logaritmisk/trueskill-tt
1
0
Fork 0
Code Issues 1 Pull Requests Packages Projects Releases Activity

refactor!: remove the inert online flag

Browse Source 
Skill.online was initialised to N_INF and assigned nowhere, so
HistoryBuilder::online(true) made every rating improper and log_evidence()
reported n * ln(0.5) — every game scored as a coin flip. The value is finite
and plausible, which is why it went unnoticed.

The default was false, so no existing result changes. A working replacement
lands next; a stored field cannot hold the quantity, because converge()
alternates sweeps and contaminates skill.forward with backward information
from the second iteration onward.

Also renames a test binding from ..._online to ..._forward: it passes the
forward flag, and the two senses being conflated is how this survived.
This commit is contained in:
logaritmisk
2026-08-27 16:11:37 +02:00
parent 187aede924
commit bf9d964cae
2 changed files with 12 additions and 36 deletions
Download Patch File Download Diff File Expand all files Collapse all files
src/history.rs
+5 -22
View File
@@ -29,7 +29,6 @@ pub struct HistoryBuilder<
beta: f64,
drift: D,
p_draw: f64,
online: bool,
score_sigma: f64,
convergence: ConvergenceOptions,
observer: O,
@@ -60,7 +59,6 @@ impl<T: Time, D: Drift<T>, O: Observer<T>, K: Eq + Hash + Clone> HistoryBuilder<
sigma: self.sigma,
beta: self.beta,
p_draw: self.p_draw,
online: self.online,
score_sigma: self.score_sigma,
convergence: self.convergence,
observer: self.observer,
@@ -87,11 +85,6 @@ impl<T: Time, D: Drift<T>, O: Observer<T>, K: Eq + Hash + Clone> HistoryBuilder<
self
}
pub fn online(mut self, online: bool) -> Self {
self.online = online;
self
}
/// Default observation noise for scored outcomes.
///
/// # Panics
@@ -135,7 +128,6 @@ impl<T: Time, D: Drift<T>, O: Observer<T>, K: Eq + Hash + Clone> HistoryBuilder<
beta: self.beta,
drift: self.drift,
p_draw: self.p_draw,
online: self.online,
score_sigma: self.score_sigma,
convergence: self.convergence,
observer,
@@ -155,7 +147,6 @@ impl<T: Time, D: Drift<T>, O: Observer<T>, K: Eq + Hash + Clone> HistoryBuilder<
beta: self.beta,
drift: self.drift,
p_draw: self.p_draw,
online: self.online,
score_sigma: self.score_sigma,
convergence: self.convergence,
observer: self.observer,
@@ -171,7 +162,6 @@ impl Default for HistoryBuilder<i64, ConstantDrift, NullObserver, &'static str>
beta: BETA,
drift: ConstantDrift(GAMMA),
p_draw: P_DRAW,
online: false,
score_sigma: 1.0,
convergence: ConvergenceOptions::default(),
observer: NullObserver,
@@ -196,7 +186,6 @@ pub struct History<
beta: f64,
drift: D,
p_draw: f64,
online: bool,
score_sigma: f64,
convergence: ConvergenceOptions,
observer: O,
@@ -223,7 +212,6 @@ impl<K: Eq + Hash + Clone> History<i64, ConstantDrift, NullObserver, K> {
beta: BETA,
drift: ConstantDrift(GAMMA),
p_draw: P_DRAW,
online: false,
score_sigma: 1.0,
convergence: ConvergenceOptions::default(),
observer: NullObserver,
@@ -399,7 +387,7 @@ impl<T: Time, D: Drift<T>, O: Observer<T>, K: Eq + Hash + Clone> History<T, D, O
let per_slice: Vec<f64> = self
.time_slices
.par_iter()
.map(|ts| ts.log_evidence(self.online, targets, forward, &self.agents))
.map(|ts| ts.log_evidence(targets, forward, &self.agents))
.collect();
per_slice.into_iter().sum()
}
@@ -407,7 +395,7 @@ impl<T: Time, D: Drift<T>, O: Observer<T>, K: Eq + Hash + Clone> History<T, D, O
{
self.time_slices
.iter()
.map(|ts| ts.log_evidence(self.online, targets, forward, &self.agents))
.map(|ts| ts.log_evidence(targets, forward, &self.agents))
.sum()
}
}
@@ -935,12 +923,7 @@ mod tests {
let w = [vec![1.0], vec![1.0]];
let p = Game::ranked_with_arena(
h.time_slices[1].events[0].within_priors(
false,
false,
&h.time_slices[1].skills,
&h.agents,
),
h.time_slices[1].events[0].within_priors(false, &h.time_slices[1].skills, &h.agents),
&[0.0, 1.0],
&w,
P_DRAW,
@@ -1180,11 +1163,11 @@ mod tests {
let f = h.keys.get("f").unwrap();
let trueskill_log_evidence = h.log_evidence_internal(false, &[]);
let trueskill_log_evidence_online = h.log_evidence_internal(true, &[]);
let trueskill_log_evidence_forward = h.log_evidence_internal(true, &[]);
assert_ulps_eq!(
trueskill_log_evidence,
trueskill_log_evidence_online,
trueskill_log_evidence_forward,
epsilon = 1e-6
);