Anti sandbagging and other areas that need development and communication

That’s a pretty blanket statement without actually considering the realities for what impact the new system will have. A ranking system is also not cruiser proof. I could exclusively enter races where my I am the strongest rider but my ranking will not be too adversely affected based on the strength of the field. People do not race enough, with enough variety, for any ranking system to accurately and consistently match for rider strength and ability. It’s completely dependent on the quality of the system, it is certainly not simple. My ZP ranking results have been really low (bad) from racing ZRL A1 all season, yet when I raced B in the FRR I was getting huge ranking results each week is much lower quality fields.

Now you know I support the development of a ranking system, but to suggest that is the only way to stop cruisers is, well, nonsense.

The purpose of the question was to discuss if the CP and MAP system will eradicate cruisers.

eFTP, let alone CP/MAP, as previously discussed would be a great way to do it, because it is not realistic or worthwhile for a rider to manage every effort on zwift to keep power down across all durations so that the system thinks they are weaker than the are, then enter a race in a category lower than they should, to continue keeping power down, to get the glory of a win. Is it possible? I guess it is, because sprint power is not considered in the model, but is it really going to happen?

The one change I would push for is to also include workouts in the calculation, because for anyone also carefully monitoring their power in workouts, well to be honest - they’re a bit weird and we should probably just leave them be.

4 Likes

Nothing is ever foolproof and someone will always find a way to work around but it’s a bloody good attempt at it.

I also think even if you did manage to find a way to keep yourself within limits it would be hard to do it while still being competitive.

3 Likes

Well, if you define a cruiser as someone who wins repeatedly without having to perform at their peak level and without getting promoted then it is indeed trivial to see that any discrepancy between actual performance and the ordering predicted from some power metric will result in cruisers. They will be precisely those people whose actual performance exceeds their theoretical order according to the power metric.

This is why this sort of performance based metric doesn’t work, cannot work, has never been used in any other sport, and is a dead end. Nevertheless, the categories as now defined do appear to be an improvement over the 20 min thing (though perhaps problematic for lightweights) and the problem of cruisers will probably be reduced, perhaps significantly.

1 Like

This is probably more a question or thinking out aloud …

If data used is only derived from completed events or saved rides (rather than live data) then am I right in thinking someone whose Cat is derived from high MAP rather than CP can game it by not saving all their high performances (or completing events) and just make sure a high proportion of lower performances dominate the model?

2 Likes

Yes and no.

Could they quit an event where they have pushed hard so it doesn’t influence the model? Yes, but then of course they will have no race result.

The model only uses best efforts for any given duration, so low performances cannot dominate the model, because they actually have no impact on the model.

3 Likes

I think the answer is no

1 Like

I think (but happy to be corrected) that the ride does not have to be saved, but would still need to be ended (you get an option at the end to bin it).

If you just quit / shut down your PC etc, the fit file ends up malformed (if anyone that has tried to upload one to Strava can attest).

1 Like

no because for a standard sized person the limits are reasonably liberal, but it probably won’t get in the way of people who aren’t

i worry about lighter riders, but most of them have flat power curves. so i should say i worry about light riders with high anaerobic ability, and there aren’t many of those.

1 Like

Is the test model cruiser proof? My take on it:

Short answer:
No, it isn’t. But you will have to be much more careful both long and short term when you cruise. We will see fewer unintentional cruisers (the ones whining in the forum right now) and casual cruisers. Only the really malicious ones (like myself) will be left. You can probably still cruise though.

The somewhat longer version:
What the test model does is to make cruising the 20 min W/kg much trickier, and for two reasons.

(I’m assuming a simple test model with only 20 min FTP and VO2Max, where VO2Max is just another point on the power curve - how many minutes, how is it defined, does anyone know? I do this just to make the argument simpler.)

Reason no 1 is the move from W/kg to Watt. You can’t e.g. exploit weight differences the same way anymore.

Reason no 2 is you can no longer maintain your rolling 20 min average, made ridiculously simple with the Zwift Monitor app, and still crush opponents in climbs or pushes, because your VO2Max will be off the charts, it will be much higher than your projected VO2Max (one that would ideally suit your 20 min FTP, give or take). Going over the top in VO2Max should push you up to a higher category now.

But there are at least two other options if you want to cruise. (If I ever get the chance next week, I promise, I WILL… and then tell everyone about it, as usual.)

The first thing you can do is to make sure ZHQ never gets a hint at your true VO2Max, just like they never got hold of your true 20 min W/kg when cruising in the past. So you need to groom your race/activity history. Just like in the past, your 90 day ZP averages, only now there is an additional parameter you need to nurture and pamper with.

