User blog:GeoFruck/Discussion on proof of Dino Fusion results

Hello All,

I've decided to create this in order to arrange a place for people to discuss this topic. Please understand, this is not an attempt to forward my point of view, or discredit someone else'. This is simply a forum for an honest and anylitical debate on whether proof is needed.

That being said, I obviously have an opinion, so I will start it out.

I do not think that proof is necessary in order to discover a pattern. If decent percentage of the users of this site start contributing, which I would expect, then we will get a lot of reports daily. Assuming that most of the people reporting are genuinely interested in helping to build this database with accurate data, then we will eventually end up with the desired "Guide on Fusing Dinos". Now this is an assumption, so I will move forward by saying that I believe most people in general TRY to be honest. Next, people who play TV are going to be the vast majority of the people that even seek out and visit this site. And, we TV players WANT to know how to get certain dinos. Therefore, the vast majority of the reported combinations will be accurate.

Addressing faulty data. Yes, there may be the occasional mis-remembering of an intentionally honest report. There may also be people that just want to report false combinations for whatever reason. I maintain that from a statistical stand point, this false data will hardly matter. For example, right now we probably get between 15-30 reports a day (total guestimation, but it doesn't matter, read on). Let's say that out of 20 reports in a day, 2 people report false data. Now this is 10% of the data, so I'd say it's a little annoying, but not terribly damaging. So, let's say we get 50 reports the next day, and the two from yesterday were just messing around and come back to report 2 more false combos. Now, it's only 4% of the data.

So let's assume that someone, or a few people are trying to skew the data. How many false reports would need to be made? Well, we will have to pick a percentage of false data that would be picked up as accurate. Assuming, there are different results for the same combo, which I think is safe, there might be 3-4 levels of "rarity", let's say something around 45%, 25%, 15%, 10%, and a remaining 5% of just totally weird results like the tier jump that has been reported. We will take the 5% for arguments sake.

If we take a conservative number of report per week, say 100 for example, This would be about 5200 reports a year. 5% of this is 240 reports. This means that one person would have to make 240 intentional false reports out of 365 days. That's a pretty decent amount of dedication. Now let's say that this movement gets going and we do around 3000 reports a week, which is a little less than 50 a day. This makes the annual reports 156,000 reports, or data points. 5% of this is 7,800, which starts to show my point. This would be around 21 posts a day for an entire year. At this point, I'm sure an admin or someone would pick up on someone spamming the site.

Anyway, that's just my opinion, and reasoning behind it. Please comment, though, because if we DO need proof, then how to do that becomes another topic to discuss. Not to mention whether that "proof" needs to be further verified or "proved", and on, and on....