Results 1 to 4 of 4

Thread: Clutch as hazardous waste product of the game-theoretical industry

  1. #1
    Join Date
    Jul 2011
    Location
    Montreal
    Posts
    342

    Clutch as hazardous waste product of the game-theoretical industry

    ok so we know that there's no such thing as a hidden clutch gene in CSFBL, unless Reds is simply a sick & damaged individual who would play with our hearts, like a very twisted Joker...

    however, I wonder if it's possible that in a functional sense certain players' abnormal clutch performances, whether positive or negative, aren't the result of changing fate (or random variation, depending on your frame of reference with respect to the universe) but rather the distribution of ratings among pitching roles occurring in definite patterns in certain leagues - either because of widespread strategic tendencies (e.g. players rostering mostly FI- dominant pitchers)" or general unconscious biases like preferring pitcters with high PO as closers, or slotting in pitchers with low SY or CO as relievers rather than starters, and on a more granular level maybe even preferring them as either set-up/middle relief... there are a number of different possibilities that could depend on things like overall skill level of an ownership group and and historical knowledge they've been bequeathed that's become unconscious orthodoxy about either the game-theoretical state or the game engine.

    Since we know that to simulate any one event the game takes into account all relevant ratings of all players on the field & all stadium dimensions and basically just spits out a probability grid onto which the die is thrown, and so we know that specific ratings can interact in specific ways with other existing ratings (by which I mean ratings not restricted only to the player himself but any rating on any player, since we know the game itself determines the outcome using all ratings interacting simultaneously as one big meta-player) and not others. For example, one player's RA affects not the probability of him "getting" to some simulated ball but rather the overall number of chances he receives which necessarily impacts the opportunities of others while keeping the proportion of balls in play constant, like it actually determines the so-called "path" the ball will travel when it's an out. We could maybe imagine something like DI existing in an inverse relationship with CO that affects only the one outcome of walks and doesn't change any other probabilities. In fact we know from the help pages that neither DI and CO affect anything except how walks vs. strikeouts split a piece of the probability pie whose size stays constant (based on other ratings) while the K/BB ratio changes*. (This isn't really true of real baseball because a skill like discipline is is an after the fact abstraction of a concrete process that's as deeply involved in the process of hitting as it is in drawing walks [namely the ability to discern the path of the ball], so the fact that clutch doesn't exist in real baseball is irrelevant to whether it exists in CSFBL with its discrete ratings that we know interact with other ratings in a single, if unknown, mathematical function)

    If this is taken to be the case, in a given league where the closers/set-up/middle relievers have an average CO that's significantly below the average starting pitching CO (innings for each group weighted I guess by the propensities of the owners in that league to turn Always Use Closer on/off & the secret probabilities of the reliever-choice algorithm), players with high DI relative to the group of players who produce similarly with the bat in this league would essentially be "clutch" players (in terms of Close & Late, not RISP - I can't think of a theoretical sense in which RISP variation isn't just noise). That's the most obvious possibilty and is obviously league dependent, but it's not hard to imagine that this could be extended to cases where other ratings are unevenly distributed among pitching roles, just with more complex interactions whose details you wouldn't even have to understand in order to detect their pattern.

    anyway, I started with a question ,but I think I ended up with an answer instead. i would be glad in my heart if anybody who actually read this crumbling, dusty, decrepit-***, ******* Western Wall of text profited from these findings. if I made a mistake in my reasoning, well, I enjoyed thinking this through, which means I'm still based,


    *afterthought: this might not be true, I assumed that since avg. changes by 0.02 points going from 50-100 DI with everything else constant pretty much only strikeouts were being affected, but I forgot that that because increasing walks diminishes ABs it's also reducing the probability of balls in play, and so it's interacting with ratings beyond just CO and vice-versa). though as i noted this doesn't really matter, even if outcomes are dependent on multiple ratings interacting you could just brute force search for leagues with different tendencies for [X] combination (and distribution) of pitcher ratings in certain roles and see if there are Close and Late outliers in [Y] combination/distribution of hitter ratings, with the number of ratings involved and the actual way they interact to create a single probabilistic outcome being unimportant. i can imagine exactly how you would automate this but it's beyond my technical abilities
    Last edited by V.I. Incandenza; April 30, 2019 at 02:17 PM.

  2. #2
    Join Date
    Jul 2011
    Location
    Montreal
    Posts
    342

    Re: Clutch as hazardous waste product of the game-theoretical industry

    more generally this kind of thing could also extend to playoff outcomes. what looks like the random variation of a 120-win team losing to 85-win team could actually be determined partly by particular rating interactions.. this might even be more extreme because individual owners' roster construction tendencies could be way more distinct from each other's than the averaged tendencies of 24 owners' from another 24
    Last edited by V.I. Incandenza; April 29, 2019 at 12:58 AM.

  3. #3
    Join Date
    Nov 2013
    Posts
    1,504

    Re: Clutch as hazardous waste product of the game-theoretical industry

    I don't know, CO-specialist on my teams usually actually end up in the bullpen, because they become garbage the 2nd time through (or rather they tend to be ineffectual SPs with few exceptions). Edit: Not that CO isn't valuable, but I find it the least valuable of the 4. As in, it is the rating that most needs a corresponding rating from the other 3 to be good, to be an effective pitcher.
    Last edited by DTCT; May 1, 2019 at 03:21 PM.

  4. #4
    Join Date
    Jul 2011
    Location
    Montreal
    Posts
    342

    Re: Clutch as hazardous waste product of the game-theoretical industry

    yeah that's exactly the tendency I'm talkin about. it might be exploitable, stay tuned

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •