People "exploiting" the CV system with the Click Fusion Bundle might want to look at this. [Updated]

Open

First and foremost, for those too quick to jump into conclusions, I'll put this at the top.

I do not advocate min-maxing to obtain as many games as possible and giving the least.

Some people have mistakenly thought this. There's a different between reporting something and endorsing it. I believe this should be obvious, as we all know that the strategy to follow in that case is to remain Level 0 forever and ever. Which is not my case. My intention is to give and take, staying as close to 1:1 ratio as possible. However, as wins are randomised and gibs are reliant on other variables (if there's a bundle I really like to buy, how many of those games are not appealing to me or my close friends, if they already own those, their mean response time, etc) there's a gap.

I will address that gap in the following week.

However I want to clarify something.

I have never entered a GA of a game I did not intend to play.

i.e. I have never entered a GA with the intention of idling a game for the cards, or to get +1 to my steam account or whatever other reason. I only enter GAs of games I actually want to , and will, play. I thought that would be the case with everyone else.

Why did you write this?

I didn't want to stir the pot more than it is. I just saw this discussion and https://www.steamgifts.com/discussion/X5teN/treat-clickteam-fusion-as-a-free-game and piqued my interest.

As many have pointed out, it's absurd to treat giveaways as an investment. That's why I was confused when I saw all the drama over people cheating by making GAs.

My intention was to mellow the discussion about trying to maximise something that, due to its very nature, is an altruistic act.

I also had to iron shirts but I didn't want to. Pending household chores are a hell of a fuel.

How many blacklists you got for posting this?

6 times more than whitelists. So I guess that the right thing to do is not to post ever again. Or at least until I fix my ratio. Or maybe some of them hate math or dislike that I didn't take into account all variables, which I don't think I can do anyway (See below). Oh, well. At least I had some fun, which is what counts in the end.

And now with the actual mathcraft.

The Analysis

Simplifying Hypotheses

Given the data obtained from the site, it is assumed that everyone starts with their September 2016 level and enters all the GAs their level allows them during the whole period T for which there is data available.

You may argue that it's an abstraction and too coarse to return any accurate data, and you would be right. Many variables are not taken into account, as many have pointed out in the comments. However, it allows us to compare data from one level against each others.

Many effects have been left out, like the points limitation, the fact that people would actually own some of the given away games, and so on. The reason they were not included in the analysis is because I wanted a quick estimation to get some data to validate what we all intuitively know and because I didn't think in a way to implement said variables. If you come up with a way to add them to the calculations, I would be most pleased to see it, so ring me, please.

One of the most glaring inaccuracies I incurred in is taking into account so many Level 0 users, as most of time are allegedly inactive. This is thankfully minimised by the fact that the number of Level 0 users don't influence the data in the tiers beyond 0, so the muddling effect is contained.

The CV requirement for level 0 users is 0, but it I used a near zero value instead, to avoid dividing by zero, which is the root of all evil.

How things were calculated (see spreadsheet as well).

WCi is calculated by taking into account the total numbers of GAs with their level restrictions and the level of the user.

e.g.

WC0 = (GAs available to users with Level 0) / (Users with level 0 or more) = (Level 0 GAs) / (All SG users) = 1222674 / 946092 = 129.23%

What does that mean in layman's terms? If you started as a Level 0 user in the beginning of SG and joined every GA available to your level, and everybody else also did, by now you should have won a game. This doesn't tell us much of use, but we use this Win Chance as a base, so we can compare better the other WCi by turning them into WC / WCLv0. There lies its true usefulness.

For the rest of WCi it goes like this:

WC1 = WC0 +((Level 1 GAs)/(All SG users except Level 0))

WC2 = WC1 +((Level 2 GAs)/(All SG users except Level 0 and Level 1))

...

And so on.

This way only the users that can actually enter a GA given their level relative to the GA restriction are taken into account.

(WC)nd / Cvreq

The CVreq has been extracted from Kiru's post [2]. It does not take into account actual money spent, but the CV value that is associated to a GA. Somebody could add a column making a correspondence between "Site CV" and "Real World Dollars", but it wasn't my initial intention and it might have muddied the results even more.

I love spreadsheets! Mind if I pick up your model and improve it?

You are most welcome to do so. As it has been mentioned, there's plenty of room for improvement. I will drop the ball now as I really need to get stuff done before Sunday's over.

OP before the last edit

I've done some math and "exploiting" the system that way doesn't make any sense.

The chance of winning over a given period divided by the CV required to reach the level you're at diminishes as you get a higher level. i.e. You get a smaller reward per CV. Yes, if you can get cheap CV with a bundle like the Click Fusion thingy more power to you, but that doesn't change this fact, as (WCi)nd / Cvreq,i has been "non dimensionalised" with CVreq,i.

To sum it up: People trying to be smart and cheat the system are just actually getting less bang for the buck than the regular Level 0 leecher or are just after brownie points.

See this table.

T Is the reference period of time. The time period for which there is data available.
WCi is the chance of winning of a Level "i" user over a period of time (T).
WC0 WC of a Level 0 user over the period T.
(WC)nd is the chance of winning of a level "i" user over a period of time (T) made dimensionless with WCLv0
CVreq,i is the minimum CV required to achieve level "i".

i (User level)	WCi	WCi / WC0	(WCi)nd / CVreq,i	CVreq,i
0	1.292	1	100000000	0.00000001
1	3.783	2.927	2.9272	1
2	8.011	6.199	0.2479	25
3	12.88	9.966	0.1993	50
4	19.093	14.774	0.1477	100
5	29.672	22.96	0.0918	250
6	39.579	30.626	0.0613	500
7	52.738	40.808	0.0408	1000
8	63.985	49.511	0.0248	2000
9	70.193	54.315	0.0181	3000
10	80.193	62.053	0.0124	5000

I spent way too much time on this, so I thought I might as well share it with the community so we can collectively get a better bang for my spent time. : )

Edit: Reduced the number of decimals from eleven thousand to just 3.

I used a non-zero value for CVreq,0 to avoid dividing by zero.

Sources:
[1] https://www.steamgifts.com/stats/community/giveaways
[2] https://www.steamgifts.com/discussion/XaCbA/list-of-all-cv-levels-and-their-corresponding-real-cv post, by Kiru.

8 years ago*

Comment has been collapsed.

We all know what I did this Saturday, but what have YOU done today?

View Results

I spent too much time on the Internet.

I played videogames, as it's meant to be.

I spent most of the day on a spreadsheet and trying to get the table formatted right. Point at me and laugh!

T-that's none of your b-business, baka!!

I shot a man in Reno just to see him die.

I've been working on my magnum opus. A game where an ancient evil awakens and the potatonist defeats it.

My answer is not included in any of the above. Assume I had to do chores, or other regular stuff, like classifying a collection of buttplugs by weight in the underwear drawer.

132 Comments

raziel88

I think it is ridiculous that one set of programs has spawned so many topics, and that people think less of some giveaways. I mean a lot of the giveaways on this site are generated from the same concept.
Then looking at the "analysis", I suspect it doesn't take into many factors, and simply looking at Chance of winning over a period of time over CV req, seems wrong. CV is a one time investment (although it decays irregularly due to price drops, bundling etc), but winning probability stays the same (actually more complicated but let's assume) so it is bound to pay off in terms of expected value over time at least in the lower levels especially if we consider that there is a resource constraint (points-> can't enter everything). Also, I am highly doubtful of how the winning chance is calculated.

Also, everybody no doubt get's less bang for the buck than lv0s because as long as lv0 wins one thing, their bang for the buck is undefined.

8 years ago*