Is level 0 of Steam Gifts infested with bots?
If it makes you feel any better, I read every description of GAs I enter, unless its a huge essay's worth of words. (say, 500+ words?)
- I have Hearts of Iron IV, so I didn't get to see this nor participate in it.
So the results may not be completely accurate as it excludes people who own these games (more likely to be real humans, perhaps?). - Users who are not interested in certain games may also manually hide certain games, or only enter for games they know they'd play. (like me)
- I think some users on this site also use ESGST's one click to enter, and were unable to see the description.
So the results would show a greater percentage of bots than there actually are?
I agree that the number of bots who enter games are at quite an alarming number though.
Just my worthless? two cents :0
Comment has been collapsed.
Exactly the reasoning, it's not that 100k users saw HOI, 4k entered and .5k entered RL. But still the pool is reduced to 4k HOI and then the unique entries from Memoria and it's dlc and Insurgency, so maybe 4.5k target pool. Of that what percentage would use ESGST and not see comments? guess that's their problem in any case, lets say 50% (I'd presume far less but who knows). 2.2k (unique) entered those 4 and .5k entered RL, I think that's still pretty good and easily possible of 100% human entries since some people might have it, others not interested etc. It was just very interesting though that TEW2010 which is a far inferior game to RL got almost the same entries in a little less time.
As I said though, the next test will be the one, after I set this one up I realised it was all show and no substance, that it could never give any real result even with the increased exposure from 4 games displaying the links. The next one though will have 100% exposure of the links to 100% of the entries, even with comment skippers etc I am expecting it to be telling with at a guess 10% entering the links is my prediction.
Comment has been collapsed.
That's fair. Thanks for the explanation and the reasonings you had here!
I think ~10% is a fair estimate, I think someone did a similar test before and got around the same percentage.
It might change depending on the type of game you put up though.
(Eg If you asked a question on a commonly bundled game, people might not answer, but if you did on an AAA game, more people would.)
Comment has been collapsed.
Well some people probably owned those games so never saw the train link.
Comment has been collapsed.
Those people would not be included in the pool and have no effect on the result.
Comment has been collapsed.
1st part, true, 2nd part, not per se.
Let's say the demographic of people who like reading (descriptions) really like HOI4, Memoria and Insurgency, so majority of them already have it and won't see the giveaways. The people who are left are less likely to be 'description readers' so the fact the first group is excluded might have an effect on the result. :P
Comment has been collapsed.
Good point, I have only a cookie to give but take it as reward, eat it wisely.
Comment has been collapsed.
That's pretty much what I meant. I would have entered that Rocket League one but I have no interest in those other games (or already had it in the case of Memoria) so would never see them.
Comment has been collapsed.
try to put giveaway link not to giveaway description, but in giveaway comments section. U will be surprised
Comment has been collapsed.
Interesting stuff, but the only problem with this analysis, as mentioned by Southrobin, is we don't know the proportion of people who use a browser extension allowing to enter GAs right from the listing page. Is it 10%, in which case you did catch many bots, or is it 90%, in which case what you see here is just people using such an extension?
About the entrances in TEW2010, isn't this a removed (hence undetected) game? Possibly a bunch of entrants already owned it.
Comment has been collapsed.
TEW2010 wasn't actually part of the linked GAs, it was a fully public level 0 GA I put up as a stand alone. Just found it interesting that it got almost as many entrants as RL. Don't get me wrong, this test definitely doesn't show us a guarantee of bots in action, it just shows an interesting anomaly at work, whether it being people not seeing or ignoring descriptions (I tried to take the ignoring option out by making the invite only list huge, it would have been hard to make it stand out more, seriously), people using scripts, bots, skynet absorbing games into its steam library, whatever.
At minimum it shows that working hard on a good description for level 0 GAs isn't a productive use of time! haha
Comment has been collapsed.
What is of more interest though is the GA of Total Extreme Wrestling 2010, a $1 bundle giveaway game which I'm fairly sure a huge percentage of members would have on their filter.
Are you sure about that? Would be interesting to see actual numbers on this. But my assumption is that most people don't filter games at all, maybe except for DLC.
Comment has been collapsed.
I currently have 6,950 games filtered, I love the feature. Makes it so much easier to see games I am actually interested in winning and playing and I don't miss them in the flood of games I have no interest in. Since it is a one click and confirmation click option I presumed plenty of people would use it, some I guess just blacklist the GA creator instead but that can be very counter productive. I know there are settings to stop seeing games of certain values etc but that is also counter productive, there are good games that are cheap, as there are bad games that are expensive!
