Would a Photofeeler woman tell me to just be myself and take it slow?

[ThePhoenix]

 
 
TL;DR: Read the big print.
 
 
 
Iʼm trying to decide if I should insert more coins. Itʼs just slightly addictive, but is that all it is?

Iʼm not primarily worried about voters just click-bombing their way to credits. As Photofeeler is quick to point out, they implement a machine learning system to weed that out. Iʼm sure itʼs not perfect, but Iʼve been on the other end of that algorithm, before time constraints made me give in and insert coins, so I know itʼs at least there. I only like black girls, and it didnʼt take long for their system to accuse me of giving them garbage votes while rating dating pics. I soon had to switch to rating business and social pics. (Because only my dick is racist.  LOL)

Iʼm more worried about the women who are sincerely voting what they consciously believe to be attractive, based on the same societal conditioning that led me to GC in the first place.

The title of this post alludes to how notoriously terrible women are at giving dating advice. For as long as I did all the things women say are “sweet” and for as long as female friends described me as a guy who would “make a girl so happy”, I stayed lightyears away from pussy. And had a hard time understanding why they stuck with guys who blatantly contradicted what they said they wanted. Itʼs only when I started to take on and practice mentalities that were dead opposite of what they said they wanted, that I started to have some success with women (in real life — Iʼve never tried the on‑line thing yet).

Iʼm sure itʼs not intentional. They just, with occasional exceptions, have really poor insight into what actually attracts them!

If I had a dollar for every time a girl has exclaimed in apparent disgust, “I donʼt like that guy”, and then somehow her pussy just happened to fall on his dick, Iʼd just buy out Tinder.

So my question is, is asking a bunch of random women to rate pictures as pointless and counter-productive as it would be to ask those same women what I should do on a first date?

My concern is that a womanʼs mental process when being asked to rate a picture could be entirely different than her mental process when deciding whether or not to invite the possibility of getting a guyʼs dick.

The only way you could entirely eliminate that concern would be to test pictures in the dating apps themselves, and I know some guys do that. Well, Iʼm a day game guy and Iʼm not willing to invest such a remarkable effort into on‑line. Especially not with Tinder having recently made it a lot harder to run multiple fake accounts in parallel. So my only options would seem to be either plunk some more money into Photofeeler and combine its results with my own intuition, or just roll with what Iʼve already figured out.

Iʼve read some opinions of guys that Photofeeler is useful, while others have said itʼs pretty arbitrary. (Photofeeler advertises something like a 200% to 300% increase in matches, but obviously theyʼre going to flatter themselves.)

One thing Iʼve noticed, although not by thorough and controlled testing, is that pictures where Iʼm smiling, particularly if itʼs an open‑mouth smile, and pictures where Iʼm looking at the camera tend to do better. Pictures with Byronic or bad‑boyish poses tend to score worse, even though some of the same bad‑boy pics have actually garnered multiple likes from peripheral girls on Facebook. That makes me a little suspicious, although there are confounding factors and I donʼt have nearly enough data to say with certainty, that the pictures being ranked highest are not the ones that best apply seduction advice.

Of particular concern is a picture I tried on both Photofeeler and Hot or Not. I had a socially awkward, gay friend behind my camera, telling him when to press the buttons. When I struck this particular pose, his response was “Phoenix! You canʼt use that, it looks like you just got out of jail!!” I told him to just press the damn button, and that picture went on to earn a 7.8 on Hot or Not. Photofeeler gave it 3.1. This is a telling comparison, because the women on Hot or Not can actually potentially hook up with you, whereas those on Photofeeler canʼt.

Iʼve also noticed that the comments are not always consistent with the scores. For instance, on almost every picture where I have sunglasses on, there is one and typically several comments against the sunglasses, and yet my so‑far best rated picture has sunglasses on — and three comments against them.

One of my best rated pics is of me and a woman, both in cultural costumes that are rather revealing, especially on the girl. Well, one of the voters harshly chastised my presenting beside a partly naked woman. (Mind you, thus far, my best pic after a fair bit of experimenting and photo editing only has a 5.4/10, so even this being one of my best, at 4.9/10 it doesnʼt say all that much.)

Inspired by the studies that red is more sexually attractive, I decided to do a little experiment. I took a pic where I had on a blue top and changed it to red in Photoshop (and changed the shorts from dark blue to dark grey), and ran both. The red did score just slightly higher, but their confidence intervals heavily overlapped, so that could have been random chance.

