrated.
We've never had to apply ratings in this way for other polls, so not sure where they come into this. All I said was I don't consider a film to be underrated simply because my ratings are higher than those on IMDb: there are films I've rated highly even though I know they're dogshit; there are also films I've rated lower than IMDb that I would consider to be underrated. It's not a criticism on the solutions proposed, but on the subject/nature of the poll.

No one is asked to justify why they include any particular film in their list (unless it's a questionable genre-inclusion or they're PdA tehe ), so I'm curious as to why you think people need to justify their selections for this list. I did a brief scan through my rated feature films yesterday and found about 35 films that I've rated at least 2.0 points higher than the imdb average (on the 1-10 scale). I wouldn't consider all of them underrated, and there are other films that fall into the 1.5-1.9 difference range that I would be more likely to consider underrated that some of the 2.0+ ones.
Nor am I asking anyone to justify their selection to other people. But I'm sure when they construct their lists they'll think about whether or not a film actually justifies the "underrated" tag...like you just did.
Tree of Life and Salo may be widely regarded by a highly visible section of the film community (namely critics), but they don't seem to have connected with film audiences in general. Therefore, if you like those films, it makes sense if you consider them underrated relative to the general film community perceives them.
I disagree.

User avatar
Onderhond
Posts: 3532
Joined: Dec 23, 2012
Contact:

#45

Post by Onderhond » November 23rd, 2019, 8:56 am

Ivan0716 wrote:
November 23rd, 2019, 6:33 am
I think you're really underselling The Tree of Life to make this point. I might even argue that it's a well appreciated film on the whole, but had gotten a rather lukewarm reception by a conveniently selected section of the community. Hey, see what I did there? :rolleyes:
6.8/10 with 158,937 votes on IMDb.

No discussion. See what I did there ;)
Unless you want to use Letterboxd as a reference of course.

Weighted average of 3.79 based on 55,542 ratings.

Still no discussion. Just pick a source, the rest is easy.
Last edited by Onderhond on November 23rd, 2019, 8:59 am, edited 3 times in total.

User avatar
Fergenaprido
Donator
Posts: 3669
Joined: Jun 03, 2014
Location: Malaysia
Contact:

#46

Post by Fergenaprido » November 23rd, 2019, 8:56 am

Agree to disagree, then, Ivan. :)
Thanks for replying, though. I don't think there's much more I can say to your points right now.

But maybe I'll get around to watching Tree of Life just to see if it would make my ballot for this poll :D

User avatar
flaiky
Posts: 1463
Joined: Feb 04, 2017
Location: London UK
Contact:

#47

Post by flaiky » November 23rd, 2019, 7:15 pm

I would vote for "Less than 7 on IMDB, plus not on TSPDP or 500>400". Anything with enough critical support to make TSDPT shouldn't be called underrated, but I agree that it's worth avoiding too much crossover with our gems list.

That said, overall I'd lean towards skipping the poll as it already seems controversial and divisive. I'm not sure how meaningful the results could be. Putting a list together also sounds like a nightmare.
Let the ashes fly
ICM | Letterboxd | All-time stats

User avatar
RogerTheMovieManiac88
Posts: 1541
Joined: Feb 04, 2017
Location: Westmeath, Ireland
Contact:

#48

Post by RogerTheMovieManiac88 » November 23rd, 2019, 11:52 pm

flaiky wrote:
November 23rd, 2019, 7:15 pm
I would vote for "Less than 7 on IMDB, plus not on TSPDP or 500>400". Anything with enough critical support to make TSDPT shouldn't be called underrated, but I agree that it's worth avoiding too much crossover with our gems list.
I pretty much agree with this. I'd probably lower the upper limit to around 6.5 on IMDb though.

This is a poll I'm rather keen on so I hope it's not scrapped.
That's all, folks!