Comment has been collapsed.
And I currently have 254, mostly dlc (useless soundtracks stuff) and removed games I already have. Maybe you like the feature but it doesn't mean people do. Most people are lazy and won't take the time to filter thousands and thousands games (they don't even take time to read description after all...)
I think KillingArts is more than right.
Comment has been collapsed.
Consider the fact, that a lot of these entries will not be from actual bots, but from people who use scripts to modify SG UI and for example add enter button to main page - they will be entering GAs without ability to read description. I still do find this attitude shitty, because the least you can do for someone who gives you a game for free is to see whether they have something important to communicate in GA, but it's not even close to levels of shittiness of running a 24/7 bot.
Comment has been collapsed.
Shitty people are a part of life really though haha
Comment has been collapsed.
indeed, but what I wanted to point out is that your experiment may be faulty, to be exact your conclusions may be, as there are other more likely reasons for people not noticing your 120 links ;) Were there some bots in the pool? For sure! But as much (if not more) people use ESGST as run bots, and they will all be misinterpret in process you interpret data ;)
Comment has been collapsed.
I took part in this experiment ;) I was interested in Hearts of Iron IV but not Rocket League (two different genres) and all the Hidden Object Games from the bundle I already had. I can't see all the giveaways in the screenshot so maybe I did enter something. Surely not Rocket League though :D
I think better way to check is what I've seen done by someone else. They make public giveaways of bundled games and have a long description about 'getting more chances by different means'. One of the means is reading description and there is a link to another giveaway with the SAME game. That's very interesting to see how many people, obviously interested in the public giveaway, enter the private one (with usually more copies!).
EDIT: Silly me, I didn't think of just checking your profile xD I did enter another game (Battles of Ancient World) and I think I wanted to enter The Mahjong Huntress but somehow didn't (my comment suggested I have entered more than 1 hidden GA but oh well xD). I just won that game so all good, just skewed your results by not entering :P
Comment has been collapsed.
Same GA inside a GA was the plan for the next test, definitely would be far more effective that this one. This one sounded great in my head and turned to mud in the real world, but I put the effort in so I thought might as well put the results up!
Comment has been collapsed.
Don't worry about this experiment! I will repeat an anecdote I've heard myself: A student read a lot of books but not all of them had something to do with the thesis. The professor said: Don't worry, cite them as well, stating, they did not cover your topic. That way you prove you tried different approaches. :D
Comment has been collapsed.
ESGST doesn't just let you enter from the main page, but also displays the description if there is one.
It filters boring descriptions (like good luck), but would show yours
Comment has been collapsed.
After a giveaway chain like this I always wonder what is the percentage of people not reading descriptions, bots, and people who already have the second game/are not interested in it.
An experiment like this would probably go much better if the same game was linked in the description(that solves the problem of some entrants not being interested in one of the games)
Probably something among those lines would do the trick:
- public GA
- private GA linked from GA1(bot with a crawler would still get here)
- private GA linked from GA2 with tempered link something among /a-b-c-d-e/
- (that's probably going too far) private GA linked from GA3(link modified and hidden from the description, but the description itself saying that a link is present)
The only problem is that using a game bundled too many times obviously will filter out some of the genuine entrants as they would already have the game thus shifting the result.
Edit: I once tried with different games. 1st containing a full link to 2nd one and the second one linking the 3rd one through a "." sign. The results were quite disturbing
Comment has been collapsed.
Same GA inside a GA was the plan for the next test, definitely would be far more effective that this one. This one sounded great in my head and turned to mud in the real world, but I put the effort in so I thought might as well put the results up!
I like the modified link idea though, to make sure it is a human working out the correct link.
Comment has been collapsed.
The last time I did a test like this(the one from the edit part) I ended with a ratio of 757:6:2. The 3 games were different bundled titles, but still the result was so bad that I dropped the subject and decided to give the benefit of doubt.
I believe that this time HoI and Memoria were simply good enough to attract more human entries, as in a similar experiment based on some mediocre games I've seen results somewhere in between. The better the game the more people who are genuinely interested in it enter the GA and read the description.
Comment has been collapsed.
tbh o have no idea how this suppose to show % of bots. Dollar bundle is somewhat curse of this site. So many people used this bundle as lvl booster that i think most SG users know those tittles by heart.
Now if i join GA for game I want - eg Memoria, then I see long list of GA links in describtion, and most of those links are from dollar bundle then :
-
- im not gonna join those additional giveaways because i dont want those games, even if i wanted Memoria.
