Trade Polarity: Convergence and Divergence
#arbitrage#spread#funding
An arbitrage bundle can be opened in two directions: on the price difference narrowing or on it widening. In Arbitrageur this direction is called the trade's polarity.
- Convergence — trading on prices coming together and the spread narrowing.
- Divergence — trading on prices moving apart and the spread widening.
The polarity you choose determines on which exchange the long is opened and on which the short. The same pair of markets can be used for both directions.
Convergence: from a larger spread to a smaller one
With convergence, we open:
- a short in the market with the higher price;
- a long in the market with the lower price.
The goal is to close the bundle after the difference between the prices narrows. Full price equalisation is not required: a move within a certain range can be enough for the trade.
Convergence example
Suppose the same asset is represented by linear contracts settled in USDT on two exchanges:
| Exchange A | Exchange B | |
|---|---|---|
| Entry price | 100 USDT | 102 USDT |
| Position | Long | Short |
(102 − 100) ÷ 100 × 100% = 2%
For the example, let's take a lot of 5,000 USDT at exchange A's price.
5,000 ÷ 100 = 50 coins
On both exchanges we open positions equivalent to 50 coins. The quantity of the underlying asset is the same, but the value of the positions differs because of the price difference:
| Exchange A | Exchange B | |
|---|---|---|
| Entry action | Open long — buy | Open short — sell |
| Volume | 50 coins | 50 coins |
| Average price | 100 USDT | 102 USDT |
| Opening turnover | 5,000 USDT | 5,100 USDT |
Here 5,000 USDT is the lot size at the chosen reference price, not the total deposit or the margin for the two positions.
Later the price on exchange A stays at 100 USDT, while on exchange B it falls to 100.50 USDT. The spread narrows to 0.5%, and we fully close both positions:
| Exchange A | Exchange B | |
|---|---|---|
| Exit action | Close long — sell | Close short — buy |
| Volume | 50 coins | 50 coins |
| Average price | 100 USDT | 100.50 USDT |
| Closing turnover | 5,000 USDT | 5,025 USDT |
Long: (100 − 100) × 50 = 0 USDT
Short: (102 − 100.50) × 50 = 75 USDT
Bundle result: +75 USDT
All calculations use the average execution prices of the full volume.
For clarity, only one price changed in the example. In real trading both can move. What matters for such a bundle is that the price change brings more profit on one leg than loss on the other.
Why convergence is the base scenario of arbitrage
The classic logic of price arbitrage is built on buying the cheaper instrument and selling the more expensive one.
In an efficient market, the same asset under comparable conditions should have a similar value on different venues. In an ideal model without fees, restrictions and delays, the price ratio tends towards 1:1.
A gap creates an incentive to buy where it is cheaper and sell where it is more expensive. These actions push prices closer together. That is why convergence is the natural base scenario for trading price gaps.
But the pull towards equilibrium does not mean the spread must return to zero. Venue restrictions, differences in liquidity and in contract terms can sustain a price difference for a long time.
That is exactly why a trader needs to assess the actual behaviour of the specific bundle.
Divergence: from a smaller spread to a larger one
With divergence, the direction of the positions is reversed:
- a long in the market with the higher price;
- a short in the market with the lower price.
The goal is to close the bundle after the price difference widens.
At first glance, buying the more expensive instrument and selling the cheaper one may look illogical. But here the result depends on how the difference between them changes next.
Divergence example
We use the same markets, but start with a smaller gap:
| Exchange A | Exchange B | |
|---|---|---|
| Entry price | 100 USDT | 100.50 USDT |
| Position | Short | Long |
(100.50 − 100) ÷ 100 × 100% = 0.5%
5,000 ÷ 100 = 50 coins
| Exchange A | Exchange B | |
|---|---|---|
| Entry action | Open short — sell | Open long — buy |
| Volume | 50 coins | 50 coins |
| Average price | 100 USDT | 100.50 USDT |
| Opening turnover | 5,000 USDT | 5,025 USDT |
Then the price on exchange A stays at 100 USDT, while on exchange B it rises to 102 USDT. The spread widens to 2%, and we fully close the bundle:
| Exchange A | Exchange B | |
|---|---|---|
| Exit action | Close short — buy | Close long — sell |
| Volume | 50 coins | 50 coins |
| Average price | 100 USDT | 102 USDT |
| Closing turnover | 5,000 USDT | 5,100 USDT |
Short: (100 − 100) × 50 = 0 USDT
Long: (102 − 100.50) × 50 = 75 USDT
Bundle result: +75 USDT
In the first example the profit came from the gap narrowing. In the second — from it widening.