User avatar
PeacefulAnarchy
Moderator
Posts: 24188
Joined: May 08, 2011
Contact:

#49

Post by PeacefulAnarchy » November 24th, 2019, 4:53 am

Underrated:
Must meet all three of the following criteria:
  • Under 7.0 on imdb (the average is 6.8/6.9 so this means not particularly well though of by the public)
  • More than 400x(#of official lists) Checks (So 0 official lists, no check limit, 1 official list - must be over 400 checks, 2 official lists - must be over 800 checks, etc) This removes critically well regarded but otherwise obscure films.
  • No more than 5 Official lists, at that point it's critically well regarded (this eliminates around 1600 films, most eliminated already by the imdb criteria).
I think that should leave plenty of good films that are overlooked critically and only regarded as average or below by the public.

User avatar
Opio
Posts: 245
Joined: May 19, 2018
Location: U.S.
Contact:

#50

Post by Opio » November 24th, 2019, 6:49 am

I like these last several suggestions. Like Roger said, I'd like 6.5 as an IMDb cutoff but 7 could work too. Perhaps some comparable limit for non-imdb-ers.

User avatar
joachimt
Donator
Posts: 30932
Joined: Feb 16, 2012
Location: Netherlands
Contact:

#51

Post by joachimt » November 24th, 2019, 7:16 am

Opio wrote:
November 24th, 2019, 6:49 am
I like these last several suggestions. Like Roger said, I'd like 6.5 as an IMDb cutoff but 7 could work too. Perhaps some comparable limit for non-imdb-ers.
Non-IMDb-ers could always ask for help to filter their checks on IMDb-cutoff.
ICM-profile
Fergenaprido: "I find your OCD to be adorable, J"

User avatar
Fergenaprido
Donator
Posts: 3669
Joined: Jun 03, 2014
Location: Malaysia
Contact:

#52

Post by Fergenaprido » November 24th, 2019, 8:43 am

I feel PA's suggestions are a tad too restrictive and complicated, but they're not so bad as a compromise, and they're better than not having the poll at all.

Incidentally, I would push for one more restriction: no films from 2018 or 2019, as I think it's too soon to consider any of them underrated (or overrated).

User avatar
beavis
Posts: 2088
Joined: Jun 20, 2011
Location: Amsterdam, Netherlands
Contact:

#53

Post by beavis » November 24th, 2019, 9:11 am

I tried making a list with the MM-database available to me. setting the difference between my ratings and the userbase to at least 1,2 'stars' (2,4 points on Imdb) and a minimal number of ratings above 100, no films younger than 2017, maximum userbase score at 3.3, minimum score from me at 4 stars. After exporting the results to ICM I removed any movie that was on 4 lists or more, and came up with these 36 titles that, when looking at it, do indeed fill the description of "underrated" for me

https://www.icheckmovies.com/lists/unde ... en/beavis/

do they also comply with imdb criteria of no movies with less than 1000 ratings and none with a rating higer than 6,5??

NB when I include films on 4 ICM-lists there are also the titles Last House on the Left (original), Spring Breakers, Last Days and Copie Conforme.... they are not for everyone I guess, but I felt they were highly rated enough (or at least have a large enough group of strong fans), so that's why I went 3 lists or less.

NB2 only 3590 out of the 10700 features I've rated on MM were eligible for this, as all the other things I've seen have too few watches by the userbase... so only about 1% of these ended up to qualify as underrated!

NB3 only 10 out of my final list of 36 are on 0 official ICM-lists.

User avatar
RogerTheMovieManiac88
Posts: 1541
Joined: Feb 04, 2017
Location: Westmeath, Ireland
Contact:

#54

Post by RogerTheMovieManiac88 » November 24th, 2019, 10:19 am

Fergenaprido wrote:
November 24th, 2019, 8:43 am
I feel PA's suggestions are a tad too restrictive and complicated, but they're not so bad as a compromise, and they're better than not having the poll at all.

Incidentally, I would push for one more restriction: no films from 2018 or 2019, as I think it's too soon to consider any of them underrated (or overrated).
I agree in a way. PA's second guideline that pertains to the number of official lists/checks seems unnecessary to me. My alternative would be to exclude films that appear on two official lists or more. One official list appearance, it could be argued, might be a fluke but two appearances shows I think that the film is generally quite well regarded.