How would this be helpful to the cruiser? Well, you can’t drop your opponents like a Gauloise butt anymore, but you are still stronger than them. Do enough of managed pushes and you will wear them down eventually if the others are legit. Just keep an eye on the 1 min/5 min/custom meters in Zwift Monitor and make sure you don’t overdo it. I foresee less explosive but at the same time much more aggressive riding next week. Tempo changes FTW.

Secondly, the test model suffers from the same problem as the standard model. It doesn’t account for the implicated power curve. It’s just a row of snapshots. In the past you didn’t have to adapt your later 20 min averages in the race to a sloping power curve. You could go as hard in the last 20 min as in the first. And no one else except a cruiser can do this.

How does this help the cruiser? It’s obvious. You take advantage of the fact that your opponents will drop off in both 20 min performance and VO2Max performance towards the end of the race, whereas you can stay as artificially strong, strong but cat boundary compliant, as in the early race. That’s a huge advantage. And obviously, whether consciously cruising or not, if you have it in you, you will do this.

2 Likes

:rofl: :rofl: :rofl:

Dr-Evil-GIF-source

Yes, after discussion with Flint I’ve explained how it is calculated here: Category Enforcement - How is my category calculated? [February 2022]

This also covers off most of your other points. Your CP is calculated from all efforts from 2m-50m. Your CP also comes with your W’ (anaerobic capacity) which allows for your MAP to be modelled from 5m-8m.

So in effect, every effort of every duration (over 2m) has the possibility to impact the calculation and your assigned cat. So cruisers beware.

2 Likes

15 Likes

Interesting! Well, that makes it even harder to cruise but not impossible. You still have the problem with legit riders dropping off towards the end. That is sooo exploitable. But yes, many cruisers will get accidental upgrades by overdoing something somewhere along the power curve, and then be forced to go on a 60 day cooldown.

I’m actually not so worried about the MAP part with regards to cruising prospects. The two most common types of cruisers in the past was the heavy guys, who basically just cruised their weight advantage, and then the somewhat woke, somewhat malicious type who was typically not that heavy. And the latter weren’t necessarily very explosive, just way stronger than you. Blasting oppontents is for sandbaggers, it’s not cruiser style. I’m more worried about taking a whole power curve into account and make sure you don’t overdo it anywhere over its length. That’s the hard part. Information overload. You’re gonna need some serious overlays on the Zwift screen with numbers flashing all over… But it all depends on sensitivities. How much can you deviate from your assumed/assigned ideal power curve in one spot before getting upgraded?

1 Like

Fancy giving it a go? If it proves too much hassle for you, we can call it a success.

4 Likes

I’m awfully tempted! :smiling_imp: I just think I will be too weak to cruise my cat next week. I’d need another month. I’m merely at the top of cat. Got no cherry to put there right now.

There is a sweetspot, no, a sweetrange, when cruising. Too weak and you can’t cruise of course. Too strong and it gets hard to gauge your effort. You want to be able to feel in your legs what/how you’re doing or you might easily overdo it. As an example, Pogacar would SUCK as a cruiser in cat C. He’d get upgraded in his first race.

So if the sensitivities over the power curve are reasonably generous, then you can build up an intuition, a sort or proprioceptive intuition, of how hard you are allowed to cruise. Enough generosity in the system and there will be a lot of “unintentional” cruisers who get the hang of this. A really harsh system and people will get upgraded quite unexpectedly and feel confused and angry about it. Hard to balance. ZHQ are making rods for their own backs, do they realize this?

That’s one reason why I tend to prefer points as a solution. That is quite obviously cruiser-proof by construction: if you win regularly enough, you get promoted, regardless of how hard you had to pedal. It also guarantees (again, by construction) that all riders will get to be at the fast and slow end of races, if they race regularly. With even (especially?) the best ranking system, organisers have to be sufficiently motivated and involved to change the pen boundaries, or else you just get the same race, time after time.

I do accept however that different systems have pros and cons. I believe points will tend to generate a broader spread in abilities within a race field on any given day. If it is important to minimise the number of riders that get dropped early, that’s a bad thing. It comes back to my general question of: what is actually the goal here? If you care about a precise ranking and tightly matched races all the time, a good ranking system is probably best. I’m not convinced that those goals are truly important, and I’ve ridden in my share of very poorly matched races, front and back. What does start to grate is the tedium of the same race week after week, especially when coupled to the cruising phenomenon.

