Tech download: IndyCar fuel saving strategies

ED's note: Welcome to the first installment of a new series of IndyCar technical deep-dives that we'll be running through the winter. Over the coming weeks we'll be shining the spotlight onto all sorts of topics related to making an IndyCar go fast, many of which are rarely discussed in detail outside of engineering debriefs.

To help explain them we're pleased to have Stan Sandoval, who was dealing with these very concepts as an IndyCar engineer until quite recently, and who was first introduced to RACER.com readers last year when he wrote a column about how a shot-in-the-dark letter to our late colleague Robin Miller helped set the ball rolling on his own career in motorsport. If you missed it the first time around, you can find it here.

Fuel saving is discussed during practically every IndyCar broadcast, and it becomes the primarily focus of in-race strategy several times a season. On the surface, it may seem counterintuitive that there is a strategic benefit to not pushing 100% at all times, but the truth is that fuel saving is a powerful strategic tool that teams turn to in many different situations.

The strategic goals of every team are the same in every race: achieve the best finishing position possible. Clearly, teams are choosing to fuel save because it helps them finish higher up the order, but what situations call for fuel saving, and why is it sometimes better than pushing flat-out? While fuel saving is prevalent in IndyCar, the motivations and techniques for doing so are often hidden from fans in the name of preserving a competitive advantage.

Why do teams save fuel?

1. To remove a stop

Probably the most obvious reason to fuel save is to make less pit stops. This is the old fable of the tortoise and the hare: is it faster to sprint the whole race and stop more often, or jog the whole race and stop less? There are lot of factors that go into deciding whether eliminating a pit stop by fuel saving is the ideal course of action for a race.

A very basic example can be used to compare strategies, and then complexity can be layered onto this model later on. Say the race is 100 laps: there is only one tire compound, there is no deg, and there are only two cars. It won’t be the most exciting race ever on TV (you could argue it wouldn’t be the most boring, either), but it does allow the variables to be controlled so that there is a direct compare between a flat-out strategy and a fuel saving approach.

In this race, pushing to the absolute maximum uses 0.74 gallons each lap and achieves a 60s lap time. The time lost making a pit stop is 35 seconds, and the fuel tank is a standard 18.5 gallons. Looking at the options, the obvious choice is to push the whole race, do 25 laps in each stint, and stop three times (the hare). The other option is to fuel save in order to only stop twice (the tortoise). This means going 33 to 34 laps per stint, which will only allow 0.54 to 0.56 gallons per lap to be used for the whole race. This consumption becomes the 'fuel target' for doing a two-stopper.

So there are two options on the table, but now a decision needs to be made as to which one is faster. This is where engineers can compare the strategies by effectively racing them against each other mathematically. By adding together the time it takes to do all of the laps and all of the pit stops for a given strategy, engineers can predict which car would finish in front. In this example, the three-stopper does 100 laps at 60 seconds and makes three 35-second pit stops, for a total race time of 6,105 seconds.

The two stopper also does 100 laps, but at a different lap time (since it has to hit the fuel target, they’re going to go slower), and only makes two 35-second pit stops. During a practice session in the buildup to this race, the driver is able to achieve a lap time of 60.2 seconds while hitting the fuel target of 0.54 gallons per lap. To compare the two strategies, engineer will often visualize how the race will play out. Below is a typical strategy graph showing lap number vs. total race time. Note that while total race time would normally just count up and up as the race goes on, engineers will usually plot total race time relative to some average to make the graph more readable.

In these graphs, the vertical spacing between the two strategies tells the gap on track on that lap. (Think of it as the difference in total race time between two strategies when they’ve completed the same lap). The upward slopes of the plot shows lap time (steeper is faster), because it’s change in the total race time per lap. The sudden downward drops in time are pit stops: 35 additional seconds lost in a single lap in order to change tires and add fuel. This basic model has constant lap time all throughout the stint, which isn’t particularly realistic.

Each lap that the three-stopper completes is the same 60s lap time and each lap for the two-stopper is the 60.2 second lap time that the driver was able to do in practice. The difference in pace between the two cars is represented visually by the difference in slope. It’s pretty evident right from the get-go that the three-stopper has better pace and begins to pull away. From this graph, it says that even though the three -topper is faster than the two-stopper throughout the whole race, the three-stopper is not able to make up the time lost by making an additional pit stop. In this example, the tortoise wins by 15 seconds.