My suggestions would be these:

-Exclude films from 400>500 and TSPDT
-IMDb rating of 6.5 or under
-Exclude films that appear in 2 or more official lists
-Set a minimum number of IMDb votes that allows for the IMDb rating to be a relatively informative and representative reflection of the film's standing but DOES NOT unfairly exclude films that are still somewhat underseen. I am thinking something along the lines of 50, 100, 200 votes as a minimum requirement.
That's all, folks!

User avatar
albajos
Posts: 6821
Joined: May 24, 2016
Location: Norway
Contact:

#55

Post by albajos » November 24th, 2019, 10:33 am

Not if those two lists are DtC and 500<>400

User avatar
beavis
Posts: 2088
Joined: Jun 20, 2011
Location: Amsterdam, Netherlands
Contact:

#56

Post by beavis » November 24th, 2019, 10:36 am

If I would exclude films on 3 ICM-lists, I lose another 9 titles of my list
Now, I've seen quite a lot and my taste is not-mainstream, so I would expect my ratings to also differ somewhat from a general consensus... if I can only come up with about 30 titles, how doable is it to get enough titles that we have somewhat of a consensus on to generate a forum top-list out of this...?

How you tested your criteria Roger? how many movies did it get you?

User avatar
albajos
Posts: 6821
Joined: May 24, 2016
Location: Norway
Contact:

#57

Post by albajos » November 24th, 2019, 10:44 am

Yes, keep it simple. Max 7,0 rating with at least 1000 votes (1000 votes (not checks) are still high and will still exclude most of Norway, but 1000 is a normal statistic pool.)

User avatar
Onderhond
Posts: 3532
Joined: Dec 23, 2012
Contact:

#58

Post by Onderhond » November 24th, 2019, 10:48 am

I wouldn't put too much weight on the official checks count. TSPDT, Criterion and a few others may be an indication of a certain broader-than-niche appreciation, but a lot of official lists don't really fit that description. The Zombie 1000, the Fok 250, Bad-Movies list, the Action canon, Paste anime ... these lists are filled with niche.

User avatar
joachimt
Donator
Posts: 30932
Joined: Feb 16, 2012
Location: Netherlands
Contact:

#59

Post by joachimt » November 24th, 2019, 10:50 am

Onderhond wrote:
November 24th, 2019, 10:48 am
I wouldn't put too much weight on the official checks count. TSPDT, Criterion and a few others may be an indication of a certain broader-than-niche appreciation, but a lot of official lists don't really fit that description. The Zombie 1000, the Fok 250, Bad-Movies list, the Action canon, Paste anime ... these lists are filled with niche.
Agree.

Excluding TSPDT makes sense.
ICM-profile
Fergenaprido: "I find your OCD to be adorable, J"

User avatar
RogerTheMovieManiac88
Posts: 1541
Joined: Feb 04, 2017
Location: Westmeath, Ireland
Contact:

#60

Post by RogerTheMovieManiac88 » November 24th, 2019, 10:53 am

beavis wrote:
November 24th, 2019, 10:36 am
If I would exclude films on 3 ICM-lists, I lose another 9 titles of my list
Now, I've seen quite a lot and my taste is not-mainstream, so I would expect my ratings to also differ somewhat from a general consensus... if I can only come up with about 30 titles, how doable is it to get enough titles that we have somewhat of a consensus on to generate a forum top-list out of this...?

How you tested your criteria Roger? how many movies did it get you?
I haven't tested my criteria, beavis. I'm going by instinct.

As to albajos's point, I'd be in favour of excluding films that appear in 400>500.
That's all, folks!

User avatar
Lonewolf2003
Donator
Posts: 8423
Joined: Dec 29, 2012
Contact:

#61

Post by Lonewolf2003 » November 24th, 2019, 1:48 pm

Since there is more than enough animo I definitively will go ahead with the poll. I’m not completely decided if strict criteria are necessary ( and what they would be) or some guidelines like the ones Onderhond proposed are the way to go.