| Convergence | Divergence | |
|---|---|---|
| More expensive market at entry | Short | Long |
| Cheaper market at entry | Long | Short |
| Expected movement | Spread narrows | Spread widens |
The market labels refer to the moment the direction is chosen. After opening, the positions keep their roles even if the prices later swap places.
When divergence is used
Let's look at two scenarios: collecting funding payments and trading recurring spread swings.
Collecting funding payments
Funding is a periodic payment between holders of longs and shorts in perpetual contracts.
With a positive rate, longs usually pay shorts. With a negative rate, shorts pay longs. So the direction of the positions can be chosen based on which combination gives a positive sum of payments across the bundle.
For example:
- in the more expensive market the rate is negative — holders of longs receive the payment;
- in the cheaper market the rate is positive — holders of shorts receive the payment.
A long in the first market and a short in the second form divergence polarity. But the main goal of such a structure may be collecting funding rather than waiting for the spread to widen.
Divergence on its own does not guarantee favourable funding: everything is determined by the rates, the value of the positions and the payment schedule. In another situation, convergence may be the right choice for collecting payments.
When assessing the result, the change in the price gap must also be taken into account. The funding received may be partly or fully offset by a loss from the spread's movement.
Trading recurring spread swings
Over certain stretches of their history, some bundles repeatedly go through a similar cycle: the prices come together, then move apart, and come together again.
If the chart shows such behaviour, trading in both directions can be considered.
| Stage | Action |
|---|---|
| Spread around 0.5% | Open divergence |
| Spread widens to 2% | Close divergence |
| Spread around 2% | Open convergence |
| Spread narrows to 0.5% | Close convergence |
This way both parts of the cycle can be traded: the spread widening and narrowing.
Each completed pass includes opening and closing two legs. Therefore the range must cover the trading fees and leave enough trade potential.
Recurrence in history is a basis for assessing a scenario, but not a guarantee of the next cycle. The spread may leave its usual range or stay at one of its bounds for a long time.
Why a 50% spread does not mean you can earn 50%
A large spread draws attention, but on its own it says nothing about what part of the movement can be captured.
Suppose the price difference is currently 50%. One might assume the prices must fully equalise and open a convergence trade expecting a move to zero.
But history shows a different picture: recently the spread has only fluctuated between 45% and 50%. It has not dropped to zero.
50% − 45% = 5 percentage points
History gives grounds to consider movement within this range. Expecting full price equalisation requires separate justification.
For example, for convergence you can consider entering closer to the upper bound and exiting closer to the lower one. For divergence — the reverse scenario. The levels do not have to be placed exactly at historical extremes: the market may not reach them again.
At the same time, 5 percentage points of spread movement are not a guaranteed 5% return on capital. The money result depends on the execution prices and volumes of both legs, the fees and the payments while the positions are held.
The main question is not “how big is the spread now” but “how does it move, and which part of that movement do we expect to capture”.
Automation in Arbitrageur
In the Arbitrageur terminal you can set entry and exit levels for both convergence and divergence.
You choose a bundle, the polarity and the lot size, then place trigger levels on the chart. The terminal monitors the conditions and automatically executes the entry and exit on the parameters you set — without having to send every order manually.
This makes it possible to automate trading selected parts of the spread's movement. For example, set up a convergence entry when the gap widens to a given level and an exit after it narrows. For divergence, set the reverse sequence.
The Arbitrageur chart is built on calculations from the order books for your lot. So when choosing levels, you rely on the weighted average prices of available execution, not just on the last trade prices.
You define the direction and the terms of the trade, and Arbitrageur takes over monitoring the levels and executing the actions you set. If the bundle's behaviour changes, the levels and the chosen scenario need to be revised: automation carries out the trading logic, but it does not make a historical range permanent.