antifragility

sagefr0g

Well-Known Member
#1
i sorta, ain’t knows what i’m writing about here, just making it up as i go sorta thing, after having read Nassim Taleb’s book, Antifragile, just sort of tinkering around, as the book suggests. been trying to see if the ideas put forth in the book can be useful within the walls of the casino. but i sorta think it’s ok to make this stuff up, since it’s my own gobbly goop and i been sincerely trying to have it make sense to me, as i do have skin in the game, when it comes to casino exploits. so anyway, just saying, take this stuff with a grain of salt.

wrote a bit about antifragility in the links below:

https://www.blackjackinfo.com/community/threads/probably-be-a-good-read.54840/#post-490622

https://www.blackjackinfo.com/community/threads/probably-be-a-good-read.54840/#post-490677

also have written about the pareto principle 80/20 rule in the links below:

https://www.blackjackinfo.com/community/threads/its-all-about-time.54841/#post-490668

and further wrote this excerpt regarding the pareto principle:

also far as pareto principle, as previously stated, from my experience and data far as advantage play types engaged in, it takes on average 17% of 100% previous winnings to win another 100%. so 83%/17% ratio sorta thing on value won and value extended.
this can be visualized looking at this histogram graph depicting win/loss data, win only data & loss only data. it's obviously demonstrative of the pareto principle, imho. link to histogram below:
https://www.blackjackinfo.com/community/attachments/summarygraph-jpg.8985/

anyway i wanted to post an image that i believe sheds some light regarding the existence of antifragile states within the walls of the casino, something that i think Taleb would say shouldn’t exist (due to the artificially, mathematically rule driven set up of games in the casino), but i believe extreme events and antifragility is present in casinos, albeit constrained and limited by those mathematical rule sets, unlike the ‘wild west’ nature of things outside the walls of casinos, where black swan events are surprisingly as common as the birthday paradox (which isn’t really a paradox) is, however seemingly unlikely. the image below also interestingly enough shows a possible link between the pareto principle and antifragility. that link has further been hinted at by seven different games that i’ve analyzed with respect to antifragility. each of those games, when restricted or adjusted to an antifragile state at its transition point have an advantage = 20% (give or take a few percentage points). that 20% advantage occurs at the ‘transition’ point wherein the condition of the game transcends from a robust state to an antifragile state, further adjusting the game condition for ever more antifragility up to the limit for which it can be adjusted imparts an asymptotic rise (smiley face curve, such as mentioned by Nassim Taleb in his book Antifragile) in advantage approaching a 300% advantage or more for the game types analyzed. not meaning too paint to rosey of a picture, these game states do not represent a ‘good’ black swan sort of event (where an advantage would be described as astronomical), although unfortunately they are rare events, especially for those states where the advantage is near 300%, none the less, they do occur. there is a ‘spectrum’ of antifragile states between the 20% advantage state and the upper bound limit > 300%. not bad prospects for the typical advantage play scenario. there is optionality with respect to decision making regarding the making of an antifragile play and the degree to which it is antifragile. a certainty equivalent sort of decision, except there is no downside to influence the decision, sorta thing, other than the options may dissipate, as in opportunity risk.

having essentially nothing to go by other than having read the book, Antifragile by Nassim Taleb, i’ve had to set my own parameters far as what i’m calling, fragile, robust and antifragile, with respect to the games i’ve analyzed. just been kind of going by what makes sense to me, sorta thing.

this is pretty much what i’ve come up with for the games analyzed:

advantage%, best case, worst case expectation & expected value all have variable states and vary, increasing value wise as play is conducted.

a game state is antifragile if it has no downside, that is you can bet away at it and can’t possibly lose money. it also loves volatility, whereas volatility for the game in a robust or fragile state would normally expose you to the possibility of loss, as well as gain, the game when in an antifragile state can actually only yield better results as a result of volatility. the worst case expectation at this point only gets more and more positive in value as play ensues. the best case can always pop up at any time as a result of volitility. both the value of the best case and the value of the worst case expectation become more and more positive under antifragile state conditions as plays are made.