From this rudimentary model, engineers can begin to add complexity to consider other variables and factors to make the model more representative of reality. If the model needs to consider tire wear, that effect can be included by making each lap 0.05 seconds slower than the previous lap until they pit for new tires. Because the two-stopper and the three-stopper have different stint lengths, the deg will affect the two strategies differently, just as it does in reality. The two-stopper, with its longer stints, is going to suffer more deg.

Rerunning the race with tire deg gives a similar graph, but with a key difference: both cars slow down as the stint goes on. This is because as the tires wear out, the lap time gets slower, as shown by the slopes getting less steep as the stint goes on (and then resetting when new tires are put on). With the model now considering tire deg, the result is different: because the two-stopper has to deal with more deg as a result of its longer stints, the three-stopper is eventually able to catch the two-stopper and make the pass with 10 laps to go. When considering tire deg, the hare wins.

As the model continues to consider other factors, it can eventually incorporate all facets of strategy: tire compounds, traffic, cautions, and whatever else an engineer thinks is relevant when numerically modelling all of the considerations that go into a race strategy. These models can become hugely complex since so much can happen in a race. The situations can be endlessly complicated and no model is perfect. Regardless of the model complexity though, the overall conclusion is the same: if fuel saving in order to eliminate a pit stop appears to be the best way to finish in front, then teams will employ a fuel saving strategy. Sometimes, they do this right from the drop of the green flag.

2. Getting off the inferior race tire

IndyCar rules mandate that each car run both tire compounds in a dry race, and as a consequence both tires will always be compared to each other over the course of a full race stint to determine which one is preferred. Sometimes, the difference in tire performance will be so great that teams will choose to spend as little time on one of the compounds as possible. This usually involves starting on the inferior tire, pitting very early to get off it, and then running longer stints on the superior tire until the end of the race. Doing so has a domino effect on fuel strategy: if the first pit stop is made early, the team will have to save fuel for the rest of the race in order to lengthen the subsequent stints and avoid an additional pit stop.

Going back to the 100 lap race example, now the Alternate tire is modeled as a really poor race tire this weekend. Yes, it’s 1.0s faster at the start of the stint, but it wears so badly that pretty soon the driver is on the radio saying the tires are dead. Going with the flat-out approach would mean doing 25 lap stints, going Alternate/Primary/Primary/Primary. It’s nice because there’s no fuel saving and the driver can push the whole time, but when the Alternate tire wears badly, 25 laps is a long way to go, so maybe it’s not the fastest strategy for this race.

Stopping on lap 22 would get the Alternates off a bit earlier, and then three 26-lap stints on Primary tires would follow in order to get to the end. It would require a bit of fuel saving (to get an extra lap on each of the final three stints, the consumption would only have to go from 0.74 gallons per lap to 0.71), but it would mean three less laps on worn Alternate tires.

This idea can be taken further by stopping on lap 19, then doing three 27 lap stints at 0.685 gallons per lap. And from this, a pattern begins to emerge. The earlier a team gets off the inferior tire, the more they have to fuel save for the remaining stints on the superior tire. And so it becomes a balance between tire life and fuel saving pace. At what point is this trade-off no longer worth it?

Turning back to the race simulation model, it will now also consider the Alternates, which can do a 59s lap and degrade at 0.2s per lap. From practicing fuel saving in one of the warm-up sessions, the team knows that extending a stint one additional lap means running at a pace 0.03s slower for the whole stint: so a 26-lap Primary tire stint is run at 60.03s (plus deg) in order to hit the fuel target, a 27 lap Primary tire stint is run at 60.06s (plus deg) in order to hit the fuel target, and so on. Armed with this information, teams can begin to sweep through all the possibilities of when to take off the Alternates, and then calculate the how the remainder of the race would go.

When graphed, it really highlights just how punishing it is to stay on the Alternates tires for too long (green plot). It’s true that staying out longer means less fuel saving for the rest of the race, but sometimes that’s not enough to make up the for the time lost by staying out on too long in the beginning. The green trace (which pits on lap 25) emerges from the first round of stops in last place, and even though no fuel saving is needed from that point onwards, they only make up one spot (passing the yellow trace with eight laps to go).