I def will exclude movies from 2018 and 2019. Maybe even 2017.
Last edited by Lonewolf2003 on November 24th, 2019, 1:54 pm, edited 1 time in total.

User avatar
Lonewolf2003
Donator
Posts: 8423
Joined: Dec 29, 2012
Contact:

#62

Post by Lonewolf2003 » November 24th, 2019, 1:54 pm

joachimt wrote:
November 24th, 2019, 10:50 am
Onderhond wrote:
November 24th, 2019, 10:48 am
I wouldn't put too much weight on the official checks count. TSPDT, Criterion and a few others may be an indication of a certain broader-than-niche appreciation, but a lot of official lists don't really fit that description. The Zombie 1000, the Fok 250, Bad-Movies list, the Action canon, Paste anime ... these lists are filled with niche.
Agree.

Excluding TSPDT makes sense.
Maybe expand we could expand it with some other big, broad critics lists like the S&S poll and 1001MYMSBYD?

User avatar
joachimt
Donator
Posts: 30932
Joined: Feb 16, 2012
Location: Netherlands
Contact:

#63

Post by joachimt » November 24th, 2019, 3:10 pm

Lonewolf2003 wrote:
November 24th, 2019, 1:54 pm
joachimt wrote:
November 24th, 2019, 10:50 am
Onderhond wrote:
November 24th, 2019, 10:48 am
I wouldn't put too much weight on the official checks count. TSPDT, Criterion and a few others may be an indication of a certain broader-than-niche appreciation, but a lot of official lists don't really fit that description. The Zombie 1000, the Fok 250, Bad-Movies list, the Action canon, Paste anime ... these lists are filled with niche.
Agree.

Excluding TSPDT makes sense.
Maybe expand we could expand it with some other big, broad critics lists like the S&S poll and 1001MYMSBYD?
Excluding S&S makes sense as well. Maybe even 21st Century, because it's basically the TSPDT-list for this century.

Who is responsible for the selection of movies on 1001MYMSBYD? The critics who wrote the pieces? Or did Schneider himself make the selection (or at least did he do so for the first edition(s)? I wouldn't exclude single critics lists like Rosenbaum for example.
ICM-profile
Fergenaprido: "I find your OCD to be adorable, J"

User avatar
Angel Glez
Posts: 2064
Joined: Apr 02, 2012
Location: Spain
Contact:

#64

Post by Angel Glez » November 24th, 2019, 3:14 pm

I would choose the easiest way: Movies underrated by IMDb's users. If average rating is 6.8, then the cutoff must be <6.8

TSPDT (or any other critics' list) is a completely different criteria, I think.

mjf314
Moderator
Posts: 11044
Joined: May 08, 2011
Contact:

#65

Post by mjf314 » November 24th, 2019, 4:19 pm

I think the criteria should be:
Favorite films rated lower than 7 on IMDb

I would be ok with a 1000-vote minimum, but it's not really important. Not many films with less than 1000 votes will make the list anyway.

I'm undecided on whether or not to exclude TSPDT.

I don't think number of official lists is a good criteria because it puts certain types of films at a disadvantage.

I think "personal rating - IMDb rating > 2" isn't a good criteria, because different people use different rating scales. For some people, 7/10 = average. For other people, 7/10 = good.

User avatar
Fergenaprido
Donator
Posts: 3669
Joined: Jun 03, 2014
Location: Malaysia
Contact:

#66

Post by Fergenaprido » November 24th, 2019, 4:46 pm

RogerTheMovieManiac88 wrote:
November 24th, 2019, 10:19 am
Fergenaprido wrote:
November 24th, 2019, 8:43 am
I feel PA's suggestions are a tad too restrictive and complicated, but they're not so bad as a compromise, and they're better than not having the poll at all.

Incidentally, I would push for one more restriction: no films from 2018 or 2019, as I think it's too soon to consider any of them underrated (or overrated).
I agree in a way. PA's second guideline that pertains to the number of official lists/checks seems unnecessary to me. My alternative would be to exclude films that appear on two official lists or more. One official list appearance, it could be argued, might be a fluke but two appearances shows I think that the film is generally quite well regarded.

