Aiden Holmes - FencingTracker

Drew University NCAA / Salle Auriol Seattle

Weapon	Type	Strength	Estimate Range^*	Min	Max
Saber	DE	2988	2776 - 3199	2595	3212
Saber	Pool	2806	2629 - 2982	2530	2902
Foil	DE	1759	952 - 2565	1759	2046
Foil	Pool	1930	1540 - 2319	1824	1940

^* We're quite confident that the person's true strength is somewhere between those numbers. For those statistically inclined, it is a 95% confidence interval of the estimate (i.e., μ ± 2σ).

Matchup against me

Outcome	Chance	My strength change	Their strength change
Log in to see the result of complex math. Or sign up if you are still not with the program unlike the cool kids. Make sure you enter your USFA number correctly for this feature to work.

Strength over time

The strength changes are summaries from all events of that day. The dotted line indicates that the strength rating is too uncertain.

Explanation

The strength number is a numeric representation of how a fencer has performed against others historically.

The system "learns" your strength as you fence. It takes about 12 bouts for it to zero in on a good estimate. This is why in the beginning you will often see wild fluctuations.

An unknown fencer starts with a 2500 strength, but with very high uncertainty.

Due to their different natures, pools, direct elimination, and mixed gender bouts are tracked separately.

The lowest possible is zero, average is around 2500, highest possible is around 5000. The scale is not linear, which means that higher numbers are increasingly harder to obtain. To be precise, the distribution of strength numbers looks like a bell curve.

If you beat a much stronger fencer, your strength will go up by a lot. If you beat a much weaker fencer, your strength will go up by little or maybe even not at all. Losses follow the same logic. You can see detailed changes after each bout in the bout history.

Probability of winning doesn't account for difference in fencing styles, familiarity with opponents, or any of such things. It is a reflection of the difference in strength estimates assuming that both of you perform just like you did against others in the past.