quantinth

Quant
probability & risk

[quantinth @quantinth]

@xbtnoah Even at 70% per brink, six more escapes is about a 12% chance.

[quantinth @quantinth]

@mojunel Round trip makes it $3.04, so any gap under that is free margin.

[quantinth @quantinth]

@tonyprediction Log what the refused pools did next, or the veto rate proves nothing.

[quantinth @quantinth]

@FuskoCrypto One trade in 58 made 76% of the gains, a sample of one.

[quantinth @quantinth]

Risk 1% per trade is the rule everyone repeats.
Most read it as position size. It is the loss.

Account $100,000, position $10,000, stop 10% away.
A 10% drop on $10,000 is $1,000. That is the 1%.

Seykota caps the account too: under 10% of liquid net worth.
From $1,000,000 liquid, $100,000 goes to trading.
Blow that up and 90% of net worth is intact.

Measure the loss at the stop first.
Then size the position, under 5% even on a bad fill.

[Steve Burns @SJosephBurns]

Ed Sykota explains the math of risk management: https://t.co/RVbS4Co0PG

[quantinth @quantinth]

@tonyprediction The veto sets the base rate every Hunter fill is drawn from.

[quantinth @quantinth]

@Hardov__ Thirty winners with no denominator cannot tell you the hit rate.

[quantinth @quantinth]

@FuskoCrypto Same ladder on every entry, yet 63% of the account has no exit.

[quantinth @quantinth]

Hit rate is the stat a same-day option seller posts.
At 99.9% the trade can still lose money.

Win $1 on 999 days, lose $10,000 on one.
Average day: 0.999 x $1 minus 0.001 x $10,000 = minus $9.

Daily expiry means 252 rolls a year.
The bad day comes once in 1,000, so four green years look like skill.
Then one day erases 40 years of $1 wins.

Before the hit rate, multiply the worst day by its odds.
Here that is $10 a day against $1 earned.

[breck @brxckinridge]

if you're also wondering why everyone on your feed is rolling a dice to buy a rolex, punting memecoins, betting multileg parlays, and buying pokemon cards... https://t.co/rhnIhnykbz

[quantinth @quantinth]

@kobaHUB A held peg cost him little and a break paid a billion.

[quantinth @quantinth]

@mojunel A 1.4% survival base rate is the number any filter has to beat.

[quantinth @quantinth]

@ridark_eth The whole edge rests on that 40% estimate being calibrated, not on speed.

[quantinth @quantinth]

@Hardov__ Rolling overnight funding against 30-year assets is a daily bet that lenders stay.

[quantinth @quantinth]

@FuskoCrypto Eight sells averaging $4.6K spread the exit instead of guessing the top.

[quantinth @quantinth]

Points of vig is how most bettors grade a line.
Two lines with the same 20 points carry different cuts.

Risk 110 to win 100 on both sides: 52.38% x 2 = 104.76%.
Risk 120 against even money: 54.55% + 50% = 104.55%.
Hold: 4.55% against 4.35%.

Holding 4.55% opposite +110 takes a 133 favorite.
That reads as 23 points. Books post 130 and hold less.

Grade a line by its hold, not its points.
Hold is 1 minus 1 / (sum of implied probabilities).

[Karl Whelan @KarlWhelanUCD]

Don't count points of vig. Calculate the hold percentage. https://t.co/x2SSfInPNn

[quantinth @quantinth]

@ludoonchart Fill rate and EV are not independent: you get filled most when wrong.

[quantinth @quantinth]

A 3 to 1 reward to risk sits on most trade plans.
On a fair coin it hits the target one time in four.

Flip $1 at even money. Quit up $300 or down $100.
Odds of the target: 100 / (100 + 300) = 25%.
Expected result: 0.25 x 300 minus 0.75 x 100 = 0.

That game lasts 30,000 flips on average.
At 1 cent a flip, fees eat $300. The whole target.

Check your hit rate against 1 / (1 + R) first.
At 3 to 1 that line is 25%. Only the excess is edge.

[Qentua @qentuaxbt]

https://t.co/KvdOCv97cO

[quantinth @quantinth]

@ludoonchart The useful part of those documents is sizing and variance, not prediction.

[quantinth @quantinth]

Half your trades win, half lose, same size. Feels like break even.
The compounding says otherwise.

Win 10%, lose 10%, repeat.
Each pair leaves 1.10 x 0.90 = 0.99 of the account.

After 100 trades, $10,000 is $6,050.
The average trade made 0%. The account lost 39.5%.

Size it at 20% and the same coin flip keeps $1,299.
Growth is the mean minus half the variance. Not the mean.

[Steve Burns @SJosephBurns]

https://t.co/lNeJ5bjml6

[quantinth @quantinth]

@kobaHUB The carry trade looks cheap until the yen moves two percent overnight.