My suggestions would be these:

-Exclude films from 400>500 and TSPDT
-IMDb rating of 6.5 or under
-Exclude films that appear in 2 or more official lists
-Set a minimum number of IMDb votes that allows for the IMDb rating to be a relatively informative and representative reflection of the film's standing but DOES NOT unfairly exclude films that are still somewhat underseen. I am thinking something along the lines of 50, 100, 200 votes as a minimum requirement.
I like your suggestion even less. :D
Onderhond wrote:
November 24th, 2019, 10:48 am
I wouldn't put too much weight on the official checks count. TSPDT, Criterion and a few others may be an indication of a certain broader-than-niche appreciation, but a lot of official lists don't really fit that description. The Zombie 1000, the Fok 250, Bad-Movies list, the Action canon, Paste anime ... these lists are filled with niche.
Fully agreed. Plus, there are lists compiled by a single person (Rosenbaum, Truffaut, the director lists) that are hardly indicative of broad support.
mjf314 wrote:
November 24th, 2019, 4:19 pm
I think the criteria should be:
Favorite films rated lower than 7 on IMDb

I would be ok with a 1000-vote minimum, but it's not really important. Not many films with less than 1000 votes will make the list anyway.

I'm undecided on whether or not to exclude TSPDT.

I don't think number of official lists is a good criteria because it puts certain types of films at a disadvantage.

I think "personal rating - IMDb rating > 2" isn't a good criteria, because different people use different rating scales. For some people, 7/10 = average. For other people, 7/10 = good.
Overall agreed. The last line may be a guideline instead of a hard and fast rule, but I think it's a useful guideline that helps people think in the way the list is presumably intended.

mjf314
Moderator
Posts: 11044
Joined: May 08, 2011
Contact:

#67

Post by mjf314 » November 24th, 2019, 5:03 pm

I think "Vote for the films that maximize the rating difference" is still not a good criteria.

For example, suppose a film's IMDb rating is 1/10, and you rated it 5/10. Should you vote for it? I don't think so. I'm not interested in a list of films that people consider average. I want a list of films that people actually like.

User avatar
Fergenaprido
Donator
Posts: 3669
Joined: Jun 03, 2014
Location: Malaysia
Contact:

#68

Post by Fergenaprido » November 24th, 2019, 5:08 pm

mjf314 wrote:
November 24th, 2019, 5:03 pm
I think "Vote for the films that maximize the rating difference" is still not a good criteria.

For example, suppose a film's IMDb rating is 1/10, and you rated it 5/10. Should you vote for it? I don't think so. I'm not interested in a list of films that people consider average. I want a list of films that people actually like.
Fair point, but I'd be surprised if anyone actually did that.

User avatar
RogerTheMovieManiac88
Posts: 1541
Joined: Feb 04, 2017
Location: Westmeath, Ireland
Contact:

#69

Post by RogerTheMovieManiac88 » November 24th, 2019, 5:18 pm

Haha, okay Fergenaprido. The more I think about it, the more I like the idea of keeping it as simple as possible in terms of parameters.
That's all, folks!

User avatar
albajos
Posts: 6821
Joined: May 24, 2016
Location: Norway
Contact:

#70

Post by albajos » November 24th, 2019, 5:20 pm

Yes, I would still only vote for 10, 9 and 8 ratings.

But everyone know that Mommie Dearest will win ;p (The nost unfair Razzie winner)

User avatar
Lonewolf2003
Donator
Posts: 8423
Joined: Dec 29, 2012
Contact:

#71

Post by Lonewolf2003 » November 24th, 2019, 10:11 pm

I wouldn’t vote for anything I didn't like, rated lower than a 7

mathiasa
Posts: 2408
Joined: Aug 18, 2013
Contact:

#72

Post by mathiasa » November 28th, 2019, 11:27 am

The problem with a strict rule comparing one‘s own rating with the imdb rating is that a 7/10 (for example) has different meanings to different voters. I often see joachimt rating something 6.0 - 7.0 and then writing something like „it was ok“. For me, a rating in that range would mean a much higher personal value. And there‘s no right or wrong way of assigning numbers to valuations, so a strict enforcement of any such rule would be counter-productive.
In my rating system (as in that of Roger Ebert), a higher rating doesn‘t even necessarily mean a better movie or higher valuation of the movie. I‘m for example more lenient in my rating with movies of a genre I don‘t particularly like.