So, Iʼm wondering whether I ought to continue exploring with Photofeeler, at about 15 cents a vote, or just say, fuck it, and run with what I have. Tinder is only a crap shot anyway, to me. I prefer day game and am probably going to move to Africa so that I can do more of it.

To put things into context, because I think for some people Photofeeler really is essential, I donʼt really need ratings as to what images are or arenʼt good on a technical or general artistic level. Itʼs not like I wouldnʼt know to recognize a soft, blown‑out, or noisy image, or terrible composition. I work in a highly visual industry, Iʼm no stranger to Photoshop, and indeed Iʼve done my own editing on many of the images, in some cases extensively.

I really only need ratings for the “pussy juice flow meter” element, LOL. And thatʼs the very element where Iʼm skeptical that female votes lacking “skin in the game” actually mean anything.
 

 

[JacobPalmer]

I actually do use photofeeler, AND I have also tracked which photos get the most "likes" on Hinge, so I'll share my insights.

As a whole, yes, photofeeler does give you some good results. You want to to get high scores in all 3 categories, and that has been consistent with the data that Hinge gets me too. My highest scoring photo on photofeeler was also my most liked photo on hinge.

BUT, you also want to look at something else - comments.
If you have a good photo and are getting quite a few "I would date them" or "gorgeous" or anything else of that sort, then that photo is great. The "great smile" and "great photo" I pay less attention to.

The negative comments like the sunglasses one are also true, but a good sunglasses pic can work well as a supplementary photo, not your main one.
If you score pretty well with a picture but you have quite a few negative comments, ditch it. I've found that the negative comments outweigh the positive ones, this was also true of my data from hinge, which was this:

I had three "face" photos up, all scored high on photofeeler for attractive, but one of them had quite a few negative comments on it. And you know how many times a girl chose that photo instead of the other 2 face photos I had? NOT ONCE!

But if you are getting low scores like this, unfortunately you're going to have to improve your physical fundamentals a lot (or do a few photo shoots to get your best side), or online dating is not going to work for you. Sorry

 

[Rain]

Didn't know about photo feeler, I had heard of hot or not. I think chase andor zacadam had one or two other photo rating websites if you search.

But, I tested photofeeler for free over the last few days, and yeah, it backs up my online testing over the years with different photos. So I recommend it.
Make sure to look at the data. eg see how many says "attractive no" "attractive somewhat" 'attractive yes" "attractive very"
Don't look at the numbers, look at the votes under "data".

In photofeeler, I make sure it was in the "dating" category, and that only women could vote who were 29 or younger. I also only needed about 10-15votes[sometimes less,] to backup my experience.

I know that isn't alot of votes per photo statistically, but it backs up what I've experienced with certain photos. Go for more votes if you want to, but just saying, if one photo stands out more, 15votes might be all you 'd need to find out

Was done for free, and yeah I'd have to vote on others photos to get the "karma" up, then people vote on mine in 24hrs, then when I changed photos, I have to do more votes.

 

[Carousel]

A dating company here in Norway did this. They wanted to create top notch dating profiles and used women the user DID NOT KNOW to rate his pictures. This turned out to be more reliable than the male instructors rating the pictures. I did this myself back in the days and the results were better than any other process.

This was not women from some internet page though, but rather real-life friends of the instructors etc. I just asked women a friend of mine knew to rate my pictures. It is however important that they are not YOUR friends, that was found to destroy objectivity.

 

[ThePhoenix]

Thanks for the info guys!

I did go ahead and order a whack of credits, while keeping some healthy skepticism.

I actually do use photofeeler, AND I have also tracked which photos get the most “likes” on Hinge, so Iʼll share my insights.
...
My highest scoring photo on photofeeler was also my most liked photo on hinge.

This, validating it against “skin in the game” sort of data, is a great idea.

One concern I do have here, though, is that Hinge is very much marketed as a relationship app, so Iʼd be a little bit worried that a “like” on Hinge might basically equate to, “you look like good husband material.” If that is the case, that would be very bad from the perspective of seduction, for reasons Iʼm assuming youʼre aware of. (But I could be wrong about Hinge — Iʼm just going off the way they market it.)