the games that i’ve analyzed can transcend from, worthless, to fragile, to robust and then to an antifragile state.

the transition from robust to antifragile occurs right about the 20% advantage point and when the worst case expectation becomes >= $0.00 .

the transition from fragile to robust occurs right about the 10% thru 11% advantage point and when the expected value = (-1 * worst case expectation) . at this robust point there is money to be made if expected value is realized, volatility can help greatly, but it can also bring about harm in the case that volatility brings about the worst case expectation rather than the expected value , while volatility can bring about the best case wherein substantial gain can be realized. in other words, expected value is what we expect, and hope for, but the other more volatile expectations can occur. fine for the long run, not so much so for the current play as the actual results are up for grabs, sorta thing. that’s the nature of a robust state, it can blow up or be broken, or it can produce as expected. expected value and advantage% for the robust state are lower than those values for the antifragile state.

the fragile state exists when the advantage is > 0 < 8% thru 10% advantage point and when expected value < (-1 * worst case expectation) . at this fragile state there is money to be made if expected value is realized, volatility can help greatly, but it can also bring about harm in the case that volatility brings about the worst case expectation rather than the expected value , while volatility can bring about the best case wherein substantial gain can be realized. similar to the robust state, expected value is what we expect, and hope for, but the other more volatile expectations can occur. fine for the long run, not so much so for the current play as the actual results are up for grabs, sorta thing. that’s the nature of a fragile state, it can blow up or be broken, even more so than the robust state. expected value and advantage% for the fragile state is less than that of the robust state and much less than that of the antifragile state.

a final note of caution, the parameter worst case expectation, is just that, as it is worded, an expectation. any expectation is either an average, or is based upon an average, that fact implies fluctuation may be existent. in other words, even a worst case expectation can allow for worse than worst case expectation results, much, much, much worse. to the tune of a $4,000.00 loss, worse, should the worst of the worst case happen for the above mentioned game condition at the antifragile transition point. that said it’s a winning game in the robust state, lol. essentially for the $4,000.00 loss to happen one of two things would have to happen, either the casino is cheating or the Cubs won the 2016 world series.

well anyway, see the graph below:

i would have shared the data as well, but there is nearly 5,000 rows of data about 13 columns across.

antifragileexample-jpg.8992
 

sagefr0g

Well-Known Member
#2
so as previously mentioned, i had been wondering if there is a link between the pareto principle and the state of antifragility. that was just because i noticed that at the point where the games analyzed (seven of them) it was found that those games transitioned from robust to antifragile consistently right about the players 20% advantage point as the game state transitioned.

think i figured out what’s going on there, and it’s hilarious far as how confusion and pondering came to fruition. first off, is the observation an example of the pareto principle? yep, i’m convinced it is. how about that 20% number, significant? not so much, lol. yes & no, sorta. turns out it was confirmation bias on my part, that by nothing other than luck turned out to be correct, lol. kinda of like my favorite ‘paradox’ to mention, the birthday paradox, that isn’t really a paradox and is just how things are, funny as it may seem, sorta thing. coincidences aren’t so coincidental, lol. and that’s how it was with this 20% number. the particular numbers 20/80 far as the pareto principle are kinda of bogus. it’s just that there are loads and loads of examples where it’s pretty much 20/80 sorta thing, as Pareto himself discovered when he found that 20% of people in Italy owned 80% of the land, sorta thing. far as i think those sorts of examples goes, i believe it’s pretty much of a fairness issue with respect to humanity, sorta thing. it’s about how far our fellow man can stick it to one another and get away with it, sorta thing, lol. read the article (especially the reference “Life isn’t Fair” in the link below:

https://betterexplained.com/articles/understanding-the-pareto-principle-the-8020-rule/

but yeah, life isn’t fair. hmm, what’s that make you think of far as casinos goes? the house edge maybe? well, it turns out, the house edge of the games i analyzed for which it was found the games transitioned from robust to antifragile at a 20% advantage point for the player. guess what the house edge for the games analyzed was? yep, you guessed it, HE = 20% . the advantage player needs to overcome that edge.

