I was trying to race full Monaco with the gen2 cars. It's tough. On three races, I ran out of electrons before the end of the race. Twice between the final corner and the start/finish line and once shortly after the chicane. I was running regen at 7, 8 and 9 levels. While doing some practice runs, I turned regen down to 0. I was still seeing regen happening when lifting/braking and it seemed to be happening at the same rate. It's almost like there's a fixed rate built-in and it doesn't matter what you set it to in setups or in-car adjustments. Can someone else confirm that regen is working properly before I go crazy please.
I decided I needed to gather some data. I selected Berlin just because I know it well and I ran 30-minute private practices with regen at 0, 3, 5, 7 and 9. I used the same setup on the car for each run, did no in-car adjustments and the weather was the same for each run. This resulted in 22-23 laps for each session. I used Motec to show fuel use for each lap. Here are the results: Code: Regen Ave Fuel Used Fastest Lap Fastest Level per Lap Fuel used Lap 0 3.44 3.66 1:14.586 3 3.24 3.39 1:14.167 5 3.21 3.21 1:13.143 7 3.10 3.01 1:12.855 9 2.89 2.94 1:13.155 Some points to consider. 1. I am not a very good driver. 2. My setup may be far from idea. I found that I really struggled to do decent laps at the lower regen levels 0-3. I tuned the car originally using the default regen 5 level. It occurred to me last night that instead of setting a regen level and running the entire lap on it, I maybe should be adjusting regen for the corners I am approaching...a higher regen for a long braking section and a lower one for faster corners. I thought about redoing my steering wheel and changing the paddle shifters to regen up and down. Since there's no shifting with the FEs, there's no reason not to use the paddles for something else.
One thing that might be impacting your test is brake bias. If I remember correctly, energy regen comes from rear wheels, and thus brake bias changes significantly from regen level 0 to 9. Brake Bias is by default weighting more towards front wheels, which is not optimum, but protects from spinning too easily. By increasing regen level, you're also moving brake bias towards the rear, which will help having braking power that's closer to the optimum under normal conditions. You'll probably have better lap times with low regen levels if you move brake bias more towards the rear. In Formula E, lift (throttle) and coast is an important process to manage energy use and optimize lap time. You might find it useful to look at a performance plot, showing several laps with both lap time and energy used. The idea is that there is a curve that represents the best lap time you can take out of specific energy use. You could do this for different regen modes, and see the impact. Below is an example of an old session I did almost a year ago. It's just a ~10 laps session, and there's no comparison on this example (e.g., stint with different regen mode, etc), but it was just to give you an idea. Power mode also has a significant part on the results, but from what you wrote, that was kept constant. Hope this helps. Cheers
As a further test last night, I drove a 45-minute race at Berlin using the same setups as my earlier tests. I wanted to see what would happen if I used one Regen setting (7) to drive the entire race. Unfortunately, I ran out of fuel with 5:50 left in the race. I drove 31 laps before running out of fuel and the AI ran 37. To drive 37 laps, I would need to average 2.7 fuel each lap. My other tests say that is not possible...even Regen 9 uses an average of 2.80 per lap. There could be some explanations here. One is that the car is horribly set up. Two, I am just a really bad driver. I also should note that I have boost mode turned off for the car just to keep the number of variables down.
Following your post, I had a chance to try this combo again, to see how it panned out. I used the default setup (avg power mode 220 kW / mode 2, regen mode 5), except for brake bias (50.5:49.5) and brake force (100%). I did two stints, just to compare impact of lift and coast on lap time and energy used. I did not change power mode nor regen mode. Stint 1 was done using all energy to reduce lap time. Stint 2 was done by experimenting with lift and coast. Here is the lap times evo: On the first stint, lap times keep reducing as I improve driving with experience. Independently of that, you can definitely see an impact of ~2 seconds with the lift and coast on the second stint. In terms of energy used, on my first stint I was quite thirsty , using around 3.6 p.p. per lap: For the second stint, I experimented a bit, but ended up lifting and coasting on: - before T1 (~50 meters) - before T6 (>100 meters) - before T9 (~50 meters) - before T10 (~25 meters) This felt like a huge portion of the time not pushing Here are the lap times / energy used table for this stint: As per the above, it's possible to reduce energy used by ~1 p.p. per lap (below 2.7 p.p. as you mentioned), although that means a lot of lift and coast, and decreasing pace about 2 seconds. I didn't have time to play with regen modes (or power modes), but it should be interesting to see their impact as well. Hope this helps re-framing how to tackle a race with these Formula E cars. Cheers
I tried doing a lot more lifting and coasting last night. My laptimes increased by about two seconds per lap, but I was able to get my fuel usage down to about 2.5 per lap, which is more than adequate to finish a 45-minute race at Berlin. For testing purposes, I set regen at 7 and did not touch it during the outing. I finished the session with 8.46 units of fuel left. That is equivalent to about three laps or roughly four minutes. If I spread those 240 seconds over the 34 laps, it would have reduced my laptimes by seven seconds per lap. So this all goes back to my original assertion--that I am not a very good driver.