A minimum check count seems completely nonsensical to me. One meaning of underrated is, “something or someone not well known by the mass public but of high quality.“

And yes, „Cool as Ice“ is the most underrated film of all time.

User avatar
RogerTheMovieManiac88
Posts: 1541
Joined: Feb 04, 2017
Location: Westmeath, Ireland
Contact:

#73

Post by RogerTheMovieManiac88 » November 28th, 2019, 12:09 pm

mathiasa wrote:
November 28th, 2019, 11:27 am
A minimum check count seems completely nonsensical to me. One meaning of underrated is, “something or someone not well known by the mass public but of high quality.“
I'm not hung up on that in all honesty. The likelihood of films with such low vote counts showing up in any final results is pretty slim. I threw it out there as a suggestion because I do think it's important that the IMDb rating is relatively reflective of the sort of estimation the film is held in, if we are going to go by IMDb ratings. I'm not sure that a film with five, ten, or twenty, or perhaps even fifty votes can be with confidence classsed as underrated on the 10 point scale.

I suppose ''underrated'' can be interpreted as signifying something that does not have many votes. I believe that the word, in the context of this poll, refers to films that members feel are rated too low on IMDb. I could be wrong but that's how I read it and assumed it was intended.

A poll for great films with under 100 or 200 or 500 votes would be another interesting undertaking but that's a whole other kettle of fish!
Last edited by RogerTheMovieManiac88 on November 28th, 2019, 12:44 pm, edited 3 times in total.
That's all, folks!

User avatar
Onderhond
Posts: 3532
Joined: Dec 23, 2012
Contact:

#74

Post by Onderhond » November 28th, 2019, 12:11 pm

mathiasa wrote:
November 28th, 2019, 11:27 am
One meaning of underrated is, “something or someone not well known by the mass public but of high quality.“
I don't think it is to be honest, but some people do use it for that (and it's often an effect of being underrated).

mathiasa
Posts: 2408
Joined: Aug 18, 2013
Contact:

#75

Post by mathiasa » November 28th, 2019, 12:37 pm

Onderhond wrote:
November 28th, 2019, 12:11 pm
mathiasa wrote:
November 28th, 2019, 11:27 am
One meaning of underrated is, “something or someone not well known by the mass public but of high quality.“
I don't think it is to be honest, but some people do use it for that (and it's often an effect of being underrated).
‚underrated‘ refers to things or persons that don‘t get proper recognition. This can be either because a) it isn‘t liked by or b) it isn‘t known to the public. That‘s how the word is used and it is the usage of a word that defines its meaning. What you think the word means is no more or no less important than what other people think.

But even if we would limit this poll to the sub-meaning a), a minimum check restriction would still be an arbitrary rule.

User avatar
albajos
Posts: 6821
Joined: May 24, 2016
Location: Norway
Contact:

#76

Post by albajos » November 28th, 2019, 12:49 pm

What's arbitrary about it?

Math is a science. And there is a consencus that 1000 is a usual number of respodents to get the most useable stats. They could collect more, but it's usually not neccessary as the margin of error is already at 1% at 1000.

