I took a good look at your weighted patience factor calculator and have several problems with it:
Snyder's PF calculation is: ((BlindOffHours*OnLineFactor)^2)*RebuyFactor
where:
BlindOffHours = Number of hours it would take for the world's most patient player to be blinded off
OnLineFactor = 1 for live play; 1.667 for online play (accounts for increased speed of play online)
RebuyFactor = 1 for non-rebuy tournies; 0.8 for rebuy tournies
Your weighted PF calculation is: (BlindOffHours)*(OnlineFactor^2)*(XanaduMultiplier)
where:
XanaduMultiplier = ((StartingBankPerPlayer/(100*NumberOfPlayers) + NumberOfTables)
(see attached edited version of your Weighted PF spreadsheet for clarification/proof)
- Right off the bat, I see that rebuys are not accounted for and the effect of squaring the BlindOffHours is eliminated.
- Increasing the number of players decreases the WPF, which I believe is backwards. This error is masked by the fact that as you increase the number of players, you increase the number of tables as well. If you adjust both together, though, you'll see that the increasing the number of players actually decreases your WPF.
- The StartingBankPerPlayer shouldn't be here at all if you're just trying to adjust the PF based on number of players and players per table. It has the effect of making a 10,000 starting chip tournament look like it has a higher PF than a 100 starting chip tournament, even if the blinds are simply scaled by a factor of 1000.
- The number of players per table effect is backwards. As you decrease the number of players per table, the PF should decrease because you have to pay the blinds more frequently. It's backwards (but again, masked a bit) in your spreadsheet.
That's it for the technical stuff, now for the philosophical:
- I don't believe it makes sense for a tournament's PF to be dependent on the number of players playing in it. It's essentially a measure of the poker tournament's blinds schedule in relation to an AVERAGE player's chips (or any player's starting chips in the case of freeze-out style tournaments). With your method, an ultra-fast blinds schedule would be rated skill level 20 as long as the number of players was high enough.
- The number of players per table should affect the PF (as noted above), but it's very dicey trying to account for it. I actually had it included in my algorithm at one time, but decided to scrap the idea. One big problem is that you don't know how many players there will be at the average table at any given blinds level. And that WOULD be affected by the total number of players in the tournament because you could expect the number to be relatively stable for large tournaments, but for very small tournaments it would vary significantly.
- Snyder's PF and Skill Level calculations were not created in a vacuum. They were created in the context of specific poker tournament strategy advice that he offers in his book for each skill level. While a 1 to 20 range for skill levels sounds good for comparing tournaments, is there really a difference when it comes to strategy between a skill level 14, 15, or 16 tournament?