If that were the case, it would somewhat correlate with my own observations. Photofeeler for me seems to prefer my harmless, friendly guy poses by far to my badboy poses. It also way prefers my photos where I was in an environment that suggests having money — even though my body language was at that time screaming “Iʼm a virgin!” (Guys who are not very good at getting sex make for better, more easily controlled husbands.)

Mainstream (i.e., cuck) dating advice is to look directly at the camera with a friendly smile. Advice from experienced seducers, by contrast, seems to be to look away from the camera and not smile. If you do smile, seducers generally hold that full smiles are friendly as opposed to sexy.

I whipped up a quick spreadsheet to average the “Attractive” rating of all my Photofeeler photos across those attributes. Relative to each other, the ones where I was looking at the camera got scores 73% higher than those looking away. The ones where I was smiling got twice the average score of the ones where I wasnʼt, and the ones with a full smile averaged 12% higher scores than those with a partial smile. While I didnʼt delve into a statistically robust analysis, this does seem to suggest that Photofeeler may be preferring friendly and harmless over mysterious and sexy.

I also did a test on two different photos where, for each, I tried two versions that were completely identical except that my shirt had been digitally edited from blue to red or vice‑versa, being that multiple studies have shown red to unconsciously trigger sexual attraction in both sexes. While the results were not different enough to get strong statistical confidence without pumping in more votes than I felt like paying for, both of these tests showed a mild “Attractive” score preference for the blue shirt, with relative increases of 9.8% and 8.7%. (I had reported above that the red shirt was favored, but that was a preliminary result that reversed when statistical confidence was improved with more votes.)

What would have been expected to be less sexually attractive according to a controlled, objective measure, got a higher score — again suggesting that these scores may run very contrary to sexual attraction.

Paradoxically and perhaps tellingly, the only two “I would date them!” comments Iʼve gotten on this platform came from both of the red shirt tests even though they scored lower than their blue counterparts. I reason that these remarks are more spontaneous than the ratings and as such are more likely to originate from true sexual attraction as opposed to the sort of calculated social reasoning designed to ensnare husbands.

That also seems to support your remark that certain comments may be more relevant than scores. In particular, I take being told I look “timid” or “uncomfortable” as very bad signs, although “arrogant” Iʼm not so sure is actually bad.

I think chase andor zacadam had one or two other photo rating websites if you search.

I did see a post from Chase with a machine-based system. It seemed to be looking mostly at technical image quality — things like noise, contrast, focus, etc.

But, I tested photofeeler for free over the last few days, and yeah, it backs up my online testing over the years with different photos.

Interesting! I assume here you mean testing youʼve done on actual dating sites. Were these more hookup-oriented or relationship-oriented sites/scenarios?

Make sure to look at the data. eg see how many says “attractive no” “attractive somewhat” ʼattractive yes“ ”attractive very“
Donʼt look at the numbers, look at the votes under “data”.

Yeah, I always look at the vote data as well as the scores. I do have to point out one thing here, though. Photofeeler uses a machine learning algorithm to adjust votes based on votersʼ tendencies when computing the total score. For instance, a “Very Attractive” from a girl who tends to rate very liberally is not as valuable as the same rating from a girl who rarely votes “Very Attractive” — the numerical score takes that into consideration.

Actually, you can see the way these are computed if you look at the dark spots along the bars representing the votes. If you put your mouse over these spots, you see something like, “This voter rates low, so their vote has been adjusted upwards in calculating your score.”

For that reason I would advise using the score, but one reason to look at the vote data as well is that some photos may be very polarizing. (Mind you, Iʼve never had this happen myself.) So, you might have a photo which most girls give a medium response to, and another photo which some girls love and others hate. That would tell you something important that you canʼt see just from looking at the score, since both photos could get a similar score because {0,0,0,10,10,10} has the same average as {5,5,5,5,5,5}.

Go for more votes if you want to, but just saying, if one photo stands out more, 15votes might be all you ʼd need to find out

Hm. I tend to look at the confidence intervals, which you can see if you hover your mouse over the score. I think you donʼt need too many votes to get a general idea, but letʼs say you have two pictures youʼre trying to decide between. If the score from the one photo is within the confidence interval of the other one and vice‑versa, then statistically speaking, the evidence is not strong that the one with the higher score is actually the better one.

Iʼve seen this happen, where Photo “A” scored higher than Photo “B”, but their confidence intervals overlapped, and once I added more votes, they flipped and now Photo “B” is higher. (Iʼve paid for credits, in which case you can dump in votes pretty quickly and to multiple tests at once.)