so now let’s take the same game, change the house edge to HE = 10%. so now, what’s the transition point gonna be for robust to antifragile, this time? yep, it’s gonna happen at a 10% advantage for the player. so now the pareto ratio is 10/90, sorta thing. essentially a fairness ratio. (albeit, other types of relationships likely exist with respect to the pareto principle)

the difference for which the games analyzed transcend from worthless, to fragile, to robust and then antifragile is huge with respect to changing the original house edge. the edge we need to break through.

same game state as the game graph posted above, except the game is changed from an original 20% HE to a 10% HE. the worthless, fragile and a good bit of the robust section of the game disappears from the graph leaving a huge area of antifragility. just an example of a given parameter (in this case HE) for a game being changed, see image below:

example2he10-antifragile-jpg.8993
 

sagefr0g

Well-Known Member
#4
Southern California said:
Good grief. Will anyone here really wade through all of that and emerge with a still-working brain? :(
lol, sorry sir.
ehhm, you should be glad, to not have been the one writing it. lol.
well, at least you tried reading it, i appreciate that.

essentially, the admittedly verbose, confusing diatribe (poorly written as is my way) is supposed to be about putting oneself in a position, for which no harm can occur, for which volatility is always a good thing, never harmful, and the prospects of profit are relatively enormous, with say respect to the advantage for example (mentioning this example as i'm aware of your interest in blackjack) card counting blackjack has to offer. that's it in a nutshell, nutty as that may sound, lol.

i'm also interested in blackjack. the game has such a low house edge. commonly in the neighborhood of HE = 0.70% or so. that's an incredibly low house edge to overcome. i currently don't know if simply overcoming and matching that circa 0.70% advantage puts one anywhere near an antifragile state, as overcoming the house edge (circa HE=20%) and matching that edge for the players favor does for the other games described in the first two original posts above. i suspect that blackjack does not have a state that can be attained for which an antifragile state can be attained as readily as the game types i analyzed. none the less, the prospect is tantalizing and i believe worthy of at least thinking about. after all thinking doesn't cost much.
 

sagefr0g

Well-Known Member
#5
opps, (learning as i go here) when i made the statement in the post above, it could lead to a misunderstanding that a fully antifragile state was not possible to attain.:
https://www.blackjackinfo.com/community/threads/antifragility.54909/
“a final note of caution, the parameter worst case expectation, is just that, as it is worded, an expectation. any expectation is either an average, or is based upon an average, that fact implies fluctuation may be existent. in other words, even a worst case expectation can allow for worse than worst case expectation results, much, much, much worse….”

to where that statement could imply the implication that all worst case expectations allow for worse cases than the worse case expectation. that implication is simply not completely accurate. for instance even if the worst case expectation for the game referred to in the posts above is based on a 100% HE, there are still regions of antifragility for which it is impossible to experience harm or lose money. see, example below, for which the HE = 100% with respect to the worst case expectation (in other words, the worst case expectation can not be any worse.

worstcase-jpg.8996
 

sagefr0g

Well-Known Member
#6
so what is the difference between typical advantage play and antifragile play?

answer:

expected results for typical advantage play may or may not come now in the short run, but should present in the long run if ruin is not realized first. chances for play are frequent, relatively abundant.

antifragile results happen now in the short run, there is no risk of ruin, but there is opportunity risk. chances for play are infrequent, relatively rare.
 

sagefr0g

Well-Known Member
#7
title: ever while in a casino, have you ask yourself, is there one play in this whole joint for which one could make that play and be guaranteed with certainty to make at least one penny? the answer is of course, no. but actually the truth is, it’s possible. such a play in essence constitutes this dreamers idea, of the ‘ideal complete play’.

