The difference between FF Rating and finishing time

There are various ways to evaluate the performance of your horse. You can look at finishing position, winnings, ROI (winnings/fees), finish time and Fleet Figure (FF) rating - all easily available from PhotoFinish.Live (PFL). There are more exotic metrics you can calculate (which we will cover at a later date). There are also factors that you need to take into account (e.g. racing under ideal conditions). Ultimately this is a game of imperfect knowledge, like Poker, where you get lots of clues, many of which maybe are not very obvious, but you need to join the dots and figure it out.

Two of the more obvious measures of performance are finishing time and FF rating. You get both at the end of every race. But what is the relationship between them?

Understanding some definitions

PFL give us a definition of how the FF Rating benchmark works to an extent, but keeps the secret sauce under lock and key.

It's a combination of factors and is somewhat normalised. This probably refers to some removal of outliers(extreme performances) and/or making adjustments for some other unknown factor(s). So let's take a look at FF ratings themselves, we will leave the benchmarks for now.

We can access our latest finish times and FF ratings under the Fleet Figures tab for any horse in the game. If you compare similar distances and conditions, you will see that there is quite a bit of variance.

In fact, if you calculate the correlation (which is a statistical test to determine whether there is a relationship between two elements) between finish time and FF rating it turns out to be -0.55. A fairly strong relationship, where the negative means that as time goes down (you run faster) rating goes up - but it also suggests there are other factors at play that contribute to a significant amount of variance, and the two metrics are measuring different aspects of performance.

Anyone can be an analyst

You don't need a PhD in data science to apply some fundamental analysis of this kind of data, anyone can do it. The first non technical analysis I would perform is the good old eyeball test. What do you see in the following scatter plot?

Each dot is a winning time from a race and the respective FF rating awarded from season 10 onwards (just to give us a decent sample size) purely for 4 furlong (the shortest) races. You shouldn't calculate correlations for all distances together - each distance is a discrete event with its own relationship, and should be treated as such. Across the x-axis (bottom) we have the finishing time where the further to the left the faster the time. On the y-axis (vertical) we have the FF rating. I've clipped both axes so we can zoom into where the action is (no nefarious reason).

What I see is a series of parallel lines running top left to bottom right. This represents that overarching pattern of the faster you run, the better your FF rating. But clearly something else is going on. The fastest time doesn't have the highest rating and the same is true at the other end. If you follow one horizontal line you can see that horses were awarded a rating of 70 for times that range between 44.5 and 47.5 seconds.

If we do the same analysis on races where the conditions were Fast, we see a smaller sample of data but the same pattern emerges. This means that condition is not the driving factor behind this pattern, if anything, the pattern here is clearer.

Now at this point we could do something clever like cluster the points to reveal group membership but a simpler approach is just to think about what might be driving this pattern, and to me the obvious candidate is race type. There are lots of rumours on Discord that running in claimers is a bad idea if you want to reduce your FF benchmark so that you are more competitive in benchmark races themselves. So let's check that out. The below chart takes the same data from above and just colour codes it for each race type. What do we see now?

Well as you might expect, the fastest times (left most) run in Grade I races. See the clump of orange dots on the top left - fast runs being rewarded with high FF ratings. Maidens, the pink dots, appear in two separate lines almost a second apart. Reds, claimers, again are in two distinct groups. It looks like different race types apply different rules for awarding FF ratings for similar times. Handicaps and some other allowance style races give horses different weights that affect their finishing time - that's probably the chaos on the right side of the chart.

If we have two parallel lines for Maidens and Claimers I would suspect the difference is surface. This final chart clearly shows the affect surface has on finishing times, which is accounted for in FF rating.

What did we learn?

There is a fairly strong relationship between finish time and FF rating - but they represent fundamentally different measures of performance. I prefer to use finish time to create performance metrics for my horses because FF rating already has a lot going on in it. FF rating also factors in finishing position and penalises horses for finishing lower despite posting a faster time. That can be misleading, particularly if you are using this data to identify the ideal conditions to run your horse under.

I recently ran an S- 2 year old colt in a claimer and received an FF rating of 81, despite his previous best FF rating being 71. Is he suddenly a much better horse than I thought? Sadly not. If you are trying to pad your rating to get your score up to sell the horse - running claimers might have some merit. If you are trying to get your benchmark down to qualify for some restricted races, you don't want to enter races that inflate your FF rating. If you are evaluating horses to buy on the marketplace, just check the best FF ratings were not all achieved in claimers.

In a subsequent post, we will explore what you can do with finishing time to build derivative measures of performance and get better at spotting under-valued horses on the marketplace or entered into claimers.

Join the fun and put these insights into practice at PhotoFinish.Live and if you are considering starting your own stable please consider using my referral code: PADDOCK or just click on this link: https://signup.photofinish.live/?referralCode=8EUMC4P2

Please remember this is a web3 game where your spend your own money. Nothing I write about should be considered financial or investment advice.