@Collecter a Just looked up the numbers again: http://www.archive.themanadrain.com/index.php?topic=46058.0
So 2 games out of 20 is about the correct rate with only 2 powders. That rate falls to something like 2 out of 30 with the full set of powders.

No one has done the explicit hand size calculations though, which are arguably just as important. Or the rate of drawing double bazaar, which is increased with powders.

@vaughnbros Looks like there is still a lot I will have to look up and actually learn about maximizing dredge win %.

@Collecter ignore all that math and just run good. It's way easier, trust me.

@Collecter You don't need to know all the math, but it certainly doesn't hurt! I am extremely adverse to mull to oblivion as I've won exactly 2 games ever, out of a few hundred, under such circumstance.

How did you guys deal with draws, especially non intentional draws e.g. going to time.

I am trying to do a similar table for our events. My initial though was for it to count as a loss for both parties but for completeness sake I thought I would defer to what you did.

@VintageFamiliar For MTGO there are no draws so we didn't have to deal with it. We did, however, remove all wins resulting from byes. For the NYSE we stripped out all the draws. For legacy you could argue that that makes miracles look better than it is because it is so prone to going to time. In vintage I don't think there's an archetype like this, and that deflating the win rate of archetypes of people who draw into top8 or are slow doesn't make sense.

@VintageFamiliar What @diophan said.

The way we have our spreadsheets written is that they deal exclusively in wins - The number of match wins Shops has against Gush is the number of match losses Gush has against Shops and so counting losses is somewhat redundant. Doing this, I guess Draws would count as losses for both parties, but the net result is that the match isn't counted as no winner was determined.

If you were to count wins and losses separately and include draws as losses you would depress the overall win percentage of the field (using our method, the number of wins is always equal to the number of losses so the overall win % will be 50% - counting draws as losses means there will be more losses than wins and the overall win percentage will be less as a result). This will also disproportionately affect certain archetypes as Ryan mentioned. There isn't one correct way to do statistical analysis, though. Just whatever you do, I would mention it.