very good point

i think a big way to think of it is every time people talk about widening the net in hockey, they point out that there are 6.3 goals per game but 1.7 hits of the post or crossbar - so simply expanding the net by an inch would increase the amount of goals dramatically - probably even more so than that data implies, because it'd mean bigger holes for players to target so instead of trying to tuck the puck in a gap in the pads, they now have more wide open space to target - as let's be clear, when they hit the post, that's the sign of a miss that missed closely not that they targeted that

i think the best route is under the new interpretation, to shrink the strike zone by half a baseball or so - but maybe this is just fancy play syndrome?

If I binked the powerball, I’d visit every ballpark.

Big enough powerball, I’d have a small entourage with me.

Send PM to apply for future possible entourage.

Goal post hits could be misleading though. Technically they are hitting something the goalie doesn't need to cover, so his positioning would be different if the net was an inch bigger. Goalies have spent thousands of hours in an exact goal, so they know exactly what they're covering and not covering instinctively to some extent.

good point

Imagine if Wemby was a Hockey Goalie.

Sherzer back with the Jays for 1 more year!

Maybe they also see his presence alone being helpful, and maybe some potential to move into coaching???

3M seems like nothing though. Kind of surprised he signed for that. One might think he could get 5 to 8M for one year from a few teams.

Maybe there are some bonuses, or maybe he just wants to be with the Jays again.

$10 million in incentives. He wanted to stay with Jays.

I have imagined it, and I think he would be terrible even if he were a great skater. I'm guessing that a 7'6 goalie would get five-holed to death.

will probably format like aids but figured worth sharing - never did this before - the juice was never worth the squeeze before, but with ai coding now it's pretty simple to do while inebriated and watching something with a lot of back and forth and iteration - which i actually enjoy

found fangraphs has an interesting feature where you can input your specific league settings to get a revised rankings for your league - but they are dumb about implementation - they apply position scarcity as a flat bonus - so in a 12 team league that does 1 catcher each - so we can expect about 12-15 catchers drafted - they'll still give the catcher premium to the 20th best catcher - it's really dumb, we should only be applying it to the catchers who are good enough to be rostered if they weren't catchers (which is few) and then in minor increments - ie alejandro kirk isn't very good, but you can do worse, much worse at catcher - so he's worth reaching for a little bit despite being awful

this specific ranking system is for me own league - standard 5x5 roto league, 12 teams, deep rosters & 2C $285 budget

it overvalues rp but maybe they are worth more than we normally value them - i didn't apply any scarcity metrics to pitchers because they aren't scarce and good relievers are scarce by default being the only ones to produce saves

figured people may be interested - fg is what fangraphs assumes is the auction value, fp is what fantasy pros believes is but for standard leagues with 280 budget (options were limited for free version and lol at paying for their trash)

hitter projections are sourced from thebatx, pitchers from atc - but you can mix and match and use any and you'll be fine, they don't really disagree much anyway - fantasypros is just crowdsourcing what randos think but is nice because it does reflect market sentiments and a lot of people draft off that as well - so it's good for a sanity check

edit: it was aids - cutting off long tail and going to post image again

nice, that one works

still needs some minor work, as you can see, despite shohei being a better option than judge, we slightly favor judge because he gives you more options to work with whereas shohei clogs up your dh start from the start which is a horrendous outcome for other players but possibly not enough to to warrant judge>ohtani more so mid level dh vs mid level of

also got some issues elsewhere as our position scarcity is still very much a work in progress - but a lot of that is likely driven by the fact that this league has 5 OF

here's it for a standard yahoo league - so there's clearly some gaps out there that need to be fixed

I have no idea what’s going on.

- Towelie Mod

i'm trying to build something that adjust player rankings based on league specifics - it's clearly not working well

Is what you’re saying that it doesn’t have enough Tegridy?

Another example of why I'll never do fantasy baseball ever again. **** takes so much time. I'll stick too hunchcapping round robins every day.

txdome,


who do u have winning it all this year?

if it's any consolidation - nobody in any of my leagues ever believes it whenever they find out what i do for a living

precisely because there's so much variance that spending hours on this kind of thing only grants a slim edge in something with so much variance that the edge is basically noise - there are good drafts and there are bad drafts - but it's much like winning a tourney, you need a lot of stuff to go right for you

the real edge in fantasy sports is grinding the waiver wire, making the big pickups before others even know about whatever caused that guy to suddenly rise in value and i'm terrible about that because although i have my finger on the pulse for that stuff for dfs and sportsbetting, that's where i also have tunnel vision - there's a bunch 2p2 leagues i signed up for, paid a few hundred to enter, did the draft and then promptly forgot about because all my sports bandwidth was focused on the active stuff

i do this now because ai greatly simplifies the process and makes it more of a "no that clearly didn't work, let's try this adjustment instead" and then it does all the legwork and go back to watching tentacle porn until i get a notification that it's ready for me to test and give feedback and that's a fun process

i doubt this will give me much edge, and any edge it will give me i'll lose because i'm going to forget to set the lineup quite a bit

was in a 2p2 baseball league last year and did pretty well because for once i finally stayed somewhat active but the format rewarded streaming pitchers and while i abused that system more than most, there were some people who stayed up until 3 am each night to get first pick off waivers so any edge i may have had going into the playoffs was soon toast once i was facing teams that were willing to go the extra mile that i wasn't and i got smoked in the playoffs

i will absolutely crush the general public in a "how many passing yards will stafford have today" game because the variables are known and variance is reduce - even if he misses half the season due to injury, of those 8/9 games he plays, we'll know ahead of time and we'll know his updated injury status and account for it - only real wildcard is if the big injury happens mid game - but then it's just 1 game getting nuked vs an entire season

but... if you want me to tell you if stafford is going to be a top 5 qb next year then your guess is as good as mine and simply looking at the preset rankings will get people 95% of the way to what a sophisticated season long projection model can produce

Give me couple weeks. Now It's obv doyer til injuries pop up. No injuries=Braves.

I wish I could be as optimistic as you about the Braves, but I can’t get there, at least not yet.

He is trying to recreate Rotolab using AI.

you're forgetting that i'm also arrogant so i think whatever i ultimately produce will be better

ie there are flaws galore in fangraphs version of it despite that it is a site centered around data analytics

There is a flaw that is common to all models that produce auction values for fantasy baseball, but it's complicated to solve for only a modest amount of improvement, so I've never bothered to try and do it.

Generally speaking, all the models account for playing time risk, injury risk, and skill instability in the same way. They try and figure out what stat line would look like for a player with that skill and risk profile if you did the moral equivalent of a Monte Carlo simulation. In other words, they say, "what would the blended average stat line for this player be if we had the opportunity to see X interations of the 2026 season play out?" Then they back into a dollar value based on that "blended average" stat line. The problem with this approach is that it implicitly assumes that the the value implied by the blended average stat line is equal to the blended average value to your fantasy team of those X iterations. And when it comes to your actual performance in fantasy leagues year over year, I'm 99.9% sure that isn't the case.