One thing I donʼt know is what percentile their confidence intervals are computed for. For instance, you could have an 80% CI vs. a 95% CI vs. a 99.5% CI, etc., which makes a huge difference. They donʼt say anywhere what the percentile is. They could be using a very high percentile in order to encourage people to buy more credits, LOL. Generally, if the intervals only overlap by a small amount, then I consider it probably a valid ranking (to the extent that the votes even mean anything, of course).

This turned out to be more reliable than the male instructors rating the pictures.

Wow, thatʼs interesting and somewhat unexpected! Iʼd almost wonder if these were good instructors. As we know, many dating instructors are clueless, especially the ones that convey socially acceptable ideals.

It is however important that they are not YOUR friends, that was found to destroy objectivity.

Yeah, female friends are the worst. Well, maybe not all, but they tend to be.

I once made the mistake of letting female friends pick out clothes for me. A few years later I realized what they picked was pretty lame. On another occasion, I told a female friend some clothing I was thinking of getting, and she told me not to, that it would not look good on me. I bought it anyway and when she saw me in it she was surprised and made a 180° and told me how good it looked! LOL


Cheers,
Phoenix

 

[Carousel]

Wow, thatʼs interesting and somewhat unexpected! Iʼd almost wonder if these were good instructors. As we know, many dating instructors are clueless, especially the ones that convey socially acceptable ideals.

I don't think they were clueless, as the WORST choices were made by the customers themselves. The instructors did better in choosing pictures than the customers, but not as good as the women. This may boil down to simply hardware differences

 

[Rain]

Interesting! I assume here you mean testing youʼve done on actual dating sites. Were these more hookup-oriented or relationship-oriented sites/scenarios?

Yes, testing I've done on dating sites and when I have to send a second photo as i only have one on my profile. The sites were more dating/relationship eg pof.

With the ratings, I did see they were adjusted, but I'm not sure if they only adjust the "numbers"[scores] or if they also adjust the actual category[not attractive, somewhat attractive, yes attractive, very attractive]
If its the latter than you oculd say the results dont matter, because you're just throwing away what people are voting with. no adjustments should be made. But if its the former than just don't look at the 'scores'. I'm not sure which it is?

 

[ThePhoenix]

This may boil down to simply hardware differences

XD

Well, now, this is interesting. Much of the feedback on this thread seems to suggest, plus or minus, that women actually can and do tap into that hardware when rating pictures of guys.

Thereʼs certainly a hardware element to making such judgments. I mean, Iʼve had it happen on many occasions that a male friend who knows I like black girls will say I would love a particular girl, sheʼs exactly my type. Meanwhile, I look at her, and sheʼs aight, but sheʼs not one of those black girls I go NUTS for. I think that on some level, theyʼre simply reasoning, “*I* think this girl is pretty attractive, and Phoenix loves black girls so heʼd have to love her even more.” As if my attraction system is exactly the same as his with the sole exception of liking black girls more. But it doesnʼt quite work that way, because there are actually very specific features, especially facial features, that my particular unique brain is drawn to.

But all this begs the question, if women actually can tap into this hardware when judging a picture, then why are they so pathetically horrible at judging what male behaviour would or wouldnʼt activate those circuits? What figures, that they can accurately judge photos but not behaviour? And why wouldnʼt behavioural elements of photos fall victim to the same poor advice as behaviour itself falls victim to?

I mean, women routinely advise men to “take it slow”, but we all know what thatʼll do to a seduction. Why doesnʼt their hardware give them the conscious insight to advise men to try to swoop a girl into bed as quickly as possible?


On that same vein, smiling is really getting to me... Photofeeler definitely seems to prefer photos of me smiling to photos where Iʼm not, and Photofeelerʼs blog — no doubt written by women and/or cucks — also advocates smiling and looking at the camera, as do most other mainstream (and as such usually very bad) sources of dating advice. Meanwhile, credible seducers seem to lean fairly strongly towards not smiling, and to some extent, not looking at the camera. Itʼs so confusing, and suspicious.

With the ratings, I did see they were adjusted, but Iʼm not sure if they only adjust the “numbers”[scores] or if they also adjust the actual category[not attractive, somewhat attractive, yes attractive, very attractive]

If its the latter than you oculd say the results dont matter, because youʼre just throwing away what people are voting with. no adjustments should be made. But if its the former than just donʼt look at the ‘scores’. Iʼm not sure which it is?

