buddha: Flooring

[buddha]

I understand flooring is the preferred method for calculating the TC. Why? Is it because it tends to conservatively underestimate your advantage? But if the indexes are also generated by flooring, doesn't this offset this under-estimation?

[Don Schlesinger]

> I understand flooring is the preferred method for
> calculating the TC. Why?

Truncating double-counts the zero interval (from the right and then again from the left), making that over 40% of all counts in shoe games. Flooring is "neater." But, it performs similarly to rounding.

> Is it because it tends to
> conservatively underestimate your advantage?

No, that's not it.

> But if
> the indexes are also generated by flooring, doesn't
> this offset this under-estimation?

Yes, the idea is to always pair up the methodologies for generating the indices and then using them. Flooring actually makes the edges for a specified positive TC higher than when rounding, but the frequencies are then lower.

Don