-
- even if theres better game hidden between those links i can easly miss it, because when i see so many "dollar bundle" tittles i will assume that all games are from it (plus i simply may not be interested in this one "better" game).
Theres also no difference in adding 12 "hidden" GA and 120, you wont get better results this way. More - making more GA gives worse results because people might not have points to join so many GA.
Try to to hide GA of same game youre giving away in "starter GA". If you wanted this game in GA nr 1, then you should want same game in GA nr 2.
I once won Hollow Knight from GA like this. Same game, same lvl, same time - open GA had around 5500 entries, GA "hidden" in describtion - 128.
But still, this shows how many people pay attention to describtion, not really bots.
Comment has been collapsed.
Same GA inside a GA was the plan for the next test, definitely would be far more effective that this one. This one sounded great in my head and turned to mud in the real world, but I put the effort in so I thought might as well put the results up!
Comment has been collapsed.
Well at least You made some people happy. I always feel better when i find "hidden" GA ^^
Comment has been collapsed.
Btw if youre still willing to do tests like this i would suggest 2 giveaways (or rather 4).
- 1st open GA with clear link in describtion for hidden GA, and maybe caps note "ANOTHER GIVEAWAY"
- 2nd GA, again open, with link to another hidden GA, but this one hidden in longer describtion, with other links (to YouTube, twitch, devianart, etc), with text edition (i mean GA hidden under word "Giveaway" not whole, visable link).
All for same game. Point is that bots will (in theory) ignore all describtions, people will ignore longer ones. :p
Comment has been collapsed.
i ignore longer ones, that true. but when im busy, i just click enter then close the tab. and i bet not iam the only one. usualy no description, or 70% self/group advertisment.
Comment has been collapsed.
Oh I never thought this would be fully accurate method, theres so many stupid, little factors.
Aside of people ignoring describtion, assumption that bots won't find additional links might be false.
It's just best test I can think of and I believe it could give you at least somewhat accurate numbers.
Comment has been collapsed.
For your test, every user must have the same games. This "test" say you only how many user already have rocket league and how many need Hearts of Iron. -.-
Comment has been collapsed.
LOL I remember that I did not open this giveaway seeing 3k+ entries xD
Damn if I knew they were a treasure of 120 links inside...
Comment has been collapsed.
I directly dont want that game so I didn't enter to that giveaway, so lost the chance to enter the rest lul ;/
Comment has been collapsed.
I like those test, but damn did I miss this one?!
Thanks anyway for doing those!
Comment has been collapsed.
So the results are in for the second bot test. Disclaimer: Doing this for laughs, not as a legitimate test for bots, or even for people that don't notice 120 lines of giveaways.
I set up a level 0 public giveaway of 3 games and a dlc, Memoria, Memoria sound track, Insurgency and Hearts of Iron IV. In the description of those 4 giveaways I added a link to around 120 other invite only level 0 giveaways. 99% of those giveaways were from $1 bundles but I also included Rocket League so there would be at least something of quality there of interest. Then I also added one last level 0 giveaway of Total Extreme Wrestling 2010 a few hours later to the level 0 public giveaways. The links were in the main 4 giveaways as such:
https://i.imgur.com/7XBXWDK.jpg
So if you can see a description that was VERY hard to miss, 120 or so lines of giveaways. The giveaways were all set for around 36 hours approximately, TEW 2010 was a few hours less.
This was the result of entries into the competitions, with the invite only $1 bundle games getting between around 40-120 entries:
https://i.imgur.com/4hSzNvR.jpg
Obviously this test doesn't show obvious bot entries into tier 0 giveaways (and hence missing the linked giveaways). But it does show some interesting points, obviously the bundle games and Rocket League were only able to be seen by people that were interested in one of those 4 original giveaways and if we say it was limited to Hearts of Iron IV then from 3500 entries 484 entered Rocket League. That's not too shabby really and although I would expect more it still doesn't really show anything substantial.
What is of more interest though is the GA of Total Extreme Wrestling 2010, a $1 bundle giveaway game which I'm fairly sure a huge percentage of members would have on their filter. The level 0 giveaway got 406 entries, not too far shy of the not very hidden Rocket League, a far higher tier of game, hell it's probably almost considered AAA.
Anyway, draw your own conclusions. I aim to do another test down the track, a far simpler test which will be far more telling one way or the other. This one actually doesn't tell anything as it relied on someone entering one of the original 4 games to also be interested in Rocket League, or one of the other 120 games. I could go through every giveaway and try and find the total number of unique entries but that seriously is too much work. My next test as I said will be far smaller, cheaper and easier to see the results.
Comment has been collapsed.