Iʼm not quite sure what youʼre saying.

I donʼt really know what their formula is — theyʼre not exactly transparent about it — but what it would seem like to me, just based on what little theyʼve said and also looking at their graphical representations, hereʼs what Iʼd assume:

Say we want to turn a given voterʼs {No/Somewhat/Yes/Very} into a 0 to 10 score.

If they say “No”, we turn this into anywhere from 0 to 2.5 depending on just how hard this person votes.
If they say “Somewhat”, we turn this into anywhere from 2.5 to 5, depending on the voter.
If they say “Yes”, we turn this into anywhere from 5 to 7.5, depending on the voter.
If they say “Very”, we turn this into anywhere from 7.5 to 10, again depending on how often this voter gives a “Very”.

That is to say, I donʼt think theyʼre just changing someoneʼs “No” to a “Yes” just because they usually always say “No”. I think theyʼre just making it a weaker “No” — a “No” that is scored closer to “Somewhat”. Does that make sense? In any case, I donʼt think theyʼd flip a vote to a different response in the talleys given in the Data tab.

Come to think of it, the above makes me understand what they did to me back when I actually voted. My extreme preference for black women combined with the extreme lack of black women in their user base led to me being quite harsh..lol. Well, their system eventually accused me of giving them garbage votes. It would make sense given the above system that they canʼt use your votes if the required adjustments would exceed the above intervals (eg., a “No” needs to be higher than 2.5/10).

But I think thereʼs another step going on in making the score, as well. I think I read somewhere in their blog that the scores are percentiles. That means, if they give you 8/10, that doesnʼt mean that the average of all your votes was 8/10. It means that you got better votes than 80% of guys.

Since they use an opaque formula that we donʼt fully know, Photofeeler scores are basically not comparable with anything other than Photofeeler scores. But thatʼs going to be true of most ratings.

If you meant that the scores are meaningless because of the adjustments, I disagree. Theyʼre more likely to be meaningless without the adjustments, and hereʼs why.

Youʼre going to have some women who vote like an uppity princess and give almost everyone “No” or maybe “Somewhat” if theyʼre lucky, and only give “Very” to the occasional supermodel. And then youʼre going to have other women who vote like a thirsty walrus, and give the majority of guys “Yes” and many of them “Very”.

Well, thereʼs a principle in statistics that states that if you take the average of a large enough number of votes, the individual effects of the walruses and princesses will disappear as they will tend to cancel each other out over a large enough sample. Donʼt quote me but I think itʼs called the “Law of Large Numbers”. The catch is, for this to work, you need a very large number of data points — like tens of thousands or more.

The thing is, on Photofeeler, votes are expensive, and that means that weʼre usually trying to gain some kind of meaningful insight off a very small number of votes, maybe as little as 10 or 15.

Well, statistically speaking, if you only have 10 votes on each of two photos, say “A” and “B”, the chances that the voters for photo “A” just happen to be significantly more uppity than the voters for photo “B”, purely by random chance, are actually quite high. So, if that does happen, and photo “B” comes out with a better raw average score, do you think that means “B” is actually better? No, because itʼs probably just that two or three of those 10 voters on photo “A” happened to be snobs.

What theyʼre doing when they adjust the votes, is theyʼre allowing you to get a much better statistical power for the same number of votes. I wouldnʼt go against that. Unless you want to pay them a few thousand dollars per picture to get enough votes that your vote counts will be statistically accurate without having to make any intelligent adjustments.

 

[Fuck This]

Dredging this up to see what folks thing 7 months later on the Photofeeler for choosing pics.

My top pictures :
My latest was a "knees up" standing shot with a drink in my hand at dusk. (S/T/A) 7.0/ 9.0/ 9.1 6 notes 2 of them "would date"
Same picture with darker lighting, and sunset behind me "Waist up" (S/T/A) 7.6/ 8.9/ 8.6 7 notes 1 "would date", 3 too dark, 3 great photo
A Selfie of my head and neck, wearing a hat and smiling, showing off my dimples (S/T/A) 8.3/9.5/9.4 14 notes were 6 "great smile" 5 great photo, 1 "would date"

I use those numbers to choose one pose/photo over another...not an objective score per se...