3 Likes

Couldn’t agree more. That’s why I suggest we leave the question of ‘what is the goal here?’ to the race organiser. Provide them with options. See that tedium disappear…

1 Like

IMPLICATIONS OF THE TEST CAT MODEL
Each cat model has its implications for how you should race cat B-D if you want to podium. There is an ideal way of racing. Let’s look at three models: our current one, a results-based model and the test model.

Current model
The optimal way to race the current model is to be heavy set (naturally large muscle volume) and to cruise the race. The model promotes that type of racing because heavy cruisers have a clear advantage. They will come out on top. You get rewarded with podiums for holding back, for NOT doing your best effort. And you can keep winning like this indefinitely in your cat.

Results-based model
Your placings (or similar) determine your cat. You can race any way you like. Races could vary in texture and effort. Some may go slow at first and pick up at the end, or perhaps the other way around. It doesn’t matter. Or it depends, from race to race, what you need to do to get on the podium. You go as hard as you need, whatever that means, it will depend. You need to be flexible. Oh, and you can’t dodge the upgrade if you want to podium a lot. But it’s never forbidden to smash your opponents to pieces. If you have it in you, why shouldn’t you?

The test model
This one is the opposite of the current model (the results-based model isn’t, it’s rather completely disjoint). Let’s assume a perfect model. Perfectly balanced. No room for cruising. No unfair advantages for certain groups of riders. Everybody is happy. But what does that mean? What kind of happiness are we looking at?

We are looking at races that would push participants to their limit. All else equal (infinite attendance, everyone equally tactically skilled etc), with everyone perfectly placed in each of the perfect categories, then life at the top of any cat would be pure zone 4/zone 5. Max effort. Every time. If you don’t do a max effort, someone else will. And if that should put them out of cat, then that is where they belong. Just below them in the classification there will be someone else who didn’t overstep the boundaries but was nevertheless in a max effort. If he wasn’t (i.e. he was cruising), then this perfect system would pick up on it and put him in the next category too. If someone wins without a max effort, then the system isn’t perfect. Then there is an exploitable flaw.

Everyone (Zwift included) needs to understand this. I much prefer a results-based system myself. It’s the only sensible option. But I also much prefer the test system if the only other option was the current system. Compared to what we have, I am fine with brutally hard racing, if there is no alternative. Are you?

4 Likes

That’s not true. Plenty a bike race is won without doing a max effort both irl and Zwift. There will be points you are pushing to max effort for certain durations but overall it’s quite a possible a race is not near your max effort or near the cat boundaries. Obviously this should not be the case every race or it would be just like the cruising issue we have right now.

3 Likes

@Gordon_Rhino-Racing , you can’t compare to IRL because IRL is using results-based categorization. And you can’t compare to Zwift as it stands today, because that is the current model that I describe. In the test model that they are planning, provided it’s perfect, then it will always be advantageous to drive up the pace. If you are at the very top of the category, then you don’t want to give someone a bit weaker the chance to keep up.

Sure, you don’t want to take wind, so I suppose you could end up in some kind of Nash equilibriums where people hold back temporarily because they don’t want to be the first one to start pushing (i.e. Zwift becomes a pure sprint game). But my point is, races will strive towards max effort from those at the top, whereas our current racing strives in the opposite direction. And a results-based system strives in… whatever direction is needed to win, it could vary.

3 Likes

James and James
You are both putting in some great work here on how categories are decided.
One likes the new Power Metrics being tested next week and the other favours a race ranking system.
You are both arguing for you own preferred choice as if the two are mutually exclusive - they are not.
IMHO, and mentioned by one or two others, the use of both at the same time is the best option.
So if a rider has hit certain power metrics they get promoted OR if they have achieved a specific Race Ranking they get promoted.
Power Metrics are being tested and ZwiftPower already has a pretty good. IMO, Race Ranking system in place they just need to eventually be married up.

So will this stop cruising - I don’t know but probably better than just using one category selection criteria.

Will this stop cheating - no almost certainly not cheaters are nearly always ahead of the game. So my solution to that is don’t encourage them into your regulated racing set up. Don’t force them to race under your rules, give them their own races. (Apologies if I am supposed to use the word sandbagging but I like to call it as it is) I believe this system works quite well in the body building world of drug free competition and open steroid use competition.

It’s a Win, Win, Win situation.
We get regulated racing with only those prepared to sign up and follow the rules and category criteria.
Cheaters/Sandbaggers - they have their own races and are happy. (Benefit is that regulated riders could also take part if it suits what they want out of the race)
Zwift doesn’t lose any customers because everyone is happy

Utopia (or is that Watopia) ! - I’m not that naive as to believe I have solved all Zwift’s problems but thought I would throw my two penn’oth into the ring.