I ended up doing another ~40 laps session, this time using regen mode 9. I had to reduce brake bias to (54:46) for better grip while braking and avoid spins. With the increased regen mode, I was able to reduce energy used by 10% to 20% when in lift and coast mode, with no relevant impact on lap time. Here is a chart showing energy used per lap vs lap time. Blue dots are from previous session with regen mode 5. Red dots are new session (regen mode 9). I believe the blue dots representing faster lap times are not fully representative of "best lap time vs energy used" theoretical best possible result (with myself as a driver). They were part of the 1st stint of the 1st session, where lap times were decreasing rapidly as I was improving with initial driving experience (see previous post). Those blue dots are expected to show faster lap times next time around, meaning I think lower regen mode could enable faster lap times. Higher regen mode enables lower energy used with similar lap times it seems, even if at the cost of consistency. I felt it was harder to be as consistency with higher regen mode (visible in the chart with more red dots vertically aligned).
Last night, I fixed my regen at 5 and ran a 45 minute race. I ran out of fuel three laps short. My best time was reasonable; my consistency was awful. I ran a 1:12.8 followed by a 1:14.xx . This is a skill issue. One thing I noticed that is nice is that with the new RCs, I never got hit from behind. This seemed to happen a lot prior to the RCs because I would lift earlier than the AI would start to brake. To me, it seems the AI drives the same in FE as they would if they were any other car. I am beginning to get a better understanding of how this all works. It's going to take me a few more laps before I get it engrained in my aging brain how to drive these machines. In trying to better use regen and by lifting a ways before braking (and trying to minimize braking), I've also found that I can dial up the brake pressure setting. I was running it low to avoid locking up the brakes. Basically, I think the problem here is in the cockpit and that needs to be resolved first for me.
The fun is in the journey These cars are interestingly different, because use of one degradable variable (energy) impacts time performance significantly. Looking at performance over a race distance, if there were no other cars, we just need to consider the number of laps done during that specific time. It doesn't matter how you got there. You can have the same overall performance (total time over stint) by always going at same pace / energy, or by varying it throughout the stint (e.g., going faster for a bit and then going slower with less energy at other times). On the other hand, when assessing a (race) lap performance, the right way to do it is to calculate the distance between your data point in the performance plot (the chart in my last post), and the estimated theoretical curve of best lap time-energy used line. This curve is essentially a continuous line that follows and is just below the data points. This is how it's done by Formula E teams. Bottom line, you can have a faster lap with worse performance. In a normal race (with other cars), race strategy decisions regarding "pace vs energy" do have an impact on overall performance (total time over stint & position), given the whole new approach possible. You might be faster at some point to by using more energy (e.g., to get ahead or in clear traffic), or managing your pace using less energy (e.g. keeping others behind, in situations where it's difficult to overtake). One conclusion on the above (and your messages), is that you're not able to easily rely on AI level alone to select AI cars that are fun to race with you (meaning similar track performance). I assume the reason is likely AI level essentially just adjusts time performance, instead of correctly adjusting the full lap time-energy used curve (maybe it already does this but not very well). This time-energy use curve is the true performance indicator for Formula E cars. I would suggest you to consider shorter races to run with AIs as a quick fix to improve fun. And we haven't even discussed the attack mode!
This is not about "fun." If this weren't fun, I wouldn't do it; I'd just log into work instead. My goal now is to understand the relationships among the elements that make up a "good" lap. I drive more and try different things to gain more knowledge. I don't find running shorter races helpful to what I am trying to accomplish personally. The more I drive the FEs, the more interesting they become.
Hi there, I have to say your message came across a bit unnecessarily aggressive as I first read it. Tbf, I'm probably misunderstanding it. Anyway, I was trying to be helpful when engaging with this thread. I hope that was the case . I also find it fascinating to analyze this kind of stuff. I wish you the best finding what you're looking for. Cheers
When I re-read it, I don't see it as being aggressive. It certainly is not meant that way. That's the problem sometimes with text rather than voice communication. At any rate, my conclusion at this point is that the "problems" I see/have are directly related to how badly I drive. Trying to understand the mechanics of fuel vs laptime and collecting the data pointed this out very clearly. The inconsistency, for example, also makes it difficult to evaluate setup changes. Did you run a good lap because of the setup change or because you just happened to run the lap properly. I just think that, for me, the data are invalidated by the consistently bad (or inconsistently good) driving.
No prob. I completely agree that lack of consistency makes it very difficult to compare setup performance.