User avatar
Onderhond
Posts: 3532
Joined: Dec 23, 2012
Contact:

#77

Post by Onderhond » November 28th, 2019, 12:50 pm

Yes, everything is arbitrary, everything is relative. Genres are, countries are, best-of criteria are. But most of all the quality of the resulting ICMF lists.

mathiasa
Posts: 2408
Joined: Aug 18, 2013
Contact:

#78

Post by mathiasa » November 28th, 2019, 12:51 pm

RogerTheMovieManiac88 wrote:
November 28th, 2019, 12:09 pm
mathiasa wrote:
November 28th, 2019, 11:27 am
A minimum check count seems completely nonsensical to me. One meaning of underrated is, “something or someone not well known by the mass public but of high quality.“
I'm not sure that a film with five, ten, or twenty, or perhaps even fifty votes can be with confidence classsed as underrated.
You seem to be unaware of how the rating typically changes with additional votes. A movie with 5-50 votes generally has a much higher rating than the same movie with more votes. The first voters are usually the most enthusiastic (or they were even part of the production), leading to a higher average rating than the movie will have later on (should it get more votes).
So it‘s already naturally much harder for a movie with few votes to be (rating-wise) underrated. So, if a movie with few voters already has a low rating, this is really significant and meaningful as we have to expect the rating to drop even further with additional votes. If you happen to highly value such movie, it would therefore make sense to call it underrated.

User avatar
OldAle1
Donator
Posts: 4084
Joined: Feb 09, 2017
Location: Dairyland, USA
Contact:

#79

Post by OldAle1 » November 28th, 2019, 1:02 pm

mathiasa wrote:
November 28th, 2019, 12:51 pm
RogerTheMovieManiac88 wrote:
November 28th, 2019, 12:09 pm
mathiasa wrote:
November 28th, 2019, 11:27 am
A minimum check count seems completely nonsensical to me. One meaning of underrated is, “something or someone not well known by the mass public but of high quality.“
I'm not sure that a film with five, ten, or twenty, or perhaps even fifty votes can be with confidence classsed as underrated.
You seem to be unaware of how the rating typically changes with additional votes. A movie with 5-50 votes generally has a much higher rating than the same movie with more votes. The first voters are usually the most enthusiastic (or they were even part of the production), leading to a higher average rating than the movie will have later on (should it get more votes).
So it‘s already naturally much harder for a movie with few votes to be (rating-wise) underrated. So, if a movie with few voters already has a low rating, this is really significant and meaningful as we have to expect the rating to drop even further with additional votes. If you happen to highly value such movie, it would therefore make sense to call it underrated.
Generally agree though I would point out that the opposite happens quite frequently as well - movies start out heavily negged, usually in a campaign by people who probably didn't see it, and often for "political" reasons, i.e. Black Panther and Captain Marvel, neither of which had anything like the high ratings that most MCU films have initially, despite getting better than average reviews and box office. Now they're probably not too far off though probably still a bit below the last two heavily fanboy-upvoted Avengers flicks. Or any Spike Lee film, or anything seen as overtly feminist. Of course I'm talking specifically about IMDb voting, it may be less extreme on other sites - but in my mind IMDb scores on recent, mainstream films are typically (even more) worthless (than usual) until the film hits 50 or 100k votes or more.

Or the short explanation: IMDb voting is heavily trolled, and therefore nearly always worthless.

mathiasa
Posts: 2408
Joined: Aug 18, 2013
Contact:

#80

Post by mathiasa » November 28th, 2019, 1:15 pm

albajos wrote:
November 28th, 2019, 12:49 pm
What's arbitrary about it?

Math is a science. And there is a consencus that 1000 is a usual number of respodents to get the most useable stats. They could collect more, but it's usually not neccessary as the margin of error is already at 1% at 1000.

Defining an appropriate sample size is a bit more tricky than simply going with 1000. I‘m not that good with statistics but
from what I know, a good sample size is a certain percentage of the population, up to about a 1000. If I understand it correctly, the population in our case would be the people who watched a certain movie. But we don‘t know that number, so we can‘t calculate sample sizes precisely. But what we know is that a movie that has been seen by less than 1000 persons, cannot have a sample size of 1000.

Further, as others have already pointed out, imdb ratings have multiple biases (eg males often make about up to 90% of the samples). If you would want to approach this poll from a scientific perspective, there would have to go a lot of work into it (correcting all the biases and the trolling pointed out by OldAle1 in the post above). I don‘t see anybody here wanting to do that, nor is it necessary.

Post Reply