Well, that's the thing. Nothing is a better study of what DBO is. Because even the tournament hasn't proved anything. Without DBO it used to be very clear who's in the lead, who needs to be focused on etc. With MBO+DBO nothing is clear. They added a whole new dimension of almost randomness. And this fact, in my opinion, speaks for itself. 

Okay so this will more less be a double post with "look at this graph"


 http://puu.sh/kG1Xk/6bf4a8c89b.png

 

here is a super fast and shitty graph from game 4, Super fast game. (5 day game)

top portion of the graph is center upgrades so when you see a dot and a 0 dot on the upper portion that implies a buy out. PB makes the first purhase of cubit at 3.5 and then another purchase at 4.5  (around just after 4.0 but the 12 hour snapshot missed the needed data)



As you can see, PH makes the first purchase of cubit and then quickly makes the second purchase of blues. 


So I did a tick my tick data grab of a specific time frame of the stock play around PB buy out of cubit. 

 

http://puu.sh/kG1WZ/f343212af6.png

 

the bar's are necessary cost to buy out cubit. (the graph is from cubits positions) so the data shown is how much his enemies need to spent to buy him out. the lines are each players cash/resource.

 

 

Okay so what we can see here is how quickly PH bought out Cubit (less than 1 day of ticks)

If we could have access to this level of data for an entire game it would be very very easy to run stock graphics.

I would be able to stack 4 of these together into 1 visualization into "stock defense" as well as showing resources comparisons and other such graphics for a total of 2 graphics.

 

 

Anyway its been a long night, so if this post makes no sense let me know I did this rather quickly. If I can make 4 of those stock graphics into 1 visuaization I think I would be on something super solid

 

(like data graphic for soren, Pb, Blues, Cubit) each one getting their own "section" graphic. 

 

yea, I do think there is a lack of "leader" metrics. But I do think that the finals (5 games) of DBO are the best of the best when it comes to examining the mechanics. 
however I am not sure they are tied specifically to DBO or MBO, But I can see why they would be for sure. 

I need more experience, to say if it applies to Non DBO as well. 

There is no easy to formulate correlation between skill level and performance in one particular game. Or 5 particular games. In one game Pb became irrelevant and got overrun when he reached level 5 and the very next game he gained a stupid lead which allowed him to make a buy at level 3 and secure his win. Cubit got a DBO and stalled because of it, while in the same game Blues got a DBO on him, and even though his investment had been less profitable (higher price + triple life support costs for 1 person worth of claims), he was able to surpass Soren and take the match. Black market, random events, resource gen and so on and so forth make it impossible to generalize. I think, getting a decently clear picture of what went down in one  particular game would be a great achievement and should be the goal and the purpose of all graphs and charts. Drawing general conclusions is beyond unrealistic due to the very nature of OTC. Regardless of sample size or its 'quality'. 

I agree on that point,

 

I do think if any games were to be looked at, finding the best match up of players is beneficial. and if no proper sample can be reached, then looking at top games again is most beneficial. 





 I will say on this point "Drawing general conclusions is beyond unrealistic due to the very nature of OTC. Regardless of sample size or its 'quality'. "

That given enough data it can be ;) 

That is quite literally what the point of big data analysis is for But of course the question of "how big" and "possible to even reach that point" are the hardest issues. 


As for graphics breaking down a match, yea its a fun puzzle, I'm coming closer, Much much closer, but most of the solutions involve data tracking that is not feasible from pausing the game and writing down numbers. 1 match is not fesible to correctly graph from hand.

There's a difference between data analysis and knowledge discovery. Again, don't get this wrong, we all appreciate your work. But personally, I think we will benefit the most from it, if you focus on data visualization at this point. Good luck and thanks again for your efforts!

deleted pointless post.

Might as well just stop.

Ill post a new thread with the newer prototype.  Ill just make the visualization and let the data speak for itself, arguing over something doesn't prove it.