0% chance of doing anything, original prediction had them winning 1 game for the season, currently sitting at 2-0, despite being 29th on the roto rankings.
Post by Rockies GM (Dan) on Apr 22, 2018 20:35:24 GMT -8
I haven't examined these at great detail yet. I slightly modified equations for determining expected R, HR, RBI and SB, which look to be more accurate (at least more accurate to the last 2 years of data)
Post by Rockies GM (Dan) on May 7, 2018 11:09:54 GMT -8
Will be posting an update shortly, but before I do, I want to announce/explain a major shift in the simulation algorithm. Until now, the program has been simulating using one number - an overall win percentage assigned to each team based on ranks in the projected (and already accumulated) statistics.
However, I took on the task of researching historical win% of specific stats. Going back to 2012, I looked at each position in the power rankings each team landed in specific stats, and compared that ranking to the percentage that category was won in any given week. Unsurprisingly, linear regressions fit extremely well.
If anyone is interested, it lends a lot of credence to the volatility of certain stats as opposed to others. Rate stats are won with much less consistency by the best teams in the league at that stat than counting stats (for example, historically, teams that have led the league in the K stat have won that category over 87% of the time on average, whereas the best team in the league at ERA has averaged a touch over 67% of the wins).
So, the newest iteration of the program simulates the 12 individual stat matchups each week and decides the winner based on the outcome.
What does this mean?
Simply put, more accuracy. Being superior in one specific stat could have skewed results, even slightly before. Now, it still helps you in the simulations, but only in the one category.
But like, what does this MEAN?
I don't know. I've noticed the odds on certain teams increase or decrease, but nothing extremely drastic. Perhaps the old system was working just fine. Idk - but more accuracy never hurt anyone.
Post by Rockies GM (Dan) on May 28, 2018 12:11:39 GMT -8
First time I've posted with the new individual category simulation. Comparing the different algorithms, this affects some teams more than others, which is probably a good thing (accuracy to be seen, but it bodes well).
Post by Rockies GM (Dan) on Jul 15, 2018 20:40:42 GMT -8
All-Star Break Update
We might be just over halfway through the MLB season, but here, that's 3/4 of the regular season. So, with the homestretch in sight, let's see where we stand.
Padres GM (Amy): @hollah, that is truly brave work
Mar 11, 2024 5:47:59 GMT -8
Reds GM (Pat H.): Hi, my name is Pat and I'm addicted to fantasy baseball.
Mar 11, 2024 6:26:35 GMT -8
Padres GM (Amy): i tried to quit and we see how that went
Mar 11, 2024 6:27:33 GMT -8
*
Cardinals GM (John C): Quote from Amy: "Just When I Think I'm Out, They Pull Me Back In."
Mar 14, 2024 6:54:31 GMT -8
Reds GM (Pat H.): We will try Round 5 of the draft on Fantrax. You are able to fill your queue with players now. It doesn't start until Round 4 is over.
Mar 14, 2024 7:24:36 GMT -8
Padres GM (Amy): Pretty sure Yankees pick is invalid as Martorella just released
Mar 17, 2024 13:08:03 GMT -8
*