It’s also evident how the difference in degradation between Alternates and Primaries plays a huge role in determining pit laps and fuel targets. In this example, the optimum first pit stop is lap 16. That’s a whole nine laps earlier than the no-fuel-save strategy, and it’s 7.74s faster. For any race that is a huge difference, all born out of strategy options that are only possible thanks to fuel saving.

Throughout the season, the preferred pit stop lap will change depending on the performance difference between the compounds and the time loss when hitting the fuel target. These factors can vary from track to track, so teams are constantly using practice sessions to gather this data.

As the model continues to gets more sophisticated, some may wonder why team don’t simply calculate the optimum strategy and do it every race. In reality, there isn’t always a consensus on what is the best strategy. Every team is using their own tools and their own data, drawing conclusions from their own analysis and experiences. Sometimes teams just have different opinions on what they think is best. There are countless examples of teams just getting the strategy downright wrong, whether it’s the number of stops or which tire is preferred. Other times, the data collected in practice isn’t fully representative of the race. In the example, the race started with the team thinking the Alternates would degrade at 0.20s per lap based on practice data, but oftentimes the ambient conditions are much different for the race. If it gets to lap 13 and calculating the degradation says it’s more like 0.35s per lap, the optimum lap to stop on is lap 10 – but it’s too late to do that now!

No strategy is so stagnant that it is decided before the race and then simply followed like an instruction manual. Information throughout the race regarding deg, fuel targets, cautions, and traffic all evolves in live time and the teams need to react accordingly.

Another reason that there is not one single strategy is because the optimum strategy is different for each team. Even at the very top level, there is a wide spread on several important factors, such as a setup’s effect on tire deg, a driver’s ability to fuel save and not lose lap time, or the amount of traffic a team will have based on their starting position. These will change what the 'best' strategy is for each team for their situation. Traffic plays a huge part in strategy decisions, since being stuck behind another car can affect pace drastically. If a model says the three-stop strategy is faster because the pace warrants the additional pit stop, that model better consider the fact that pitting an additional time will mean being behind cars and losing time while trying to pass them.

3. The undercut

Since traffic can be so detrimental, teams turn to a strategic move called the undercut, where they will intentionally stop earlier than what the model says is 'ideal' in order to avoid traffic. Think of an undercut as an advance loan on pace: the team gets to go to new tires earlier and push in clean air, but will have to reduce fuel consumption later on in the stint to make up for having stopped earlier. Still, if the time gained through the pit cycle allows the team to get past a car they otherwise wouldn’t be able to overtake on track, the benefit can be massive.

In the previous example, the model said the optimum pit stop lap was on 16, but stopping on lap 15 would make total race time is only 0.36s slower, which admittedly is not that much compared to the loss that can be caused by traffic. Being behind a car that is running 1.0s slower that they can’t overtake because of dirty air can ruin the entire race. So the faster car that is stuck in traffic decides to pit on lap 15, one lap earlier than optimum according to the model. By the time the slower car decides to pit on the optimum lap, the faster car has gained enough time comparatively in clean air that the slower car comes out of the pits behind. Now in front, the superior pace allows the faster car to pull away despite having to save more fuel in order to go one lap longer on this stint.

Undercuts can be used offensively as just shown, or defensively. There’s nothing to stop the slower car in front from pitting a lap earlier and doing the exact same thing in order to keep the position. Yes, the car in front is now slower and saving more fuel, but if the quicker car still cannot get past, the tactic has successfully maintained position.

This is a critical decision in-race that is really difficult to model. If both teams are trying to undercut each other, a game of one-upmanship can ensue, leaving both cars pitting absurdly early just to cover each other and having to drastically reduce their consumption. While this may do well to keep them in front of the car they’re battling directly, this could leave the team vulnerable to other cars. While models are good at considering factors intrinsic to the team itself (compounds, number of stops, deg), determining the actions of other cars in the race is very difficult to predict.

Final thoughts

These are just some of the many ways in which saving fuel can give teams a strategic advantage as the race plays out. While teams are always looking for 'on-car' performance, a team’s ability to predict and numerically model all the considerations that go into choosing a strategy is equally as important. Being the best fuel-saver in the world doesn’t win much when it’s done at the wrong times. Still, as with anything in IndyCar, if there is a performance benefit to be had then teams will do everything in their power (and budget) to maximize it. Teams will go to great lengths to be better at saving fuel than their rivals, as these have examples have shown it can play a huge part in finishing ahead.