You really need to do a uni course in regression analysis to understand.
We plot all the comp scores and battle scores. It looks like this:
Once you've plotted the results, you want to find the line of best fit. You need an intercept and a slope for that. You run a regression (what I've done in the excel sheets) and you get your line of best fit.
You then do t-stat on the slope to see if it's significantly different from zero. If it is not significantly different from zero, then we cannot say it is not zero. In other words, it could very well be zero, in which case comp has no effect on battle.