Just me maybe, the point of an advantage play is that the value of a bet extended is exceeded by the value of the payoff of the completed play. bets extended value < payoff value .

herein lays the rub. advantage players deal with EV, expected value and EV is an average.

point being, upon the completion of an advantage play, expected value is not necessarily realized. our result (call it R) may be such that R>=<EV. advantage players know this and accept this, thing is though, they are ideally always shooting for EV, far as their concern is regarding a given play. in other words, EV is all they can really expect.

advantage players don’t so much mind this inaccuracy when it comes to actually reaching the ‘target’ of EV, mainly because of the fact that they know they will be performing such plays many, many, many,…. times, over and over and over,…. again. and that means that the results of those play events shall average out, so that EV is realized. so, ok, fine. well, no, maybe not so much fine. still there is an issue. that issue being time. and as that time unfolds, an inescapable misery of sorts unfolds in that one has to deal with the psychological conflagration of a given play not reaching EV at a given time. in short, an advantage player lands up having to pay for his ‘free’ lunch, with time and the psychological state induced by not being on target (visa v realization of EV now), sorta thing, and the misery of giving up time, if nothing else.

again, all the above, doesn’t mean there doesn’t exist utopia, in other words, the possibility of employing an ‘ideal complete play’ for which a bets extended value < payoff value for that ‘complete play’ with certainty and for a reasonably short extended period of time. such a play is what is meant by the ‘ideal complete play’, sorta thing. the point at which expected value is not an average but a certain value, achievable with certainty that is known ahead of time. an ideal complete play such as described if it was the only kind of play ever executed would mean that the advantage player would realize an ATH (all time high) after each and every play made.

far as this dreamer knows, it is possible for such an ideal complete play to present. this dreamer has in fact witnessed such plays (well ok, not completely ideal, but pretty darn close to ideal) out in the wild and exploited said plays. such a play is ultimately a play that is one bet from the point where expected value, overall advantage, worst case expectation and the best possible case converge or meet. the furthest point to the right on the graph as can be seen in this example:

https://www.blackjackinfo.com/community/attachments/worstcase-jpg.8996/

that’s just a specific example of what i mean by the idea of an’ ideal complete play’, sorta thing. just me maybe, there ought to be a way to mathematically envisage that idea in general, but it’s beyond me.
 
Last edited:
#8
Antifragile (2012), Fooled by Randomness (2001), The Black Swan (2007), and The Bed of Procrustes (2016), and the four-volume philosophical essay on uncertainty titled the Incerto ... how could DonS not love this Taleb guy? Look what Taleb has done for our own SageFrog!
 

sagefr0g

Well-Known Member
#9
@xengrifter ,
i can see how anyone would find themselves at odds with Taleb. thing is with me, i just try and separate the wheat from the chaff. edit: and one man's chaff is another man's treasure, as the saying goes. end edit
that said sir, have no doubt i'm as lost a puppie as ever there was, just as i ever have been.
funny though how sometimes lost puppie's some how survive.
anyway, ever when you was a kid, have two friends, each of which despised one another?
 
Last edited:

sagefr0g

Well-Known Member
#11
holypoly95 said:
Wouldn't antifragility in this case simply be an insane rainbow room?
pretty much so, imho. the trick being to analyze the spectrum with the purpose of steering clear of the insanity.
 
#12
sagefr0g said:
pretty much so, imho. the trick being to analyze the spectrum with the purpose of steering clear of the insanity.
DonS called, he said in order to steer clear of the insanity you must steer clear of Taleb and ALL of his concepts including Antifragility.
 

sagefr0g

Well-Known Member
#13
xengrifter said:
DonS called, he said in order to steer clear of the insanity you must steer clear of Taleb and ALL of his concepts including Antifragility.
tongue in cheek, no doubt sir.
anyway, crazy as i may doubtless be, i feel confident that both gentlemen mentioned above are utterly sane. they are both winners and know how to be such.
unlike the gentleman in the following link apparently was for a period of time (no disrespect meant at all for anyone suffering such matters): https://www.scientificamerican.com/...-nash-s-schizophrenia-disappeared-as-he-aged/
just me maybe, but messing with game theory would drive anyone up the wall. competition, uhhgggg :eek:
 
Top